<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom">
<channel>
  <title>SendTech Times - Latest News</title>
  <link>https://stechtimes.com</link>
  <atom:link href="https://stechtimes.com/rss.xml" rel="self" type="application/rss+xml" />
  <description>AI, semiconductors, cloud, cybersecurity and Gulf technology markets from SendTech Times.</description>
  <language>en</language>
  <lastBuildDate>Sat, 20 Jun 2026 11:52:04 GMT</lastBuildDate>
  <ttl>15</ttl>
<item>
  <title>Iran’s World Cup Becomes A Test Of AI Media And Fan Mobilization</title>
  <link>https://stechtimes.com/en/article/irans-world-cup-becomes-a-test-of-ai-media-and-fan-mobilization-mqmas8sj</link>
  <guid isPermaLink="true">https://stechtimes.com/en/article/irans-world-cup-becomes-a-test-of-ai-media-and-fan-mobilization-mqmas8sj</guid>
  <pubDate>Sat, 20 Jun 2026 11:52:04 GMT</pubDate>
  <category>sports</category>
  <description><![CDATA[Iran’s 2026 World Cup campaign has turned into a digital contest, with pro-regime AI videos, activist web apps and coordinated social posts showing how global sport can become infrastructure for political messaging.]]></description>
  <enclosure url="https://stechtimes.com/_article-images/4c64112feb2d3439-iran-s-world-cup-becomes-a-test-of-ai-media-and-fan-mo-1781956319283.jpg" type="image/jpeg" />
  <content:encoded><![CDATA[World Cup Attention Has Become A Digital Megaphone
Iran’s 2026 World Cup campaign is showing how global sport can become a distribution system for political media, not just a stage for matches. The contest around the team is taking place on social platforms, in AI-generated videos and through web tools designed to turn fans’ phones into coordinated displays.

The clearest example came minutes before Iran’s opening match against New Zealand on Monday. Explosive Media, a group associated with pro-Iran AI videos that resemble The Lego Movie, released a clip that placed Iran’s players beside schoolchildren who were killed in a US attack in March. A hashtag tied to the school in Minab then began trending globally on X, with a message that read: “With you from the stadium.”

The speed of the spread is the important technology point. Within less than one hour, the video was viewed millions of times and drew hundreds of thousands of shares and comments. Iran’s state-linked news outlets quickly amplified videos when Iran scored, and Iranian embassies also shared the material, extending the campaign beyond ordinary fan discussion.

That amplification chain matters because it joins layers that usually get treated separately: synthetic video production, hashtag distribution and official diplomatic reposting. The result is a campaign that can begin as fan-facing content but quickly gain the reach and authority of state-aligned channels.

Activists Are Building Around Platform Rules
Opposition activists are using a different technical playbook. A Fifa rule kept Iran’s pre-Islamic Revolution flag out of World Cup matches, so activists created a web-based app called IranSync. The tool lets fans synchronise phones to show a digital version of the older monarchist flag.

One creator said IranSync was developed by Iranian activists who volunteer their time and expertise to build tools that amplify Iranian voices. The app also lets users display messages aligned with Reza Pahlavi, the exiled son of Iran’s last shah. Its homepage describes the goal as turning smartphones into one unified display.

After IranSync gained attention, the same activist network created another web portal, IranUp. That site is built for coordinated participation on X during one of the world’s biggest media events. It offers pre-written messages on anti-regime themes including democratic leadership, human rights, cultural erasure and regime overthrow, with quick-sharing links for X, Telegram, WhatsApp, Facebook, LinkedIn and Truth Social.

That design matters because it moves online activism from spontaneous posting into a more structured interface. The activists are not only writing slogans; they are building lightweight tools that reduce friction, standardise messages and time participation around moments when global attention is already concentrated.

The phone-display tactic also shows how a stadium restriction can be challenged without physically bringing a banned symbol through the gate. Some fans still managed to bring the older flags into the stadium, but the web app created another route: make the screen itself the sign.

The Match Is Only One Layer Of The Story
The sports result became part of the media cycle, but it was not the whole story. Iran drew 2-2 with New Zealand in Los Angeles on Monday. After the match, Fifa president Gianni Infantino visited the Iranian team in the locker room, and a social video showed him telling the players they were “writing history” and that the whole world was watching.

For political media operators, that kind of attention is valuable because it turns a sports moment into a shareable proof point. For platform companies and tournament organisers, it creates a harder governance problem. The same systems that help fans coordinate support can also be used to push state-linked messaging, exile politics or emotionally charged AI media at high speed.

The next watchpoint is Iran’s match against Belgium on Sunday. The source does not say whether Fifa or social platforms plan new enforcement steps before then. What is already clear is that the World Cup has become more than a tournament feed for this story. It is a live test of how AI media, fan apps and coordinated posting can compete for attention under the cover of global sport.]]></content:encoded>
</item>
<item>
  <title>GoPro’s Slide Shows How AI Hardware Is Tilting Toward Chinese Brands</title>
  <link>https://stechtimes.com/en/article/gopros-slide-shows-how-ai-hardware-is-tilting-toward-chinese-brands-mqm8gz7o</link>
  <guid isPermaLink="true">https://stechtimes.com/en/article/gopros-slide-shows-how-ai-hardware-is-tilting-toward-chinese-brands-mqm8gz7o</guid>
  <pubDate>Sat, 20 Jun 2026 10:47:22 GMT</pubDate>
  <category>chips-semiconductors</category>
  <description><![CDATA[GoPro’s market-share collapse and iRobot’s loss of Roomba control show how Chinese hardware makers are challenging Western consumer-tech pioneers as AI features move into cameras and home robots.]]></description>
  <enclosure url="https://stechtimes.com/_article-images/4f7c979a0300ce3b-gopro-s-slide-shows-how-ai-hardware-is-tilting-toward-1781952434548.jpg" type="image/jpeg" />
  <content:encoded><![CDATA[GoPro’s Camera Lead Has Become A Hardware Warning
GoPro’s fall in action cameras is no longer only a story about one company missing a product cycle. It has become a warning about how quickly Chinese consumer-hardware makers can turn a category pioneered in the United States into a tougher fight over manufacturing scale, product speed and AI-enabled features.

The source example is stark. GoPro launched the action-camera category in 2002 and still held an estimated 75% of the global market in 2022. In early June, the company warned in a regulatory filing that its future operations were in question. Its share now sits near 18%, a three-year collapse that shows how little room was left after weaker sales, memory-cost pressure and rivals with faster hardware refresh cycles.

The shift is visible in the brands consumers now encounter. DJI and Insta360, both Chinese companies, are described as the two names repeatedly recommended by divers, travel creators and outdoor users. Together, they now account for over 80% of the action-camera market. The competitive lesson is not only price. Faster model rollouts and AI-powered editing features make the camera category depend on software experience as much as rugged hardware.

For GoPro, that means the memory-chip shortage landed on a business that was already exposed. The company’s own filing pointed to rising costs, but the market history shows a longer squeeze: smartphones reduced the need for a separate camera for many users, product bets failed to restore momentum, and Chinese brands kept improving the devices most visible to creators and outdoor buyers.

GoPro cited rising costs from the global memory chip shortage. The broader pressure, however, is competition from DJI and Insta360 after years of sales declines and failed product bets. GoPro peaked in 2015, DJI entered action cameras in 2019, and Insta360’s arrival in 2022 intensified the pressure as the Chinese companies moved faster with new models and AI-powered editing features.

Robot Vacuums Show The Same Competitive Pattern
The same pattern is appearing in home robotics. Chinese robot-vacuum brands including Dreame and Roborock have become major players in a category once defined by iRobot and its Roomba line. iRobot, which invented the home cleaning robot that became Roomba, struggled to keep its lead, filed for bankruptcy in December last year and was acquired by its Chinese manufacturer, Picea Robotics.

That comparison matters because action cameras and robot vacuums are not identical products. One is built around creators, sports and travel; the other around autonomous home cleaning. Yet both markets show Western pioneers losing advantage as Chinese manufacturers compete through faster product cycles, lower-cost hardware execution and increasingly software-driven features.

The AI angle is not that every device becomes strategically important overnight. It is that more consumer hardware now depends on cameras, sensors, microphones, mapping, editing tools and automated decision-making. As these capabilities move into everyday devices, hardware competition starts to look less like a simple gadget fight and more like a contest over who controls intelligent products inside homes, vehicles and creative workflows.

Security Concerns May Become The Western Counterweight
Chinese consumer-tech gains are also running into policy and trust limits. Robot vacuums have drawn data-security concerns because they use cameras, sensors, microphones and detailed maps of users’ homes. The source also points to limits already placed on other China-made technology products in the United States, including drones and electric vehicles.

Those concerns create a possible opening for Western brands, but not an easy one. GoPro is exploring a pivot toward defense and aerospace markets, a move that suggests security-sensitive buyers may value trusted hardware even when consumer buyers choose faster or cheaper products.

The unresolved test is whether that narrower trust advantage can compensate for lost consumer scale. DJI, Insta360, Dreame and Roborock show that Chinese brands can compete beyond low-cost manufacturing when product cycles, sensors and AI features matter. For Western consumer-tech pioneers, the harder question is whether security positioning is enough after the mainstream gadget market has already moved.]]></content:encoded>
</item>
<item>
  <title>Apple&apos;s Memory Warning Shows AI Server Demand Reaching Consumer Devices</title>
  <link>https://stechtimes.com/en/article/apples-memory-warning-shows-ai-server-demand-reaching-consumer-devices-mqm4cdf1</link>
  <guid isPermaLink="true">https://stechtimes.com/en/article/apples-memory-warning-shows-ai-server-demand-reaching-consumer-devices-mqm4cdf1</guid>
  <pubDate>Sat, 20 Jun 2026 08:51:46 GMT</pubDate>
  <category>chips-semiconductors</category>
  <description><![CDATA[Apple is preparing customers for higher prices as AI data-center demand tightens memory supply, exposing how HBM demand from Nvidia-class systems can raise costs for phones, PCs and tablets.]]></description>
  <enclosure url="https://stechtimes.com/_article-images/2df6285642338689-apple-s-memory-warning-shows-ai-server-demand-reaching-1781945500616.jpg" type="image/jpeg" />
  <content:encoded><![CDATA[AI Memory Demand Is Moving Into Device Prices
Apple's warning on memory costs turns the AI infrastructure boom into a consumer hardware problem. The pressure is not only inside data centers. It is also reaching smartphones, PCs and tablets that rely on the same limited supplier base for memory capacity.

Apple CEO Tim Cook told the Wall Street Journal that product price increases are planned because of ongoing memory shortages. He did not say when the increases would begin or which devices and models would be affected. Apple also declined to comment to CNBC, leaving the immediate pricing map unclear.

That uncertainty is the central business tension. Apple has usually been viewed as better protected than other device makers because of its scale, supplier relationships and pricing power. Ranjit Atwal, an analyst at Gartner, said the situation shows the depth of the problem and that even Apple cannot fully shield itself from the shortage.

The supply squeeze begins with AI hardware. Nvidia-class data-center chips use high-bandwidth memory, and that demand is pulling capacity toward AI servers. Consumer device makers still need DRAM and NAND for phones, PCs and other products, but they must compete for output from the same memory industry. The article identifies Micron, SK Hynix and Samsung as the three primary suppliers behind that capacity.

HBM Economics Create A Consumer Trade-Off
The reason this shortage is difficult to absorb is that suppliers face a production trade-off. When Micron makes one unit of HBM, it gives up three units of more conventional smartphone memory. HBM is also described as more profitable, so new capacity may continue to favor AI systems even as fabs are built.

That makes the memory shortage different from a normal component cycle. Data-center AI processors rely on HBM, a faster and more power-hungry memory class than the DRAM used in phones. One Nvidia Blackwell B200 chip carries 192GB of HBM, while large clusters can reach over 2,000 servers. An iPhone, by contrast, uses 8GB or 12GB of DRAM, so the gap between server memory demand and handset memory needs is enormous even before supplier allocation decisions are made.

Those figures explain why consumer devices are exposed to AI buildouts even when they do not use the same memory type. The capacity decision happens upstream. Suppliers must decide which products receive wafers, tools and factory output, and the most profitable AI memory can crowd out conventional memory needed by handset and PC makers.

Francisco Jeronimo, an IDC analyst, framed the effect as consumers paying for AI before they receive the full benefit of on-device AI. Apple is putting more RAM into each phone for AI features, while some newer Siri and dictation capabilities are expected to be limited to newer iPhones, iPads and Macs because older or less expensive devices cannot handle the memory-hungry options.

Apple Still Has Pricing Choices
Several Apple responses remain possible, but there is no confirmed price list. Jeronimo expects the $999 iPhone Pro and the $1,199 iPhone Pro Max to absorb the increase pressure, with lower-end devices left unchanged. BofA Securities analysts also expect price increases for most Mac and iPad models.

Apple could also use the shortage competitively. It has recently targeted budget-conscious buyers with the $599 MacBook Neo and $599 iPhone 16e. Some analysts expect Android manufacturers to cut specifications or raise prices, while IDC expects average smartphone prices to increase by 20% this year.

That creates a strategic fork. Apple can protect margins by raising prices on premium devices, use budget models to pressure Android rivals, or spend from its balance sheet to help expand supply. Cook said Apple is willing to use its balance sheet as part of the solution, but the source does not describe a specific supply deal.

The next test is whether memory shortages remain a data-center cost story or become a broader consumer electronics reset. If AI server demand keeps pulling capacity toward HBM, device makers may have to explain why phones and PCs cost more before users see the promised benefits of on-device AI.]]></content:encoded>
</item>
<item>
  <title>Finastra’s Banking Unit Sale Tests Whether Core Modernization Needs A Separate Owner</title>
  <link>https://stechtimes.com/en/article/finastras-banking-unit-sale-tests-whether-core-modernization-needs-a-sep-mqm262e5</link>
  <guid isPermaLink="true">https://stechtimes.com/en/article/finastras-banking-unit-sale-tests-whether-core-modernization-needs-a-sep-mqm262e5</guid>
  <pubDate>Sat, 20 Jun 2026 07:50:56 GMT</pubDate>
  <category>uae-economy</category>
  <description><![CDATA[Pollen Street Capital plans to acquire Finastra’s Universal Banking business, giving the core banking unit independent ownership while Finastra narrows its focus to payments and lending.]]></description>
  <enclosure url="https://stechtimes.com/_article-images/ce0111b001930182-finastra-s-banking-unit-sale-tests-whether-core-modern-1781941848004.jpg" type="image/jpeg" />
  <content:encoded><![CDATA[A Core Banking Unit Is Being Pulled Out Of Finastra

Pollen Street Capital plans to acquire Universal Banking, the global core banking software business inside Finastra. The transaction would turn Universal Banking into an independent business led by its existing management team, while Finastra sharpens its focus on payments and lending.

The move is not just a portfolio reshuffle. Universal Banking supplies core banking technology for account and deposit management, payments, lending and treasury operations across retail, commercial and corporate banking. That makes the deal a test of whether a business built around bank modernization can move faster with a dedicated owner rather than remaining one unit inside a broader financial-software group.

Universal Banking Brings Customers, Countries And A Platform

Universal Banking supports over 150 customers worldwide across more than 100 countries. The customer base includes large and regional financial institutions, digital banks, Islamic banking providers and building societies. Its central platform is Essence, described by Finastra as a cloud-first, open banking platform for banks trying to modernize legacy systems and improve operational efficiency.

Those details explain why the buyer is not simply acquiring a software product line. Core banking systems sit close to deposits, customer accounts and payment flows, so replacement or modernization usually has to balance new capability with operational continuity. The transaction therefore points to a practical modernization problem: banks want newer platforms and data capabilities, but they also need older and newer systems to coexist while critical operations continue.

Finastra says Universal Banking helps customers manage that transition through technology designed for progressive modernization rather than a single disruptive replacement. That claim is important because core banking buyers rarely judge a platform only by product ambition. They also need delivery discipline, migration support and credible continuity planning before moving functions that affect deposits, loans, payments and treasury operations.

Pollen Street Is Buying Into The Modernization Gap

Pollen Street says its backing will support Universal Banking as a standalone company, with investment aimed at product innovation, GenAI and data capabilities, customer delivery and broader capabilities for banks. The company did not disclose the purchase price or give a closing date.

Finastra chief executive Chris Walters framed the sale as a way to give Universal Banking dedicated focus and investment, while allowing Finastra to concentrate on payments and lending. Pollen Street partner Anastasia Kovaleva described Universal Banking as a business with longstanding customer relationships and a modern platform, and said the firm wants to support AI-led innovation and customer modernization.

The Watchpoint Is Execution After Separation

The transaction remains subject to customary regulatory approvals. Arma Partners advised Finastra and Vista Equity Partners, while Kirkland & Ellis served as legal adviser. Nomura advised Pollen Street, and Clifford Chance served as legal adviser.

Pollen Street was established in 2013 and says it manages over €8bn in assets under management across private equity and credit strategies. It also says it has over 90 professionals. For Universal Banking customers, however, the important question is narrower: whether independent ownership translates into faster delivery, clearer product investment and safer modernization of systems that banks cannot afford to interrupt.]]></content:encoded>
</item>
<item>
  <title>Instacart’s Grocery AI Rollout Tests Whether Agents Can Build Baskets Without Breaking Trust</title>
  <link>https://stechtimes.com/en/article/instacarts-grocery-ai-rollout-tests-whether-agents-can-build-baskets-wit-mqm08qtz</link>
  <guid isPermaLink="true">https://stechtimes.com/en/article/instacarts-grocery-ai-rollout-tests-whether-agents-can-build-baskets-wit-mqm08qtz</guid>
  <pubDate>Sat, 20 Jun 2026 06:56:59 GMT</pubDate>
  <category>ai</category>
  <description><![CDATA[Instacart has rolled out an AI shopping assistant to millions of U.S. customers, with U.S. and Canada expansion planned in the coming months.
The assistant turns prompts, photos and deal requests into carts using live inventory from nearly 100,000 stores and data from more than 1.6 billion lifetime orders.
The tension is whether larger baskets and personalization can scale while customers still review every decision before checkout.]]></description>
  <enclosure url="https://stechtimes.com/_article-images/3cfc3735fc32fd79-instacart-s-grocery-ai-rollout-tests-whether-agents-ca-1781938612795.png" type="image/jpeg" />
  <content:encoded><![CDATA[Instacart is moving agentic commerce from a partner-platform promise into its own grocery marketplace, rolling out an AI assistant that can turn a meal idea, shopping prompt or uploaded list into a ready-to-review cart.

Instacart Turns Grocery Search Into Cart Building
The assistant is now available to millions of U.S. customers through Instacart’s app and website. A full rollout across the U.S. and Canada is planned in the coming months, after months of testing.

The product is built for a specific retail problem: grocery recommendations fail quickly when items are out of stock, household preferences are ignored or substitutions feel random. Instacart says the assistant uses grocery-specific machine learning, live inventory and customer history to build baskets that reflect what is available at a chosen retailer.

Users can ask for easy weeknight dinners, deals on usual items, a party menu or a cart based on a handwritten list. The assistant can generate meal ideas with shoppable ingredient lists, translate a list photo into items, surface promotions and adapt suggestions to brand preferences and dietary choices.

The Data Claim Is The Competitive Moat
Instacart’s case rests on the depth of its grocery data. The company says the assistant draws on more than 1.6 billion lifetime orders and live inventory from nearly 100,000 stores across North America. It also points to signals from household behavior, millions of consumers and thousands of retail banners.

Those figures matter because agentic shopping is only useful if the agent understands availability, substitutions, recurring purchases and price sensitivity. A general chatbot can suggest dinner; a grocery agent has to convert that suggestion into an in-stock basket that a household is willing to buy.

Instacart also frames the assistant as part of a wider AI strategy for retailers and advertisers. The same grocery intelligence is being extended through enterprise tools, Storefront Pro, ads optimization tools and integrations with ChatGPT, Claude and Gemini.

Early Usage Points To A Basket-Size Bet
The company says early testing showed customers using the assistant for more complex tasks such as recipe discovery and meal planning, not only faster search. Orders placed with the AI assistant are, on average, larger than Instacart’s typical basket, while the platform’s average order value is $113.

Survey data included with the launch explains the demand side of the bet. An Instacart-commissioned Harris Poll of more than 2,000 U.S. adults found that 83% say deciding what to make for dinner causes stress. The same survey found more than two in three would be interested in an AI-powered assistant for that task, while just 8% currently use AI for meal planning or grocery shopping.

The commercial question is whether higher-value baskets come from genuine convenience or from nudging customers toward more items. Instacart says the tool does not finalize anything without explicit action and that every decision is reviewed before checkout. That review step is important if agentic shopping is to feel like help rather than automated upselling.

Rollout Risk Is About Trust, Not Only Accuracy
Grocery is a difficult test case for consumer AI because mistakes are personal and immediate. A wrong size, missing ingredient, unsuitable brand or unavailable item can make the whole basket less useful.

That is why Instacart’s strongest claim is not simply that the assistant can generate carts in seconds. The bigger test is whether it can keep improving from customer accepts, refinements and rejections while preserving user control. If customers trust the assistant enough to use it for weekly planning, agentic commerce moves closer to a repeat habit. If they feel the cart needs too much correction, the feature becomes another search layer with a conversational interface.]]></content:encoded>
</item>
<item>
  <title>Revolut’s UAE Licenses Put Its Super-App Model Into A Regulated Gulf Test</title>
  <link>https://stechtimes.com/en/article/revoluts-uae-licenses-put-its-superapp-model-into-a-regulated-gulf-test-mqly0r7j</link>
  <guid isPermaLink="true">https://stechtimes.com/en/article/revoluts-uae-licenses-put-its-superapp-model-into-a-regulated-gulf-test-mqly0r7j</guid>
  <pubDate>Sat, 20 Jun 2026 05:54:46 GMT</pubDate>
  <category>uae-economy</category>
  <description><![CDATA[Revolut has completed UAE licensing for stored value facilities and Category II retail payment services, setting up a local launch that will test whether its multi-currency app can translate global banking ambition into a tightly regulated Gulf payments market.]]></description>
  <enclosure url="https://stechtimes.com/_article-images/e12e2c5ed75a6efd-revolut-s-uae-licenses-put-its-super-app-model-into-a-1781934880929.webp" type="image/jpeg" />
  <content:encoded><![CDATA[UAE Approval Moves Revolut From Intent To Launch Prep

Revolut has completed its regulatory licensing process in the United Arab Emirates, securing approval for stored value facilities and Category II retail payment services from the Central Bank of the UAE. The company is preparing a full-scale local launch after receiving in-principle approval in September.

The approval gives Revolut a clearer path to offer a local version of its financial app in one of the Gulf’s most competitive digital-banking markets. Once live, the service is expected to combine multi-currency balances, card controls and domestic and cross-border transfers inside one app for UAE customers.

The UAE Pitch Is Payments Choice, Not Just A New App

The strategic question is whether Revolut can stand out in a market where banks, wallets and exchange houses already compete heavily around remittances, cards and mobile payments. Its pitch rests on combining foreign-exchange utility, payments and card management in one interface rather than selling a single narrow payment feature.

That matters because UAE consumer behaviour gives payments a visible role in merchant choice. PYMNTS Intelligence found that 84% of UAE shoppers said payment options affect their choice of merchant, compared with a global average of 66%. Those figures do not guarantee Revolut adoption, but they show why payment experience is a serious commercial lever in the market. They also make the launch less about app novelty and more about whether Revolut can become a practical payment option merchants and consumers notice.

Regulation Sets The Boundary For Expansion

The licensing also shows how Gulf fintech expansion is shifting from broad market-entry announcements to regulated execution. Revolut cannot rely only on its international brand or user experience; it has to operate within Central Bank of the UAE rules while adapting its global platform to local compliance, payment behaviour and customer expectations.

Ambareen Musa, Revolut’s GCC chief executive, framed the UAE as a forward-looking financial market and linked the launch plan to consumer choice and control over money management. That message fits the UAE’s wider digital-economy agenda, but the proof will come after launch through actual product scope, pricing, customer acquisition and regulatory performance.

The Watchpoint Is Local Depth

Revolut has described itself as moving toward a truly global bank, and the UAE approval supports that direction. The local test is narrower and more practical: whether a global super-app model can become useful enough for residents who already have banking apps, card products and remittance options.

For Gulf fintech, the important signal is not simply that another international player is entering the UAE. It is that regulated digital-payment competition is becoming more crowded, and new entrants must prove they can add daily utility rather than just brand recognition.]]></content:encoded>
</item>
<item>
  <title>Vercel’s Eve Framework Tests Whether Agent Tools Can Escape Shadow AI</title>
  <link>https://stechtimes.com/en/article/vercels-eve-framework-tests-whether-agent-tools-can-escape-shadow-ai-mqlvt8w7</link>
  <guid isPermaLink="true">https://stechtimes.com/en/article/vercels-eve-framework-tests-whether-agent-tools-can-escape-shadow-ai-mqlvt8w7</guid>
  <pubDate>Sat, 20 Jun 2026 04:52:57 GMT</pubDate>
  <category>cloud-data-centers</category>
  <description><![CDATA[Vercel introduced the open-source eve agent framework and Passport controls for employee-built AI apps, putting its developer platform strategy up against enterprise concerns over unmanaged agents, data exposure and cloud cost premiums.]]></description>
  <enclosure url="https://stechtimes.com/_article-images/b74f4254bde1404d-vercel-s-eve-framework-tests-whether-agent-tools-can-e-1781931171733.jpg" type="image/jpeg" />
  <content:encoded><![CDATA[Eve Turns Agent Building Into A Platform Bet

Vercel introduced eve, an open-source agent framework, at its Ship event in London and paired the launch with Passport controls for employee-created AI applications. The move pushes Vercel beyond web deployment into a more contested question for enterprise software teams: whether custom AI agents can be simple enough for builders while remaining visible to IT and security teams.

The framework uses TypeScript and Markdown. In Vercel’s model, an agent is organised as a directory of files that define instructions, skills, model provider, tools, authentication, channels and schedule. Agents run in isolated virtual machines by default, and the code is available on GitHub under the Apache 2.0 license.

Simplicity Is The Pitch, Control Is The Risk

Vercel CTO Malte Ubl presented eve as a way to hide orchestration complexity from developers. The framework is designed to manage lifecycle details such as sandboxes and context-window handling, while builders place configuration in the expected locations and deploy agents through the same Vercel command used for web applications.

That simplicity also creates enterprise risk. If employees can quickly build and run agents, companies need a way to know which tools exist, what data they touch and whether the applications follow internal controls. Passport targets that shadow AI problem by giving organisations a way to bring employee-created AI applications under company governance.

Model Choice And Cloud Economics Stay Open Questions

Eve can connect to model providers through Vercel’s AI SDK, and it can also use Vercel’s AI Gateway, which offers a single endpoint for multiple model providers and can switch providers to improve reliability. Model choice therefore becomes part of the platform story rather than a fixed dependency on one large language model.

Cloud economics are more complicated. Ubl acknowledged that using AWS indirectly through Vercel carries a cost premium, but argued that more efficient use of compute resources can offset it. The claim will matter for teams evaluating whether agent workloads should run through a developer platform, directly with a cloud provider or in an internal environment.

Passport Shows Where Enterprise AI Is Headed

The launch reflects a broader shift in enterprise AI adoption. The first wave focused on giving employees access to chatbots and coding tools. The next phase centres on custom workflows that act on company systems, raising the stakes for authentication, approvals, auditability and data boundaries.

For Vercel, the watchpoint is adoption beyond early developer enthusiasm. Eve needs to prove that a directory-based agent framework is useful in production, while Passport needs to show that governance does not become a bottleneck. If both pieces work, Vercel gains a stronger claim on the enterprise AI development stack; if not, shadow AI remains a problem that simple deployment alone cannot solve.]]></content:encoded>
</item>
<item>
  <title>HSBC’s Gulf IPO Pipeline Tests Whether Market Calm Can Restart Listings</title>
  <link>https://stechtimes.com/en/article/hsbcs-gulf-ipo-pipeline-tests-whether-market-calm-can-restart-listings-mqltf98j</link>
  <guid isPermaLink="true">https://stechtimes.com/en/article/hsbcs-gulf-ipo-pipeline-tests-whether-market-calm-can-restart-listings-mqltf98j</guid>
  <pubDate>Sat, 20 Jun 2026 03:46:06 GMT</pubDate>
  <category>uae-economy</category>
  <description><![CDATA[HSBC Menat chief executive Selim Kervanci said the bank has 45 M&A and IPO mandates across the Gulf and expects listing activity to resume in the fourth quarter.
Companies in food, consumer, retail and technology sectors delayed rather than cancelled listing plans during the US-Iran conflict and wider regional lull, leaving a backlog that depends on Q4 valuation clarity to reopen.
The open question is whether improved sentiment can translate into valuations strong enough for Gulf issuers to reopen equity capital markets.]]></description>
  <enclosure url="https://stechtimes.com/_article-images/3933afd29a8358b6-hsbc-s-gulf-ipo-pipeline-tests-whether-market-calm-can-1781927159883.jpg" type="image/jpeg" />
  <content:encoded><![CDATA[Gulf equity markets are moving from a relief rally to an execution test as HSBC points to a still-intact pipeline of mandates that could restart listings once investors are ready to price risk again.

HSBC Sees a Fourth-Quarter Listing Window
HSBC Menat chief executive Selim Kervanci said the bank has 45 mergers and acquisitions and initial public offering mandates across the Gulf. He expects listing activity to resume in the fourth quarter, after uncertainty tied to the US-Iran conflict and wider regional risk slowed equity capital market transactions and deal flow.

For Gulf exchanges, the key distinction is that issuer demand did not disappear. Kervanci said companies in the food, consumer, retail and technology sectors delayed listing plans, leaving a backlog that still requires valuation expectations and investor appetite to align.

Kervanci estimated that investor sentiment, equity capital market activity and deal flow would need at least one quarter to recover as regional conditions stabilize. The fourth quarter therefore looks less like a guaranteed boom than a financing window that still has to prove buyers and sellers can agree on price.

Delayed Deals Put Valuation Back at the Center
Valuation uncertainty caused the slowdown. Kervanci said uncertainty prevented investors from forming a clear valuation view, which became the main obstacle for equity capital market transactions coming to market.

Gulf listing pipelines have increasingly depended on domestic champions, family businesses and growth companies using public markets to fund expansion. If buyers demand a war-risk discount while sellers still expect pre-conflict valuations, mandates can remain active without turning into completed IPOs.

The 2025 baseline shows why banks and exchanges are watching the restart closely. Dealogic figures put wider regional IPO proceeds at $7.1 billion from 61 listings in 2025, versus $13.1 billion in 2024. The amount raised in 2025 was also described as the lowest since 2020, when companies pulled in $22 billion.

Capital Needs Extend Beyond IPOs
The banking angle is wider than listings alone. Kervanci said lenders such as HSBC are expected to be central to postwar efforts to rebuild infrastructure across the Middle East. He said the region will need double-digit billions of dollars, or more, for damaged infrastructure, diversification plans and projects that reduce dependence on the Strait of Hormuz.

Kervanci pointed to spending across infrastructure schemes, AI projects, electricity storage, batteries and port capacity, although he said immediate preference would be for hydrocarbons infrastructure, especially oil and gas pipeline projects. Banks could see fee opportunities across advisory, debt and equity markets, not just IPO execution.

The question now is whether demand is deep enough to clear deals without heavy discounts. HSBC’s pipeline suggests issuers have not walked away, and regional officials and executives expect markets to recover. But the next signal will be whether fourth-quarter transactions price successfully, especially for delayed companies in technology, retail and consumer sectors.]]></content:encoded>
</item>
<item>
  <title>SK hynix Uses HPE Discover to Push AI Memory Beyond HBM</title>
  <link>https://stechtimes.com/en/article/sk-hynix-uses-hpe-discover-to-push-ai-memory-beyond-hbm-mqlkxv6o</link>
  <guid isPermaLink="true">https://stechtimes.com/en/article/sk-hynix-uses-hpe-discover-to-push-ai-memory-beyond-hbm-mqlkxv6o</guid>
  <pubDate>Fri, 19 Jun 2026 23:48:37 GMT</pubDate>
  <category>chips-semiconductors</category>
  <description><![CDATA[SK hynix used HPE Discover 2026 in Las Vegas to showcase HBM, CMM-DDR5, eSSD and server DRAM products for AI infrastructure buyers.
The company said HPE-certified products already deployed in HPE servers include PS1010 E3.S eSSDs based on 176-layer 4D NAND and 64GB DDR5 RDIMM modules built on 1c process technology.
The clearest commercial point is HPE certification and supply; the booth display does not by itself show broader customer adoption.]]></description>
  <enclosure url="https://stechtimes.com/_article-images/dd97f67eec1ba75a-sk-hynix-uses-hpe-discover-to-push-ai-memory-beyond-hb-1781912911277.jpg" type="image/jpeg" />
  <content:encoded><![CDATA[SK hynix used HPE Discover 2026 to present a broader AI memory portfolio for HPE server systems, placing CXL memory, enterprise SSDs and server DRAM alongside its better-known HBM line.

HPE Discover Becomes an AI Memory Showcase
HPE held HPE Discover Las Vegas 2026 from June 15 to 18 at The Venetian Convention and Expo Center in Las Vegas, with the program focused on AI, cloud and networking. SK hynix used the event to reinforce its partnership with HPE and to showcase products it said are already certified for HPE server systems.

The story extends beyond high-bandwidth memory. SK hynix organized its booth into four product zones: HBM, CMM-DDR5, eSSD and server DRAM. Training and inference systems need more than accelerator memory. They also require expandable system memory, fast storage and server modules that fit into data-center platforms.

HBM4 Sits Beside CXL Memory
The HBM area centered on a model of Nvidia’s Vera Rubin superchip, showing where HBM4 would sit in a next-generation AI computing design. SK hynix presented its HBM lineup across three reference points: 16-layer 48GB HBM4 as the highest-capacity product shown, 12-layer 36GB HBM4 as another next-generation option, and 12-layer 36GB HBM3E as the current comparison point.

The CXL section carried a different message. SK hynix showed a first-generation CMM-DDR5 product with 128GB and CXL 2.0+, then placed it next to a second-generation version with 256GB and CXL 3.2. The company also showed CMM-DDR5 inside Liqid’s CXL Pooled Memory Server, offering a practical interoperability example for buyers evaluating memory expansion beyond standard server slots.

HPE Certification Gives the Booth a Commercial Anchor
The most concrete commercial detail is HPE certification. SK hynix said it has completed certification and is supplying HPE with PS1010 E3.S data-center eSSDs based on 176-layer 4D NAND, along with 64GB DDR5 RDIMM server DRAM built on 10-nanometer-class sixth-generation 1c process technology.

The eSSD display covered several server-storage form factors. SK hynix listed U.3 products from the PS1010/1030 and PE1010/1030 lines, E3.S products from the same PS and PE families, and additional SE5110 SATA3, PE9010 M.2 2280 and PEB210 M.2 2280 drives. On the DRAM side, SK hynix showed 128GB, 96GB and 64GB DDR5 RDIMM modules, plus a 256GB 3DS RDIMM and 192GB SOCAMM2.

SK hynix also used a technical session to connect the hardware display to memory architecture. Byeonho Koo of DRAM Solution and Yunjay Hong of DRAM System Analysis presented CXL pooled memory, Processing-Near-Memory and the HMSDK heterogeneous memory management tool as ways to address rising memory-capacity and bandwidth pressure from AI inference workloads.

The remaining question for buyers is the breadth of deployment. HPE certification and named modules make the partnership more tangible than a generic booth claim, but SK hynix did not disclose additional customer names, shipment volumes or production-scale adoption for the full portfolio shown in Las Vegas.]]></content:encoded>
</item>
<item>
  <title>Buzz HPC’s £220 Million Deal Puts Canada’s Sovereign AI Push on Nvidia Racks</title>
  <link>https://stechtimes.com/en/article/buzz-hpcs-220-million-deal-puts-canadas-sovereign-ai-push-on-nvidia-rack-mql849qd</link>
  <guid isPermaLink="true">https://stechtimes.com/en/article/buzz-hpcs-220-million-deal-puts-canadas-sovereign-ai-push-on-nvidia-rack-mql849qd</guid>
  <pubDate>Fri, 19 Jun 2026 17:49:40 GMT</pubDate>
  <category>cloud-data-centers</category>
  <description><![CDATA[Buzz HPC signed a three-year sovereign AI contract with Bell Canada and Cohere, with 2,304 Nvidia Grace Blackwell GPUs planned for Bell Canada’s Merritt data center.]]></description>
  <enclosure url="https://stechtimes.com/_article-images/6c9c08a95afbb21d-buzz-hpc-s-220-million-deal-puts-canada-s-sovereign-ai-1781891376223.webp" type="image/jpeg" />
  <content:encoded><![CDATA[Canada’s AI Factory Deal Gets a Compute Layer
Hive Digital Technologies subsidiary Buzz HPC has signed a three-year sovereign AI contract with Bell Canada and Cohere, moving Canada’s AI infrastructure debate from policy language to a named data-center deployment. The deal is valued at £220 million and centers on GPUs that Buzz HPC will provide to Cohere inside Bell Canada’s facility in Merritt, British Columbia.

The customer use case is specific. The Merritt cluster is intended to give Cohere Canadian compute for foundation-model work and AI products aimed at public-sector and enterprise buyers. The announcement is more concrete than a general cloud-capacity pledge: it names the infrastructure operator, the telecom and data-center partner, the model company, the site, and the intended customer base.

The Rack Details Are Specific
Buzz HPC’s hardware order covers 2,304 Nvidia Grace Blackwell GPUs configured around GB200 NVL72 rack-scale systems. Nvidia Quantum InfiniBand networking links the systems, giving the project a clear high-performance computing profile rather than a vague “AI cloud” label.

The financing trail is also visible. The GPUs were bought using proceeds from Hive’s April 2026 $115m convertible note financing round. Aydin Kilic, Hive’s CEO, expects the cluster to go live in late 2026 or early 2027 and add approximately $70m of annualized revenue. Plans for the 6.5MW Merritt deployment were disclosed in March 2026.

Those figures give the project measurable operating markers. Buzz HPC has to turn purchased GPUs, Bell Canada data-center capacity, and Cohere demand into a running cluster on the disclosed timeline. The annualized revenue figure is meaningful only if the deployment enters service, maintains enough utilization, and supports Cohere workloads without slipping into another infrastructure announcement cycle.

Sovereignty Claim Still Has Operating Questions
Frank Holmes, executive chairman of Hive Digital Technologies, framed the partnership as a way to commercialize Canadian AI talent with domestic industrial infrastructure. Craig Tavares, Buzz HPC’s president and COO, tied the stack to Bell’s national platform, Cohere’s enterprise AI solutions, Hypertec’s Canadian-built GPU servers, and Nvidia’s full-stack AI infrastructure.

The operating case remains unfinished. Buzz HPC has not named Cohere workload volumes, government customers, corporate customers, utilization commitments, or service-level terms. Hive/Buzz is already working with Bell Canada on several deployments, including earlier Nvidia AI infrastructure work in Manitoba, and Hive is also planning a 320MW data center in the Greater Toronto area.

For now, the Merritt project gives Canada’s sovereign AI push a concrete cluster to judge. The key test is whether the late 2026 or early 2027 go-live happens, whether Cohere runs production workloads for government and corporate customers in Canada, and whether Hive can turn the cluster into the approximately $70m annualized revenue Kilic described.]]></content:encoded>
</item>
<item>
  <title>Arm and Supermicro Put Agentic AI Servers to a CPU Test</title>
  <link>https://stechtimes.com/en/article/arm-and-supermicro-put-agentic-ai-servers-on-a-cpu-test-mqkzkvbz</link>
  <guid isPermaLink="true">https://stechtimes.com/en/article/arm-and-supermicro-put-agentic-ai-servers-on-a-cpu-test-mqkzkvbz</guid>
  <pubDate>Fri, 19 Jun 2026 13:50:39 GMT</pubDate>
  <category>chips-semiconductors</category>
  <description><![CDATA[Supermicro has introduced new server platforms built around Arm’s AGI CPU for inference-heavy and agentic AI workloads across cloud, enterprise and edge deployments.
Arm says the AGI CPU includes up to 136 Arm Neoverse V3 cores, 12 DDR5 memory channels running at up to 8800 MT/s and PCIe Gen6 connectivity within a 300W power envelope.
The key test is whether operators can use these CPU-heavy designs to add inference capacity without creating new pressure on power and cooling.]]></description>
  <enclosure url="https://stechtimes.com/_article-images/1f6289712490dceb-arm-and-supermicro-put-agentic-ai-servers-on-a-cpu-tes-1781877033016.jpg" type="image/jpeg" />
  <content:encoded><![CDATA[Supermicro has introduced a new server portfolio built around Arm’s AGI CPU, giving AI infrastructure buyers another option for inference-heavy and agentic workloads that need more than GPU acceleration alone.

Supermicro Pitches A CPU-Heavy AI Rack Strategy
The announcement focuses on servers for cloud, enterprise and edge deployments. Arm describes agentic AI workloads as persistent systems that coordinate reasoning, retrieval, memory access, planning and communication across services and models. In that workflow, the CPU is not just a support chip beside the accelerator. It handles orchestration, I/O movement and general-purpose compute, roles that can become more prominent as inference spreads across more applications.

Arm introduced the AGI CPU in March 2026. Its published specification centers on a large general-purpose compute block: up to 136 Arm Neoverse V3 cores, 12 DDR5 memory channels, memory speeds of up to 8800 MT/s, PCIe Gen6 links and a 300W envelope. Arm also makes a direct rack-level comparison, estimating up to 2x higher performance per rack than comparable x86-based systems.

The portfolio is relevant to data-center operators facing a practical constraint: inference demand can grow even when facilities cannot keep adding power and cooling at the same pace. The announcement does not include customer deployments, benchmark logs or production volumes, so the performance claim remains an Arm estimate until buyers provide real installation data.

The operational question is workload fit. A CPU-dense rack can look attractive on paper, but agentic AI systems still need to move data between retrieval tools, models, storage and application services without creating new bottlenecks. Memory bandwidth, I/O capacity and software scheduling are as important as the headline core count.

The Rack Figures Are Specific
Supermicro’s liquid-cooled Open Rack Wide platform, the ARS-142TP-QNR-LCC, can support up to 336 AGI CPUs in a fully populated rack. A second liquid-cooled Open Rack V3 system, the 2U4N ORV3 ARS-242TP-QNR-LCC, supports up to 168 AGI CPUs per rack. Both systems are targeted for sampling in Q1 2027 and production availability in Q2 2027.

The company is also extending the design to air-cooled systems. The single-socket ARS-212HE-FNR short-depth server is aimed at edge deployments with tighter power and space limits, with sampling targeted for Q4 2026 and production in Q1 2027.

For more conventional data-center work, the dual-socket 2U ARS-222H-NR supports up to 8 NVMe drives and accelerator expansion in a standard 19-inch form factor. The 5U ARS-522GP-NR targets AI inference deployments with up to eight accelerator cards, dual AGI CPUs and high-density NVMe storage.

The Installation Burden Shifts To Power, Cooling And Workload Fit
The pitch is narrower than a broad AI boom story. Supermicro and Arm argue that agentic AI will need balanced systems, with CPUs, accelerators, memory bandwidth, I/O capacity and efficient rack design working together. That is a real operational question for enterprises that want inference closer to applications, databases or edge locations.

The next evidence should come from sampling, production availability and buyer deployment details. Operators will need to see whether these systems can deliver the promised density, manage heat in liquid-cooled and air-cooled environments, and improve inference throughput for real agentic workloads rather than only in supplier estimates.]]></content:encoded>
</item>
<item>
  <title>Festina Finance Secures Birchway Capital for a Pension-Core Upgrade Push</title>
  <link>https://stechtimes.com/en/article/festina-finance-gets-birchway-capital-for-a-pensioncore-upgrade-push-mqkxdj7z</link>
  <guid isPermaLink="true">https://stechtimes.com/en/article/festina-finance-gets-birchway-capital-for-a-pensioncore-upgrade-push-mqkxdj7z</guid>
  <pubDate>Fri, 19 Jun 2026 12:48:56 GMT</pubDate>
  <category>uae-economy</category>
  <description><![CDATA[Festina Finance has secured a growth investment of more than €25 million from Birchway Capital, valuing the Danish pension-technology company at about €200 million.
The company says its platforms support customers responsible for about 10 million pension policies and 3 million banking customers across Europe.
The funding tests a core operating question for legacy pension administration: whether cloud-native modular systems can replace ageing infrastructure without weakening resilience or control.]]></description>
  <enclosure url="https://stechtimes.com/_article-images/4111a5663fbe0836-festina-finance-gets-birchway-capital-for-a-pension-co-1781873331703.jpg" type="image/jpeg" />
  <content:encoded><![CDATA[Festina Finance has secured a growth investment of more than €25 million from Birchway Capital, giving the Danish pension-technology company fresh capital for European expansion and product development. The transaction values Festina Finance at about €200 million.

The Funding Targets Pension Infrastructure, Not a Consumer App
Festina Finance sells modular cloud software used by pension, life insurance and financial-planning providers. The company says its systems sit behind customer books that include about 10 million pension policies. They also touch about 3 million banking customers in Europe. Festina Finance employs more than 200 people across Denmark, the Netherlands, the UK and Norway.

The capital targets a narrow but difficult part of financial technology: replacing old administration systems that handle long-running pension and life insurance obligations. Festina Finance said the investment will support continued development of its life and pension platform, team expansion and international growth, with the UK named as a key strategic market.

This is not a payments-rail story or a retail banking launch. The problem is older back-office infrastructure that pension funds and insurers cannot casually switch off. Policy records, member data, fee rules and scheme changes must keep working while the underlying platform changes.

Netcompany Keeps a Direct Stake in the Stack
Netcompany will hold a 22% stake in Festina Finance after the investment. The companies plan to keep working together through AMPLIO Life & Pension and Netcompany Banking Services, where Festina Advisor is part of the collaboration.

The relationship runs deep in the product because pension administration is rarely a clean software swap. Providers must move rules, policies, member records and operating workflows from older systems into platforms that can be changed without losing control. Founder Morten Schantz said Festina Finance is focused on cloud-native modular systems that let providers adapt faster, improve operating efficiency and maintain resilience.

The Netcompany connection also gives the announcement a broader infrastructure angle. AMPLIO Life & Pension ties Festina's core pension capabilities to a larger implementation channel, while Netcompany Banking Services links the work to financial-infrastructure operations rather than a stand-alone software module.

Birchway Is Buying Into a Legacy-Systems Problem
Birchway Capital is a UK growth equity investor focused on enterprise technology, fintech and AI businesses. Founding Partner Kilian Pender said customers confirmed the strength of Festina Finance's technology and its role as providers move away from fragmented older infrastructure.

The commercial claim is specific enough to test. Festina Finance already points to large policy and banking-customer exposure in Europe, but the funding announcement does not name new pension-provider wins, migration deadlines or signed UK customers tied to this investment. Implementation evidence remains the best measure.

The UK Expansion Has to Prove Migration Discipline
The UK is the market Festina Finance chose to identify by name. Richard Davies, Country Managing Partner at Netcompany UK, said the transaction supports work to update core pension infrastructure through cloud systems, larger-scale deployment and data-led tools.

The next proof will come from deployments, not valuation. Pension providers will need evidence that Festina Finance can move complex administration workloads, support scheme changes and maintain control over member-critical operations while expanding beyond its current European footprint.]]></content:encoded>
</item>
<item>
  <title>Goldman’s AI Growth Bet Runs Into a Harder Global Economy</title>
  <link>https://stechtimes.com/en/article/goldmans-ai-growth-bet-runs-into-a-harder-global-economy-mqkv5icw</link>
  <guid isPermaLink="true">https://stechtimes.com/en/article/goldmans-ai-growth-bet-runs-into-a-harder-global-economy-mqkv5icw</guid>
  <pubDate>Fri, 19 Jun 2026 11:46:44 GMT</pubDate>
  <category>ai</category>
  <description><![CDATA[Goldman Sachs chief economist Jan Hatzius said AI infrastructure spending could lift productivity growth, while economists warned that war, debt and trade fragmentation are weighing on the global outlook.
The World Bank cut its 2026 global growth forecast to 2.5 per cent and lowered its projection for the Middle East, North Africa, Afghanistan and Pakistan to 1.6 per cent.
AI may improve long-run productivity, but energy shocks, export cuts and debt concerns are immediate constraints for emerging markets and Gulf exporters.]]></description>
  <enclosure url="https://stechtimes.com/_article-images/d81c3dfbdd6cfe00-goldman-s-ai-growth-bet-runs-into-a-harder-global-econ-1781869598206.jpg" type="image/jpeg" />
  <content:encoded><![CDATA[Goldman Sachs is making an optimistic case for AI at a difficult moment for the global economy. Chief economist Jan Hatzius said investment in AI infrastructure is already supporting activity and could lift productivity over the coming decade, while other economists pointed to war, debt and trade disruption as near-term brakes.

AI Optimism Meets a Lower Growth Forecast
The World Bank cut its 2026 global growth forecast to 2.5 per cent from an estimated 2.9 per cent last year. It also lowered its 2026 outlook for the Middle East, North Africa, Afghanistan and Pakistan to 1.6 per cent from 1.8 per cent.

The Goldman view is more than a simple technology call. Hatzius told a Council on Foreign Relations event in New York that most topics around the global economy are negative, but he still sees “one big positive” in faster underlying productivity growth.

His AI argument rests on infrastructure spending and productivity, not hype around consumer apps. Hatzius estimated that AI could provide about one and a half percentage points of annual productivity-growth support over a 10-year transition period. Goldman Sachs has raised its long-term GDP growth forecast to 2.5 per cent from 1.75 per cent before the pandemic.

Debt and Energy Are the Immediate Friction
Hatzius also flagged a different risk: the assumption that debt can keep rising without political resistance. That warning sits alongside the World Bank downgrade, where wars, fragmented trade and weaker investment are already reducing growth expectations.

Natasha Sarin, founder of Yale University's Budget Lab, said emerging markets are taking the hardest economic hit from the Iran war. She pointed to disrupted energy markets, higher oil prices and delays in refined-product shipments, with recovery for infrastructure and supply chains measured in months or years rather than days.

The Gulf angle is direct. Opec said it still expects world oil demand to reach 113.3 million barrels per day in 2030, and Gulf exporters have had to make significant export cuts this year because of the Strait of Hormuz closure and attacks on key energy sites.

Productivity Gains Still Need a Stable Operating Base
AI infrastructure spending can support economic activity before productivity gains are fully visible. Data centres, chips, cloud capacity and enterprise deployments all require capital, power and long planning cycles. The investment-and-productivity link is clear, but it does not show how quickly AI gains will reach weaker emerging-market economies.

Doug Rediker of International Capital Strategies argued that the Iran war's consequences may become more strategic than economic. IMF managing director Kristalina Georgieva wrote that inflation expectations, financial conditions and commodity prices have been affected, while also warning that intensified conflict or disruption would create a clear risk to global growth.

Measuring More Than AI Promises
The next test will not be another broad AI claim. It will be whether companies can turn infrastructure spending into measurable productivity while governments manage debt, energy costs and trade fragmentation. For Gulf exporters and emerging markets, the harder question is whether AI-led efficiency gains will arrive fast enough to offset the current pressure from shipping delays, oil-market disruption and weaker investment.]]></content:encoded>
</item>
<item>
  <title>Alterra Takes UAE Climate Capital Into Peru’s Power Grid</title>
  <link>https://stechtimes.com/en/article/alterra-takes-uae-climate-capital-into-perus-power-grid-mqkt1q63</link>
  <guid isPermaLink="true">https://stechtimes.com/en/article/alterra-takes-uae-climate-capital-into-perus-power-grid-mqkt1q63</guid>
  <pubDate>Fri, 19 Jun 2026 10:47:47 GMT</pubDate>
  <category>uae-economy</category>
  <description><![CDATA[Alterra, the UAE climate fund, has made its first direct Latin American renewable energy investment through a co-investment in Peru’s Inkia Energy with I Squared Capital.
The deal uses Alterra’s $1.2 billion Opportunity Fund and targets an operator with 2.6GW of generation capacity and a renewables pipeline of about 4GW.
The investment gives the UAE another overseas climate-finance test: whether long-term capital can support power-sector growth in a market where mining, infrastructure and industrial demand are rising.]]></description>
  <enclosure url="https://stechtimes.com/_article-images/7366e1129282a624-alterra-takes-uae-climate-capital-into-peru-s-power-gr-1781866062517.jpg" type="image/jpeg" />
  <content:encoded><![CDATA[Alterra is taking UAE climate capital into Peru’s power sector through a co-investment in Inkia Energy with I Squared Capital, its first direct investment in Latin America’s renewable energy market.

UAE Climate Finance Moves Into Latin America
Alterra is the UAE’s $30 billion climate fund. The Inkia transaction is being made through its $1.2 billion Opportunity Fund, a co-investment vehicle that includes a $250 million contribution from BBVA.

The companies did not disclose the size of the Inkia investment. The transaction is being presented less as a single-asset purchase and more as a test of how UAE-backed transition capital can enter high-growth power markets outside the Gulf.

Inkia is based in Lima and operates 2.6GW of generation capacity, equal to about a quarter of Peru’s electricity supply. Its renewables pipeline is about 4GW, including an expansion of 1GW in solar and wind projects tied to Peru’s decarbonisation plans.

The Peru Bet Is About Power Demand
Peru gives Alterra a different kind of transition exposure from Gulf infrastructure or mature-market renewable platforms. Infrastructure development, mining activity and industrial growth are driving electricity demand, making reliable generation capacity part of the economic story as much as the climate story.

Alterra chief executive Majid Al Suwaidi linked the investment to long-term capital for growth markets and to gigawatt-scale energy platforms. Inkia already has generation capacity in place, while its pipeline gives investors a path to expand renewable supply if projects move from planning into delivery.

The deal also extends Alterra’s post-Cop28 investment mandate. The fund was launched during the 2023 Cop28 climate conference in Dubai and has said it aims to raise $250 billion by 2030 for projects that support the global energy transition.

The Numbers Put Inkia In A Wider Renewable Build-Out
The timing lines up with a broader renewable capacity surge. In 2025, renewable generation worldwide was about 8.5 per cent higher than a year earlier. New renewable capacity reached 800GW, and solar made up 75 per cent of that build-out while meeting more than 25 per cent of demand.

The market backdrop is large enough to attract sovereign and institutional capital. Global renewable energy is estimated at nearly $1.14 trillion in 2026; the cited projection reaches about $1.84 trillion by 2034, with growth of almost 6.2 per cent a year. That scale explains why investors are chasing power platforms that can expand, not isolated demonstration projects.

The Real Test Is Execution In Peru
For the UAE, the investment adds another outward-facing climate-finance deal to its economic diplomacy. For Inkia, the question is whether new capital can turn a large pipeline into operating power without losing the reliability needed by Peru’s mines, industrial users and infrastructure projects.

I Squared Capital chairman and managing partner Sadek Wahba framed Inkia’s next phase around a gigawatt-scale renewables programme. What matters now is concrete: disclosed investment size, project-level approvals, solar and wind expansion timelines, and whether Inkia’s additional capacity actually keeps pace with Peru’s electricity demand.]]></content:encoded>
</item>
<item>
  <title>Block’s Builderbot Shows Where AI Coding Tools Hit The Enterprise Wall</title>
  <link>https://stechtimes.com/en/article/blocks-builderbot-shows-where-ai-coding-tools-hit-the-enterprise-wall-mqkqyp2o</link>
  <guid isPermaLink="true">https://stechtimes.com/en/article/blocks-builderbot-shows-where-ai-coding-tools-hit-the-enterprise-wall-mqkqyp2o</guid>
  <pubDate>Fri, 19 Jun 2026 09:49:27 GMT</pubDate>
  <category>ai</category>
  <description><![CDATA[Block says its Builderbot framework coordinates AI agents across internal repositories, Slack threads, issue trackers and continuous-integration workflows.
The company says the system runs over 200,000 commands each day, merges about 1,500 pull requests each week and accounts for roughly fifteen percent of company code changes.
The stronger claim is not code generation alone. Block is testing whether agentic software work can handle permissions, context, CI failures and customer-data isolation inside a large engineering organisation.]]></description>
  <enclosure url="https://stechtimes.com/_article-images/5a7ac0d95390f826-block-s-builderbot-shows-where-ai-coding-tools-hit-the-1781862561822.jpg" type="image/jpeg" />
  <content:encoded><![CDATA[Block has deployed Builderbot, an internal agent framework built to move AI coding assistance beyond single-repository suggestions and into cross-service software work.

Slack Becomes The Control Surface
Builderbot starts inside Slack. An engineer tags the @builderbot account, writes a short description of the task and keeps the work inside the same thread while the system researches, plans and acts.

That design is important because Block is not describing a code-completion tool. The company says Builderbot coordinates multiple agents across internal services, project-management systems and continuous-integration workflows. Several team members can watch the same thread and steer the work while the agent handles the mechanical parts of the task.

Block built the system after standard coding assistants ran into the limits of a large corporate software estate. The company describes hundreds of active services and hundreds of millions of lines of proprietary code, where a change in one product area can depend on repositories and APIs owned by another team.

The Scale Claim Is Operational
Builderbot has permission and context to work across company-managed repositories. Block says a Cash App engineer can use it to trigger changes in a Square backend service without prior knowledge of that subsystem, because the orchestration layer supplies the architectural context.

The workflow also connects to Linear and Jira. Builderbot can retrieve assigned tickets, create the initial branch, generate code and open pull requests. It then watches the continuous-integration suite, responds to automated test failures or human feedback and iterates until the change meets production standards.

Block attaches concrete volume to the deployment. The company says Builderbot executes over 200,000 operational commands each day and merges about 1,500 pull requests each week. It also says autonomous code contributions represent roughly fifteen percent of all structural modifications across the company network.

The numbers give the story weight than a laboratory demo. They still do not prove that every merged change is complex or strategically valuable. They show that Block has moved agentic coding into a live engineering pipeline where throughput, test repair, review discipline and repository permissions all matter.

Customer Data Is Outside The Agent Boundary
The security boundary is central to the deployment. Block says Builderbot operates within source-code repositories and system-configuration domains, while the architecture prevents the agent from reading, processing or transmitting raw customer data.

The company says Builderbot has no technical access to live payment information or personally identifiable information inside production servers. That isolation is the difference between a useful engineering agent and a compliance problem for a company that runs payments and financial-app infrastructure.

Brad Axen, Block's head of AI capabilities, framed Builderbot as a missing layer between AI coding tools and large-scale engineering practice. He said the system handles orchestration, context and environment so engineers can focus on higher-value decisions. He also said Square teams used it to move seller-requested features from a backlog into production in days instead of months, with engineers still shaping product decisions.

Goose And MCP Are The Public Pieces
Builderbot sits on top of Block's open-source goose agentic framework. Block first developed goose internally, then contributed the code base to the Agentic AI Foundation.

The company also links Builderbot's development to Model Context Protocol. Block says internal integration challenges around goose led to collaboration with Anthropic on MCP, which is used to connect autonomous agents to development tools and data sources.

What matters now is not whether Builderbot can produce more pull requests. Block has already disclosed volume. The harder enterprise test is whether agentic engineering keeps code quality, repository permissions, CI discipline and customer-data isolation intact as more teams hand routine software work to agents.]]></content:encoded>
</item>
<item>
  <title>Dream Raises $260 Million For The Hard Sell In Sovereign AI</title>
  <link>https://stechtimes.com/en/article/dream-raises-260-million-for-the-hard-sell-in-sovereign-ai-mqkoz93i</link>
  <guid isPermaLink="true">https://stechtimes.com/en/article/dream-raises-260-million-for-the-hard-sell-in-sovereign-ai-mqkoz93i</guid>
  <pubDate>Fri, 19 Jun 2026 08:53:55 GMT</pubDate>
  <category>ai</category>
  <description><![CDATA[Dream has raised $260 million at a $3 billion valuation to expand sovereign AI and national cyber defence platforms across several regions. For Gulf buyers, the missing proof is named deployments, hosting arrangements and procurement scope.]]></description>
  <enclosure url="https://stechtimes.com/_article-images/b58475baa4524be1-dream-raises-260-million-for-the-hard-sell-in-sovereig-1781859228626.jpg" type="image/jpeg" />
  <content:encoded><![CDATA[Dream raises money for national AI systems
Dream has raised $260 million in new funding at a $3 billion valuation, adding capital for an expansion plan built around sovereign AI and national cyber defence. The company says the round lifts total funding to $412 million.

The financing was co-led by Bicycle Capital and Group 11. Antler, Bain Capital Ventures, Tru Arrow Partners and other investors also participated. Dream is based in Tel Aviv and said the new money will support work across Europe, the Middle East, Asia and the Americas, although it did not name target countries.

Sovereign AI is the pitch to governments
Dream's argument is aimed at governments and sensitive public-sector systems. Its chief executive Shalev Hulio said nations need control over data, infrastructure and strategic technologies rather than depending on systems they do not govern.

That message fits the Gulf policy environment, where sovereign AI has become tied to cloud location, public services, defence applications, health care, financial systems and government data. Dream is not selling a consumer AI tool; it is positioning AI as national infrastructure.

The round buys Dream attention. It still has to win deployments
The $260 million round gives Dream more capital than many security-focused AI start-ups, and the $3 billion valuation signals investor appetite for companies that package AI with cyber defence and sovereignty. The round is also the company's fifth financing, according to Crunchbase data material.

The harder test is deployment evidence. Dream has named broad regions for expansion, but it has not disclosed which Middle East governments, agencies or enterprises it expects to serve next. The source material also does not give contract values, customer counts or production benchmarks.

What Gulf buyers need to see
Sebastian Kurz, Dream's president and co-founder, framed sovereign AI as a question of whether governments will own the systems they use. That is a harder claim than ordinary AI adoption language because it points to operational control: where data sits, who governs infrastructure and how national-interest decisions are made.

For Gulf governments and regulated sectors, the next evidence should be specific deployments. Named customers, hosting arrangements, security controls and procurement scope would show whether Dream's funding round becomes a regional sovereign AI business or remains a capital raise with a broad geographic map.]]></content:encoded>
</item>
<item>
  <title>NymCard Pitches One Stack For MENA Banks Still Stuck With Patchwork Payment Systems</title>
  <link>https://stechtimes.com/en/article/nymcard-pitches-one-stack-for-mena-banks-still-stuck-with-patchwork-paym-mqkmmptc</link>
  <guid isPermaLink="true">https://stechtimes.com/en/article/nymcard-pitches-one-stack-for-mena-banks-still-stuck-with-patchwork-paym-mqkmmptc</guid>
  <pubDate>Fri, 19 Jun 2026 07:48:09 GMT</pubDate>
  <category>chips-semiconductors</category>
  <description><![CDATA[Dubai-based NymCard launched nCore FullStack, a platform that puts card issuing, lending, money movement, settlement, financial-crime controls and reconciliation behind one integration.
The company says the system can run on public cloud, hybrid, on-soil or on-premise deployments, a key point for banks working under strict regional data-residency rules.
NymCard says it powers programmes for more than 60 banks, fintechs and enterprises across eight markets, but the launch still has to prove that banks will replace fragmented vendor stacks rather than add another layer.]]></description>
  <enclosure url="https://stechtimes.com/_article-images/a7332503575b57dd-nymcard-pitches-one-stack-for-mena-banks-still-stuck-w-1781855284874.webp" type="image/jpeg" />
  <content:encoded><![CDATA[Dubai-based NymCard has launched nCore FullStack, a payments-infrastructure platform aimed at banks that want to add financial products without stitching together a new vendor system each time.

One Integration Across Cards, Lending And Settlement
NymCard says nCore FullStack brings issuing processing, lending, money movement, settlement, financial-crime controls and reconciliation into a single in-house platform. The company is pitching the product at banks that still run older card-processing systems and then add separate tools for payments, fraud, compliance and back-office matching.

The operating pitch is narrower than a broad digital-bank rebrand. A bank connects to nCore once, then activates capabilities such as card issuing, lending or cross-border payments on the same integration. NymCard says the bank's core banking system remains the system of record while nCore connects directly into it.

The product list is broad. The issuing layer covers prepaid, debit, credit, wallet, virtual and tokenized programmes. The lending layer includes digital onboarding, credit decisioning, loan origination and servicing. Money movement covers domestic payments, cross-border payments, foreign exchange, remittance and open finance. Settlement includes fiat, stablecoin and real-time rails, while the financial-crime layer covers card fraud prevention, anti-money-laundering controls, sanctions screening, identity verification and authentication.

Data Residency Is Part Of The Sales Pitch
For MENA banks, deployment model is not a side detail. NymCard says nCore can run on public cloud, hybrid, on-soil or on-premise infrastructure, depending on what local regulators require for banking data.

For regional banks, that claim matters more. A modern payment platform that works only in a public-cloud setup can be hard to use where regulators require local hosting, stricter sovereignty controls or direct links into existing core banking systems. NymCard says the integration remains the same whether a bank chooses on-premise sovereignty or a public-cloud setup.

CEO and founder Omar Onsi framed the product as an answer to banks that keep adding connections on top of ageing infrastructure. Chief product officer Mario Wehbe said a bank could launch cards first and turn on lending or cross-border payments later without rewriting its core systems. Chief technology officer Srikanth Achanta said the technical challenge was to keep deployment flexible while keeping the capability set inside one platform.

Migration Is The Hard Part To Prove
NymCard also says banks can move from a legacy processor to nCore without disrupting live programmes. The company says it uses an agentic AI process and a purpose-built AI engine to move existing card programmes and data onto the platform, and that it has already done this for banks running in production.

That claim is important because payment infrastructure decisions are rarely won by feature lists alone. Banks need to know whether migrations preserve live card programmes, reconciliation records, fraud controls, sanctions workflows and customer-service continuity. NymCard says the migration process exists; it does not name the production banks or disclose migration volumes, error rates or implementation timelines.

NymCard has some operating scale behind the launch. The company says it powers payment programmes for more than 60 banks, fintechs and enterprises across eight markets, processes billions of transactions a year and has raised more than $70 million in funding to date. The cost pressure is also rising. NymCard cites a forecast that global banks will spend $57 billion on legacy payment-technology maintenance in 2028; the comparable 2022 figure was $36.7 billion.

The Hard Proof Is Bank Adoption, Not Feature Count
The market case is clear enough. NymCard is trying to replace a fragmented bank-payments stack with one configurable platform that can handle cards, lending, settlement, financial crime and reconciliation under regional deployment rules.

The commercial proof will come from named bank migrations, live product launches on multiple modules and evidence that customers can retire old processors rather than simply connect nCore beside them. Until those details are public, the launch is best read as a payments-infrastructure bid for banks that want modernization but cannot ignore data residency, core-system integration and operational risk.]]></content:encoded>
</item>
<item>
  <title>Claros Turns Samsung Foundry Into Its AI Power-Chip Test</title>
  <link>https://stechtimes.com/en/article/claros-turns-samsung-foundry-into-its-ai-powerchip-test-mqkko42w</link>
  <guid isPermaLink="true">https://stechtimes.com/en/article/claros-turns-samsung-foundry-into-its-ai-powerchip-test-mqkko42w</guid>
  <pubDate>Fri, 19 Jun 2026 06:53:17 GMT</pubDate>
  <category>chips-semiconductors</category>
  <description><![CDATA[Claros says Samsung Electronics will manufacture its integrated voltage regulator at the Austin, Texas fab, giving the startup a U.S. production route for chips meant to reduce power loss near AI processors and support 800 VDC data-center designs.]]></description>
  <enclosure url="https://stechtimes.com/_article-images/6bef5f065bc763f8-claros-turns-samsung-foundry-into-its-ai-power-chip-te-1781851990375.jpg" type="image/jpeg" />
  <content:encoded><![CDATA[Samsung gives Claros a manufacturing path
Claros says it is working with Samsung Electronics to scale production of its integrated voltage regulator, a power chip aimed at AI data centers. The company says the IVR will be manufactured at Samsung’s fab in Austin, Texas, giving Claros a U.S. foundry path as it tries to move from design into customer delivery.

The product target is specific: power conversion close to the processor. Claros argues that older data-center power architecture wastes electricity before it reaches compute silicon, while its IVR moves the conversion point much nearer to the chip. The company describes that shift as cutting the distance electricity travels from inches to millimeters.

The 800 VDC claim depends on the last step near the chip
Claros ties the Samsung collaboration to 800 VDC data-center architecture. Its argument is that higher-voltage distribution can improve rack-level efficiency, but the gain is incomplete if voltage regulation still happens too far from the point of demand. In Claros’s framing, the IVR is meant to finish that power chain by converting electricity millimeters from the processor.

The strongest number in the disclosure is the company’s claim that Claros IVRs can cut energy loss by up to 30%. That is a product claim, not a disclosed customer result. The source does not name a buyer, installation site, qualification milestone, shipment volume or measured deployment outcome. For now, the disclosed progress is manufacturing access and an expected sample path, not broad data-center adoption.

U.S. foundry capacity is the hard constraint
The manufacturing detail matters because Claros is not only selling a circuit idea. It needs production slots in a semiconductor market where foundry capacity is scarce and lead times are long. The company says IVR samples from Samsung fabs will start shipping this year, with full manufacturing in sight.

That does not settle the commercial test. Samples still have to reach prospective customers, survive qualification, fit server and accelerator designs, and prove that energy savings near the processor can justify the integration work. The source supports a manufacturing milestone; it does not prove customer demand, cost advantage, thermal performance under deployment load or data-center-wide energy reduction.

Claros also links the power argument to heat and cooling. Its stated chain is simple: less waste at the chip should mean less heat, less cooling, lower draw and less strain on the grid. Those are plausible operating targets for dense AI rooms, but the disclosure stops short of measured customer evidence. The Samsung arrangement gets Claros closer to hardware that can be tested in the field; it does not by itself validate the full data-center energy story.

The proof moves from fab access to qualification
Claros is trying to solve a real AI-infrastructure bottleneck: more compute means more power conversion, more heat and more pressure on data-center electrical design. A chip-level IVR can be attractive if it reduces loss before electricity reaches processors, but the operating evidence will come after samples leave the fab.

The useful evidence is concrete: whether Samsung-built IVR samples reach customers this year, whether those customers qualify the parts for AI systems, and whether Claros can show measured power-loss reduction without creating new integration or reliability problems inside dense data-center hardware.]]></content:encoded>
</item>
<item>
  <title>FedNow May Open A Faster Dollar Leg For Cross-Border Payments</title>
  <link>https://stechtimes.com/en/article/fednows-crossborder-plan-puts-instant-payments-up-against-compliance-mqkidaxa</link>
  <guid isPermaLink="true">https://stechtimes.com/en/article/fednows-crossborder-plan-puts-instant-payments-up-against-compliance-mqkidaxa</guid>
  <pubDate>Fri, 19 Jun 2026 05:48:52 GMT</pubDate>
  <category>fintech-digital-payments</category>
  <description><![CDATA[The Federal Reserve wants to let FedNow participants use non-Federal Reserve intermediaries for cross-border transfers. Fintech firms see a faster U.S. dollar leg; bank groups want sanctions screening, message codes and a staged rollout before instant payments move deeper into cross-border flows.]]></description>
  <enclosure url="https://stechtimes.com/_article-images/ef0cb8caeec04fff-fednow-s-cross-border-plan-puts-instant-payments-up-ag-1781848127044.webp" type="image/jpeg" />
  <content:encoded><![CDATA[FedNow Could Handle The U.S. Leg Of Cross-Border Transfers
The Federal Reserve is considering a Regulation J change that would let FedNow participants use intermediaries other than Federal Reserve Banks when sending funds through the instant-payment service. In plain terms, a U.S. bank could use FedNow for the domestic settlement leg while a correspondent bank or another intermediary handles the international part of the transfer.

That sounds technical because it is. The policy question is more familiar: fintech firms want the dollar leg of cross-border payments to move at instant-payment speed, while banks do not want that speed to outrun sanctions checks and risk controls.

Since July 2023, FedNow has operated as a 24/7 year-round rail for instant payments. Its current rules block non-Federal Reserve intermediaries, which keeps the network effectively domestic when a payment needs another bank in the chain. The proposed change would make FedNow look a little more like Fedwire's intermediary model, but without turning it into a global payment network by itself.

The demand signal is speed. The cited research found that 43% of U.S. small and medium-sized businesses that source internationally named faster settlement as their top cross-border payments priority. It also found that 27% are open to switching providers. That is why Stripe, Wise and other payment companies care about a rule change most customers will never read.

Fintech Support Comes With A Residency Problem
Stripe backed the proposal and said it closes a gap in U.S. payment infrastructure. Its argument is straightforward: businesses moving dollars across borders do not have continuous access to real-time USD settlement because Fedwire and same-day ACH are not always available, while FedNow remains largely domestic.

Stripe also pointed to a remaining constraint. FedNow's on-behalf-of residency restrictions still require the ultimate end customer in some flows to be a U.S. resident or U.S.-domiciled entity. If that rule stays in place, the cross-border benefit could be narrower than fintech providers want.

Wise also supported the proposal and described FedNow's domestic-only design as an artificial disparity with Fedwire. For firms that already connect to payment systems in multiple countries, the appeal is not a new consumer app. It is the chance to join real-time dollar settlement inside the United States to international payment workflows that already exist elsewhere.

Banks Want Guardrails Before Instant Cross-Border Scale
Bank groups broadly supported the goal, but their comments focused on operational risk. The Clearing House and the Bank Policy Institute warned that FedNow participants would likely need real-time sanctions screening for cross-border transactions because instant payments settle immediately and are generally irrevocable after processing.

The associations asked for a specific message code to identify cross-border transactions and a two-phase implementation strategy. In their preferred model, institutions would first opt into cross-border FedNow activity during an initial pilot period. Wider obligations would come later, depending on industry readiness and pilot results.

BNY called for clear guardrails and defined responsibilities for intermediaries and risk allocation. That is where the real argument sits. A faster message is useful only if banks can still satisfy sanctions, know-your-customer, data-sharing and local payment-system rules across jurisdictions.

The proposal would remove one U.S. connectivity constraint. It would not make foreign regulators, central banks or payment-system operators move at the same speed. If the Fed gets the balance right, FedNow could become a useful real-time dollar leg for international payment providers. If the compliance model is too vague, banks may support the idea in principle while keeping actual usage cautious.]]></content:encoded>
</item>
<item>
  <title>DataVolt’s Uzbekistan Data Center Tests Saudi Expansion Beyond The Gulf</title>
  <link>https://stechtimes.com/en/article/datavolts-uzbekistan-data-center-tests-saudi-expansion-beyond-the-gulf-mqkg57m1</link>
  <guid isPermaLink="true">https://stechtimes.com/en/article/datavolts-uzbekistan-data-center-tests-saudi-expansion-beyond-the-gulf-mqkg57m1</guid>
  <pubDate>Fri, 19 Jun 2026 04:46:35 GMT</pubDate>
  <category>cloud-data-centers</category>
  <description><![CDATA[Saudi Arabia-based DataVolt is developing the 12MW TAS-1 data center in Tashkent after securing $150 million in project financing, with larger Uzbekistan capacity still tied to preliminary agreements and future investment plans.]]></description>
  <enclosure url="https://stechtimes.com/_article-images/36ad84150d64cce0-datavolt-s-uzbekistan-data-center-tests-saudi-expansio-1781844390224.jpg" type="image/jpeg" />
  <content:encoded><![CDATA[TAS-1 Gives DataVolt A Central Asia Buildout
DataVolt is pushing its Saudi data center business into Uzbekistan with a 12MW facility in Tashkent. The company has secured $150 million in project financing for TAS-1. One funder is the German Investment and Development Company. European Bank for Reconstruction and Development also appears in the funder list, alongside Proparco and the OPEC Fund for International Development.

The project is useful because it connects a Saudi infrastructure developer to a Central Asian cloud and data-center market rather than another Gulf buildout. It is not yet a hyperscale proof point. TAS-1 is a first facility, and the disclosed capacity is modest compared with the larger expansion framework around it.

One timing question remains open: the investment receipt date is unclear, while the Uzbek Ministry of Digital Technologies had already described the project in May 2024 as fully financed through 100 percent foreign direct investment worth $150m. DataVolt has said TAS-1 is expected to come online in late 2026.

The Larger Plan Is Still Mostly Framework
Uzbekistan’s ministry has described TAS-1 as the first phase of a broader data center programme. The larger plan includes two additional 250MW data centers on two 25-hectare land parcels, with a foreign partner expected to invest $5 billion by 2030.

Those numbers are large, but they are not the same as operating capacity. The firmer near-term asset is the 12MW Tashkent project. The bigger 250MW sites and the $5 billion plan remain dependent on future execution, partner commitments, power availability and local demand for compute capacity.

DataVolt has also signed two memoranda of understanding with Veon-linked Beeline Uzbekistan. One could make Beeline Uzbekistan a major tenant of TAS-1. The other sets up evaluation work for the joint development and operation of a potential data center in Bukhara.

Saudi Capital Meets Local Execution Risk
DataVolt was founded in 2023 and is controlled by Vision Invest, a Saudi Arabian group focused on critical infrastructure. Its chief executive, Nanda, was previously CFO of ACWA Power, and much of the company’s board and executive team came from ACWA.

That background gives DataVolt a power-and-infrastructure profile, which matters for data centers because grid access, cooling, financing and land are often as important as server procurement. In Uzbekistan, the next evidence will be practical: TAS-1 completion, a confirmed tenant structure, and whether the Bukhara memorandum turns into a binding development plan.

Veon adds a regional operating link. The company is headquartered in Dubai and operates in Bangladesh, Kazakhstan, Pakistan, Ukraine and Uzbekistan. Its Kazakhstan unit, Beeline Kazakhstan, is also building a 2MW data center in Almaty that is set to go live by the end of this year, giving the group a wider Central Asia data-center thread.]]></content:encoded>
</item>
<item>
  <title>HKEX And HKMA Test e-HKD For After-Hours Derivatives Margin</title>
  <link>https://stechtimes.com/en/article/hkex-and-hkma-put-ehkd-into-a-live-marginpayment-test-mqkc26c8</link>
  <guid isPermaLink="true">https://stechtimes.com/en/article/hkex-and-hkma-put-ehkd-into-a-live-marginpayment-test-mqkc26c8</guid>
  <pubDate>Fri, 19 Jun 2026 02:52:15 GMT</pubDate>
  <category>fintech-digital-payments</category>
  <description><![CDATA[HKEX and the Hong Kong Monetary Authority are testing e-HKD for after-hours derivatives margin payments, using a wholesale CBDC payment leg to address a specific 3:00 p.m. funding deadline in HKCC clearing.]]></description>
  <enclosure url="https://stechtimes.com/_article-images/318dcb8a22f6ba9f-hkex-and-hkma-put-e-hkd-into-a-live-margin-payment-tes-1781837529941.jpg" type="image/jpeg" />
  <content:encoded><![CDATA[A CBDC Test Inside Derivatives Clearing
HKEX and the Hong Kong Monetary Authority are launching a joint pilot for digital margin payments in Hong Kong's after-hours derivatives market. The project uses e-HKD as a wholesale central bank digital currency for advance margin payments during the After-Hours Trading session.

This is not a broad consumer wallet trial. It is a narrow clearing test aimed at one awkward part of market plumbing: getting margin money where it needs to be after the normal banking day has effectively closed. HKFE Clearing Corporation Limited Clearing Participants can join Real-Value Trial Transactions on an optional basis, while any wider rollout still depends on regulatory approval, market readiness and other considerations.

That narrowness is the point. A retail CBDC pilot can be judged by wallets, merchants and consumers. This pilot will be judged by whether clearing participants can move money at the right time without creating new operational risk.

Why After-Hours Margin Is The Hard Part
The existing process gives Clearing Participants a deadline problem. Advance margin deposit requests must reach HKCC by 3:00 p.m. if those funds are to count for the following After-Hours Trading session.

e-HKD is being tested as a 24/7 wholesale settlement tool that could make margin funding more flexible outside regular banking hours. HKEX and HKMA say the design is meant to preserve existing operational workflows while improving risk management in derivatives clearing.

That is the hard part. The payment leg has to work, but it also has to fit how participants already submit margin requests, manage liquidity and handle obligations after the banking day has closed. A faster token does not help much if it adds reconciliation, compliance or workflow friction somewhere else.

Hong Kong's derivatives market gives the pilot a stronger reason to exist. Average daily volume reached a record 1.66 million contracts in 2025 and exceeded 1.78 million contracts in the first five months of 2026. More volume means more pressure on the systems that handle margin and risk outside regular hours.

The Pilot Still Has Approval Risk
HKEX Chief Operating Officer Vanessa Lau framed the project as a way to address operational pain points outside regular business hours and strengthen capital-market infrastructure. HKMA Deputy Chief Executive Howard Lee described it as a wholesale CBDC application in a live market environment.

Those comments matter because Hong Kong is not testing e-HKD only as a policy symbol. It is putting the instrument into a specific market operation where failure would be visible to clearing participants, not just to a sandbox team.

The limits are just as important. HKEX is inviting HKCC Clearing Participants to join on an optional basis, not announcing mandatory participation across the market. The same announcement ties wider adoption to regulatory approval, market readiness and other considerations.

The next evidence should be concrete: how many Clearing Participants join, whether after-hours margin payments settle without disturbing existing workflows, and whether regulators allow the trial to move beyond optional participation. If that happens, Hong Kong's CBDC strategy gains a more convincing use case than another retail-payment demo. If it does not, e-HKD remains useful on paper but unproven in the kind of market plumbing where digital settlement is supposed to matter.]]></content:encoded>
</item>
<item>
  <title>Google’s Brazos Sidecar Brings Liquid Cooling To One AI Rack At A Time</title>
  <link>https://stechtimes.com/en/article/googles-brazos-cooling-sidecar-targets-the-ai-retrofit-problem-mqk9rmoh</link>
  <guid isPermaLink="true">https://stechtimes.com/en/article/googles-brazos-cooling-sidecar-targets-the-ai-retrofit-problem-mqk9rmoh</guid>
  <pubDate>Fri, 19 Jun 2026 01:48:04 GMT</pubDate>
  <category>cloud-data-centers</category>
  <description><![CDATA[Google has introduced Brazos, an open-source liquid-to-air cooling sidecar for existing air-cooled data centers. The rack-level design supports OCP ORv3 racks, a 60kW nominal thermal load and a retrofit path for operators that need AI cooling capacity without rebuilding whole halls.]]></description>
  <enclosure url="https://stechtimes.com/_article-images/563dd1aa7f3389ca-google-s-brazos-cooling-sidecar-targets-the-ai-retrofi-1781833678926.webp" type="image/jpeg" />
  <content:encoded><![CDATA[Brazos Is Built For Air-Cooled Data Centers
Google has introduced Brazos, an open-source rack-mounted liquid-to-air cooling system for existing air-cooled data centers. The target is a real data-center headache: AI racks are getting hotter faster than many facilities can rebuild their mechanical rooms.

Brazos is a closed-loop sidecar. It captures heat through liquid at the component level, then rejects that heat into the hot aisle through liquid-to-air heat exchangers. The practical appeal is simple: operators can add high-density liquid cooling rack by rack instead of converting an entire hall in one project.

Google said Brazos is already generally available, and its manufacturing suppliers are ready to work with the broader industry to produce the design. The company also plans to open-source the technical specifications, design principles and visual assets. That open-source angle is not charity alone; wider adoption can help standardize the retrofit layer around AI cooling before every operator invents a slightly different box.

The Numbers Are Rack-Level, Not Facility-Level
The system uses three cooling units and integrated rack manifolds. Each chassis takes 11 Open Units of rack height and is built for standard ORv3 racks from the Open Compute Project.

At rack level, Brazos is rated for a 60kW nominal thermal load across its three modules. Operators can run it with deionized water or a 25 percent propylene glycol mixture, while power comes through a 40-60V DC input connected to standard rack busbars.

The safety and operations details are more useful than the usual launch language. Brazos is certified to UL/CSA/IEC 62368-1 standards, includes leak detection and pressure relief valves, and has a built-in human-machine interface with remote management through Modbus over TCP. Those are the details facility teams will care about before they let liquid cooling sit next to production AI hardware.

Sidecars Are Becoming AI Infrastructure Tools
Google is not introducing the sidecar in isolation. Brazos extends work around Open Compute Project rack adaptations, where power and cooling are being separated from the compute rack to handle AI power densities.

In 2024, Microsoft and Meta announced Mount Diablo, an open rack design for AI data centers that separates power and compute into different cabinets. Google has also worked on Mount Diablo with Meta and Microsoft, including an AC-to-DC sidecar power rack that leaves the main rack space for GPUs, TPUs and CPUs.

AWS has explored a similar path with its In Row Heat Exchanger, a custom liquid cooling design that can be installed without changing air-cooled mechanical designs. The industry signal is hard to miss: AI clusters are forcing operators to upgrade old rooms in smaller, modular steps because full rebuilds are expensive, slow and disruptive.

The Hard Part Is Installation At Scale
Brazos gives data center operators a defined retrofit option: open specifications, OCP rack compatibility, 60kW rack-level cooling and a supplier path for production. That is useful for facilities that cannot wait for entirely new AI halls.

The real test is less elegant than the diagram. Operators will need to show that Brazos can be installed with limited disruption, maintained safely, integrated with existing airflow and produced at a cost that makes one-rack-at-a-time liquid cooling worth the trouble. If those economics work, Brazos becomes more than a Google design note. It becomes a practical bridge between yesterday's air-cooled rooms and the AI racks already arriving at the door.]]></content:encoded>
</item>
<item>
  <title>IFC Backs Sify’s India Data Centres With $371 Mn Package</title>
  <link>https://stechtimes.com/en/article/ifc-backs-sifys-india-data-centres-with-371-mn-package-mqk7uno1</link>
  <guid isPermaLink="true">https://stechtimes.com/en/article/ifc-backs-sifys-india-data-centres-with-371-mn-package-mqk7uno1</guid>
  <pubDate>Fri, 19 Jun 2026 00:54:26 GMT</pubDate>
  <category>cloud-data-centers</category>
  <description><![CDATA[IFC has committed $371 Mn to Sify Infinit Spaces Ltd. for two AI-ready data centres in Navi Mumbai and Chennai, adding another financing signal to India’s cloud infrastructure build-out.]]></description>
  <enclosure url="https://stechtimes.com/_article-images/39a94d9a57618f41-ifc-backs-sify-s-india-data-centres-with-371-mn-packag-1781830460498.jpg" type="image/jpeg" />
  <content:encoded><![CDATA[IFC Puts Debt Behind India’s AI Data Centre Build-Out
International Finance Corporation has committed $371 Mn to Sify Infinit Spaces Ltd., a Sify Technologies subsidiary, to develop two data centres in Navi Mumbai and Chennai. The financing gives India’s AI and cloud infrastructure market another large debt-backed expansion signal.

The package includes a $71 Mn loan and $300 Mn in additional support through debt mobilisation. IFC said the money will support SISL’s growth and expansion in India.

The two planned facilities will have a combined capacity of 103 MW. They are being designed for AI-ready workflows, energy-efficient cooling and renewable-energy operations, with construction aligned to Indian Green Building Council Platinum rating specifications.

Capacity, Cooling And Capital Are Moving Together
The Sify financing is not only about adding server space. IFC framed the project as part of the World Bank Group’s Country Partnership Framework for India. It said the investment would broaden cloud and AI-ready infrastructure access, draw more private money into data centres and support thousands of jobs.

SISL chief financial officer Ganesh Sankararaman said IFC’s role provides capital and a vote of confidence for the company’s expansion plans. The useful detail is the structure: a smaller direct loan paired with a larger debt-mobilisation component, rather than a simple equity headline.

For data centre operators, the source points to three linked constraints. AI workloads need power capacity, cooling design and financing depth. Sify’s project names all three, but it does not disclose anchor customers, construction schedules or expected utilisation.

India’s Data Centre Funding Is Getting Denser
The announcement followed another large India data centre deal. Canada Pension Plan Investment Board said it was investing ₹4,000 Cr to acquire an 8.2% stake in CtrlS, and CPPIB and CtrlS also plan a joint venture for hyperscale campuses across India.

Under that separate plan, CPPIB committed up to ₹3,000 Cr to the joint venture. CPPIB would hold 48% ownership, while CtrlS would hold the remaining 52%.

The source also points to wider demand drivers: AI demand, cloud expansion and data localisation requirements. India’s data centre capacity is projected to cross 8 GW by 2030, a target that explains why infrastructure investors are moving beyond single-facility announcements.

The Next Test Is Who Uses The Capacity
Sify now has a financing package, two named cities, a 103 MW capacity plan and sustainability specifications. Those are strong infrastructure facts, but they are not the same as customer proof.

The sharper evidence will come from named cloud customers, AI tenants, commissioning dates, power-procurement details and utilisation once the Navi Mumbai and Chennai sites move from financing to operation. Until then, IFC’s commitment shows capital formation around India’s data centre market, while the operating proof still depends on who fills the capacity.]]></content:encoded>
</item>
<item>
  <title>Alibaba Cloud Adds Tokyo Capacity As Japan AI Demand Rises</title>
  <link>https://stechtimes.com/en/article/alibaba-cloud-adds-tokyo-capacity-as-japan-ai-demand-rises-mqk5u5fm</link>
  <guid isPermaLink="true">https://stechtimes.com/en/article/alibaba-cloud-adds-tokyo-capacity-as-japan-ai-demand-rises-mqk5u5fm</guid>
  <pubDate>Thu, 18 Jun 2026 23:58:12 GMT</pubDate>
  <category>cloud-data-centers</category>
  <description><![CDATA[Alibaba Cloud has opened its fifth data center in Japan and made Model Studio available in the country. The expansion gives Japanese enterprises local access to Qwen3.7-Plus, AI-native database services and a larger regional cloud footprint.]]></description>
  <enclosure url="https://stechtimes.com/_article-images/1e145bc7c08c5db7-alibaba-cloud-adds-tokyo-capacity-as-japan-ai-demand-r-1781827077582.png" type="image/jpeg" />
  <content:encoded><![CDATA[Tokyo Gets More Local Cloud Capacity
Alibaba Cloud has opened a fifth data center in Japan, adding Tokyo capacity for companies working with cloud and AI workloads. The new site follows a fourth Japanese facility launched in March 2026, so the story is a local infrastructure expansion rather than a single product announcement.

The company is aiming the added capacity at Japanese customers in retail, gaming, entertainment and manufacturing. Those sectors matter because the release ties the new facility to agentic AI workloads, not only conventional compute, storage and networking demand.

Japan is now part of a broader Alibaba Cloud network that spans 105 availability zones across 32 regions. The new Tokyo site also extends the local service stack to elastic computing, storage, containerization, networking, security, databases and developer tools.

Model Studio Moves Into The Japan Region
The Japan expansion includes Model Studio, Alibaba Cloud’s enterprise AI development platform. Japanese developers and businesses can access Qwen3.7-Plus and third-party LLMs through the service, with online inferencing backed by enterprise-grade security.

The model menu is broader than one chatbot interface. Qwen3.7-Plus is presented as a multimodal agent model with vision-language, coding, tool-use and productivity workflow capabilities. HappyHorse is a video generation model, while Qwen3.5-Omni is designed for multimodal uses such as live streaming, voice assistants and video captions for gaming and entertainment.

Alibaba Cloud also launched AI-native database and data analytics services in Japan. The offerings include Data Agent for Analytics, Meta, DAS and DataBridge, covering analytical insights, data asset management, database operations and data ingestion.

The Operating Question Is Local Adoption
The useful test is whether Japanese enterprises use the new Japan region for production AI systems rather than trials. The source names no customers for the fifth data center and does not disclose capacity, power draw or capital spending for the Tokyo facility.

Takeshi Kurita, Alibaba Cloud’s general manager for Japan and South Korea, linked the new site to demand from Japanese enterprises for agentic AI. The company also plans workshops, hackathons and events for Japanese developers and startups in the coming months, giving the expansion a clear follow-up point: developer activity and disclosed production deployments in Japan.]]></content:encoded>
</item>
<item>
  <title>Policloud’s AI Cloud Deal Puts GPUs On Renewable Sites</title>
  <link>https://stechtimes.com/en/article/policlouds-ai-cloud-deal-puts-gpus-on-renewable-sites-mqjsus56</link>
  <guid isPermaLink="true">https://stechtimes.com/en/article/policlouds-ai-cloud-deal-puts-gpus-on-renewable-sites-mqjsus56</guid>
  <pubDate>Thu, 18 Jun 2026 17:54:40 GMT</pubDate>
  <category>cloud-data-centers</category>
  <description><![CDATA[Policloud has secured a €580 million framework contract with CloudGrid Energy to deploy 280 modular units, 29,000 GPUs and 35MW of compute capacity across 16 European sites by the end of 2027.]]></description>
  <enclosure url="https://stechtimes.com/_article-images/8133581df40f5e8f-policloud-s-ai-cloud-deal-puts-gpus-on-renewable-sites-1781805271930.png" type="image/jpeg" />
  <content:encoded><![CDATA[A Distributed AI Cloud Contract
Policloud has secured a €580 million framework contract with CloudGrid Energy for a European sovereign AI platform. The deal covers 280 modular Policloud units, 29,000 GPUs, two million CPUs and 35MW of compute capacity.

The structure is different from a single large data-center campus. Policloud’s units are containerized, waterless-cooled micro data-center systems. Each unit can house up to 400 GPUs, use up to 100 sqm of space and draw 500kW of power capacity.

CloudGrid Energy’s role gives the project its infrastructure angle. The company deploys compute at renewable energy sites and aims to colocate HPC infrastructure directly into a partner’s asset. Policloud CEO David Gurle said the deployment will be at renewable power farms.

Sixteen Sites, Five Countries
The CloudGrid Energy agreement covers 16 secured sites in France, Germany, Italy, Spain and Sweden. Policloud said the first unit in Bonne Voisine, France, is already operational, while all of the sites covered by the agreement are scheduled for deployment by the end of 2027.

That timeline turns the announcement into a delivery test. Europe’s sovereign AI debate often focuses on national control, but this contract is also about the practical mechanics of installing compute where power is available. The disclosed capacity is large enough to matter, yet still depends on execution across multiple countries, sites and energy assets.

Policloud also has to prove that modular systems can scale without losing operational discipline. The source does not name customers for the planned capacity, disclose utilization commitments or state which GPU models will be installed. Those gaps matter because sovereign AI infrastructure is judged by usable compute, customer adoption and reliability, not only by announced capacity.

Seed Funding Meets A Larger Buildout
Policloud was founded in 2025 by David Gurle, a former Microsoft executive and founder of Symphony. The company previously raised €7.5m in seed funding led by Global Ventures, with participation from MI8 Limited, OneRagtime, Inria and private investors.

The new CloudGrid Energy contract pushes Policloud beyond its earlier target. The company had previously said it was aiming for 100 systems with more than 25,000 GPUs by the end of 2026. Gurle said the latest deal and other planned announcements move the company beyond that 100-unit objective and toward a 1000-Policloud target slated by the end of 2030.

For cloud buyers and infrastructure partners, the evidence to watch is concrete: site delivery, GPU installation, power availability, named customers and workload performance across the distributed network. The Bonne Voisine unit gives Policloud a starting point; the larger question is whether the 16-site rollout can turn a framework contract into dependable sovereign AI capacity.]]></content:encoded>
</item>
<item>
  <title>Northslope Picks Abu Dhabi For Its Gulf AI Buildout</title>
  <link>https://stechtimes.com/en/article/northslope-picks-abu-dhabi-for-its-gulf-ai-buildout-mqjm6269</link>
  <guid isPermaLink="true">https://stechtimes.com/en/article/northslope-picks-abu-dhabi-for-its-gulf-ai-buildout-mqjm6269</guid>
  <pubDate>Thu, 18 Jun 2026 14:47:27 GMT</pubDate>
  <category>ai</category>
  <description><![CDATA[Northslope has opened its first Gulf office in Abu Dhabi after building a six-engineer UAE team. The move puts the former Palantir-linked AI software company closer to regional customers in logistics, asset management, energy, manufacturing and aerospace.]]></description>
  <enclosure url="https://stechtimes.com/_article-images/2b8d4cfd334d6bda-northslope-picks-abu-dhabi-for-its-gulf-ai-buildout-1781794041194.jpg" type="image/jpeg" />
  <content:encoded><![CDATA[Abu Dhabi Becomes Northslope’s Gulf Base
Northslope has moved from a small UAE engineering presence to a permanent Abu Dhabi office, giving the US AI software company its first base in the Gulf. The company was founded in 2024 by former Palantir employees and says its custom AI software runs on Palantir’s operating system.

The office is a regional expansion story, not just a branding exercise. Northslope already had six engineers in the UAE before it decided to deepen its presence. The company said the expansion followed strong early demand from organisations across the region.

Harvey Young, managing partner for Northslope’s UAE division, will be based full-time in Abu Dhabi. His role covers regional customer engagements, partnerships and business development across the Middle East.

Industrial AI, Not A Consumer App Push
Northslope’s disclosed customer focus is industrial and enterprise-heavy. The company serves clients in logistics, asset management, energy, manufacturing and aerospace, areas where AI tools are judged by deployment reliability, workflow fit and operational outcomes rather than app-store adoption.

That customer mix explains why Abu Dhabi is useful to Northslope. The company is placing engineers and commercial leadership near organisations that are trying to turn AI programmes into working systems. The source does not name specific UAE customers, contracts or revenue tied to the new office, so the expansion should be read as a regional positioning move rather than proof of local sales scale.

Northslope also brings a Palantir connection into a market already paying attention to deployed AI platforms. The company says it is backed by former Palantir leadership. Palantir has separately partnered with Dubai Holding and launched Aither for AI-powered transformation across Dubai’s public and private sectors.

Funding Gives The Office A Longer Runway
The Abu Dhabi move follows Northslope’s April Series A funding round. The company raised approximately $22 million, with Friend & Family Capital leading the round and Fifth Down and Leblon Capital joining.

That capital matters because forward-deployed AI work is labour-intensive. Northslope says its Abu Dhabi team combines customer-aligned engineering with production-grade delivery experience. For regional buyers, the test is whether the company can move from engineering access to named deployments, measurable workflow gains and durable partnerships.

Northslope has 136 employees and is based in Denver, Colorado. Its hub offices now include New York, London and Abu Dhabi. The next evidence to watch is concrete: disclosed Middle East customers, sector-specific deployments, partner names or implementation results from the Abu Dhabi team.]]></content:encoded>
</item>
<item>
  <title>Adobe Puts AI Assistants Inside Creative Cloud. The Test Is Control.</title>
  <link>https://stechtimes.com/en/article/adobe-puts-ai-assistants-inside-creative-cloud-the-test-is-control-mqjk7n7o</link>
  <guid isPermaLink="true">https://stechtimes.com/en/article/adobe-puts-ai-assistants-inside-creative-cloud-the-test-is-control-mqjk7n7o</guid>
  <pubDate>Thu, 18 Jun 2026 13:52:41 GMT</pubDate>
  <category>ai</category>
  <description><![CDATA[Adobe is rolling out AI assistants in public beta for Photoshop, Premiere, Illustrator, InDesign and Frame.io. The launch pushes conversational editing into major creative tools, but adoption will depend on whether professionals trust the assistants with real project work.]]></description>
  <enclosure url="https://stechtimes.com/_article-images/83507ddf95d617c8-adobe-puts-ai-assistants-inside-creative-cloud-the-tes-1781790756043.webp" type="image/jpeg" />
  <content:encoded><![CDATA[AI Moves Into The Editing Workspace
Adobe is rolling out AI assistants in public beta across Photoshop, Premiere, Illustrator, InDesign and Frame.io. The assistants are built for individual Creative Cloud applications rather than a single generic chatbot, so each product gets a version tuned to the work normally done inside that tool.

The launch puts conversational editing closer to the core creative workflow. In Premiere, the assistant is described as useful for reorganizing a video timeline. In Photoshop, it can understand layer structure and help turn written prompts into image edits. Illustrator’s assistant can help create editable vectors, while the InDesign version is tied to layout adjustments and document styling.

Frame.io also gets an assistant, but its job is different. The product is a collaboration and review platform, so Adobe is using the assistant for project organization, asset summarization, review tracking and help finding the right creative files. That makes the release less about novelty chat and more about whether AI can reduce the time spent navigating large creative projects.

Professional Adoption Depends On Precision
Creative tools are not the same as general consumer chat apps. A designer or editor may accept a rough draft, but production work depends on layers, timelines, formats, brand rules and review history. Adobe’s approach tries to handle that by making each assistant operate as a specialist inside the application where the work happens.

The practical question is control. If an assistant changes a timeline, adjusts a layout or edits an image, professionals need to understand what changed and how to reverse it. The source confirms prompt-based editing, app-specific assistants and public beta availability; it does not prove that the assistants can handle complex production files without manual correction.

That limitation is central to the product test. AI can make creative software feel faster when it handles repetitive steps, but it can also create friction if users must inspect every automated change. For agencies, media teams and freelancers, the value will come from reliable task completion, not from the presence of another chat panel.

Adobe Is Turning Firefly Into Workflow Infrastructure
The assistants are powered by Adobe’s conversational creative agent and connect with the company’s broader Firefly strategy. Adobe has been trying to position generative AI as a controlled part of creative production, not only as an image-generation feature. Putting assistants inside flagship apps gives that strategy more reach because Photoshop and Premiere are already daily tools for many professional users.

The rollout also shows why platform placement matters. A standalone creative chatbot has to pull users away from their projects. An assistant inside Photoshop, Premiere or InDesign can read the working context and offer edits where the file already lives. That does not guarantee accuracy, but it gives Adobe a distribution advantage over tools that sit outside the production environment.

The Useful Evidence Comes From Real Projects
The public beta is the first test of whether Adobe’s assistant model can survive professional habits. The near-term evidence will be concrete: whether editors keep using the Premiere assistant after the first experiment, whether designers trust Photoshop layer edits, whether Illustrator outputs remain editable, and whether Frame.io summaries actually help teams move reviews faster.

Adobe has named the applications and the assistant roles. It has not shown broad adoption data, paid conversion figures or independent productivity results for this beta. Until that evidence appears, the launch is best read as an important product-integration step: AI is no longer sitting next to Creative Cloud; Adobe is putting it inside the tools where creative work is already made.]]></content:encoded>
</item>
<item>
  <title>SK hynix Ships HBM4E Samples. AI Memory Buyers Still Need Volume Proof.</title>
  <link>https://stechtimes.com/en/article/sk-hynix-ships-hbm4e-samples-ai-memory-buyers-still-need-volume-proof-mqjhzi0n</link>
  <guid isPermaLink="true">https://stechtimes.com/en/article/sk-hynix-ships-hbm4e-samples-ai-memory-buyers-still-need-volume-proof-mqjhzi0n</guid>
  <pubDate>Thu, 18 Jun 2026 12:50:22 GMT</pubDate>
  <category>chips-semiconductors</category>
  <description><![CDATA[SK hynix has sent 12-layer HBM4E samples to major customers, citing 16Gbps per pin speed and a 48GB stack. The announcement moves the AI memory race from specification claims toward customer qualification and production timing.]]></description>
  <enclosure url="https://stechtimes.com/_article-images/bbd7981cc5444c55-sk-hynix-ships-hbm4e-samples-ai-memory-buyers-still-ne-1781787016946.jpg" type="image/jpeg" />
  <content:encoded><![CDATA[HBM4E Moves From Roadmap To Customer Labs
SK hynix said on June 18, 2026, that it has shipped samples of HBM4E, its next-generation DRAM for AI systems, to major customers. The company described the shipment as a scheduled 12-stack sample delivery, not a full production launch.

That distinction keeps the story grounded. Sample shipments put memory in customer qualification pipelines, where chipmakers, accelerator vendors and data-center system builders test performance, thermals and reliability before committing designs at scale. SK hynix said it will work with partners for mass production in a timely manner, but it did not give a mass-production date or name the customers receiving the samples.

What The Numbers Show Speed, Capacity And Heat
The 12-layer HBM4E product has a maximum data processing speed of 16Gbps per pin, according to SK hynix. The company also said power efficiency is up more than 20 percent from previous models, a claim aimed at AI training and inference systems where memory bandwidth and energy use are both design constraints.

Capacity is another part of the pitch. SK hynix said its Advanced MR-MUF process lets the HBM4E stack reach 48GB while maintaining structural stability. The same packaging technology improved heat resistance by 17 percent compared with the preceding HBM4, which matters because high-bandwidth memory is stacked vertically and must move heat out of dense compute packages.

Packaging Is Becoming Part Of The AI Compute Sale
HBM competition is no longer only a density contest. The SK hynix announcement puts packaging, latency and thermal handling into the same commercial frame as raw speed. The company said the latest interface and design optimization reduce data-transfer latency while keeping operation stable in high-bandwidth environments.

For AI data centers and large-scale computing systems, those claims point to a practical bottleneck: accelerators can sit idle if memory cannot feed them fast enough, and power or heat limits can cap performance before peak specifications are reached. SK hynix links HBM4E to that bottleneck by saying the product is designed to improve data processing for AI training and inference.

The company also frames HBM4E as a supply-chain continuation rather than a one-off laboratory part. It says its work on HBM3, HBM3E and HBM4 gives it mass-production and supply experience for customers planning next-generation infrastructure. That is the commercial argument behind the sample shipment: buyers need performance, but they also need a supplier that can qualify, package and deliver memory reliably enough for AI systems.

Qualification Is The Hard Part
SK hynix has already supplied HBM3, HBM3E and HBM4, and it is using that record to argue that customers can move toward next-generation infrastructure with less supply risk. Ahn Hyun, the company’s President and Chief Development Officer, said HBM4E builds on its technological capabilities and manufacturing expertise.

The company is headquartered in Korea and supplies DRAM and NAND flash for global customers. That background matters here only because HBM4E depends on both memory design and stacked-chip manufacturing discipline; a sample that meets a speed target still has to survive customer validation, thermal testing and production planning.

The missing evidence is customer-side validation. The source confirms samples, headline specifications and the packaging approach; it does not confirm purchase orders, named accelerator platforms, volume commitments or a production schedule. The next evidence to watch is narrow and concrete: customer qualification results, mass-production timing, and whether HBM4E appears in announced AI server or accelerator designs.]]></content:encoded>
</item>
<item>
  <title>Data Centers Have A Climate-Risk Problem. The Map Is Uneven.</title>
  <link>https://stechtimes.com/en/article/data-centers-have-a-climaterisk-problem-the-map-is-uneven-mqjfnyls</link>
  <guid isPermaLink="true">https://stechtimes.com/en/article/data-centers-have-a-climaterisk-problem-the-map-is-uneven-mqjfnyls</guid>
  <pubDate>Thu, 18 Jun 2026 11:45:24 GMT</pubDate>
  <category>cloud-data-centers</category>
  <description><![CDATA[A First Street study puts 79% of global data center capacity at elevated risk from acute climate hazards. The findings make resilience, power access and water systems part of the AI infrastructure buildout test.]]></description>
  <enclosure url="https://stechtimes.com/_article-images/0b02e99f97d3ac54-data-centers-have-a-climate-risk-problem-the-map-is-un-1781783119040.jpg" type="image/jpeg" />
  <content:encoded><![CDATA[Climate Exposure Moves Into The Data Center Ledger
A First Street study released Thursday puts 79% of global data center capacity in markets with elevated exposure to acute climate hazards. The hazards include flooding, extreme winds and wildfires, each capable of interrupting operations, increasing downtime and raising insurance or repair costs.

The climate-risk analytics firm examined 97 global data center markets. Its conclusion is a warning for a sector that is expanding around AI demand, cloud services and enterprise computing: location risk is no longer only a question of land, tax incentives, fibre routes or power price.

First Street CEO Matthew Eby said most real-asset underwriting still relies on historical data even though climate conditions have changed. His point narrows the issue for data center investors. A site can look attractive under older models while carrying exposure that shows up later through weather disruption, water stress, cooling costs or power-system fragility.

Acute Events Are Only Part Of The Exposure
The study separates sudden climate shocks from slower operating pressure. Just over half of all data centers globally sit in markets exposed to chronic climate stress such as extreme heat and drought. Those stresses can reduce energy efficiency and lift costs even without a single catastrophic event.

Jeremy Porter, First Street's chief economist, said backward-looking models miss climate-adjusted sources of risk. He pointed to precipitation models that rely on past rainfall patterns even as a warmer atmosphere holds more moisture and produces heavier rainfall.

That changes the diligence question for developers and investors. Data centers are typically expected to operate for 20 to 30 years, so a project underwritten around today's utility, cooling and insurance assumptions may face a different risk profile over its operating life.

Cooling Choices Show The Adaptation Problem
Some operators are already designing around resource constraints. Digital Realty CEO Andrew Power said earlier this year that almost all of the company's 300 data centers use either waterless systems or closed-loop water systems. He described those systems as avoiding evaporation after the company chooses to make the investment.

That example shows why climate adaptation is not only about stronger walls or roofs. Building envelopes can reduce damage from severe weather, but Porter said the more important question is how the facility connects to external systems: infrastructure, access routes, power access and the surrounding community.

For a data center, a hardened building is not enough if power delivery, roads, emergency access or local water availability become the weak point. The study pushes resilience planning beyond the parcel and into the network of services that keeps compute capacity online.

Asia-Pacific Carries The Highest Regional Share
The regional split is uneven. First Street found that Asia-Pacific has 89% of its data center capacity at risk, the highest share among the regions measured. The Americas were listed at 50%, while Europe, the Middle East and Africa stood at 46%.

The study also identified several fast-growing markets as exposed, including Northern Virginia in the U.S., Johor in Malaysia and Marseille in France. Nordic markets had the lowest climate risk.

For AI infrastructure buyers, the finding does not say those markets will fail. It does say that compute location, cooling design, power reliability, insurance assumptions and community-level infrastructure now belong in the same investment review. The next useful evidence will be project-level disclosure: which operators show resilience spending, power-access planning and water strategy before capacity is sold.]]></content:encoded>
</item>
<item>
  <title>UAE Sets A Social Media Age Gate. Enforcement Is The Hard Part.</title>
  <link>https://stechtimes.com/en/article/uae-sets-a-social-media-age-gate-enforcement-is-the-hard-part-mqjdyi2o</link>
  <guid isPermaLink="true">https://stechtimes.com/en/article/uae-sets-a-social-media-age-gate-enforcement-is-the-hard-part-mqjdyi2o</guid>
  <pubDate>Thu, 18 Jun 2026 10:57:37 GMT</pubDate>
  <category>cybersecurity</category>
  <description><![CDATA[The UAE Cabinet has barred children under 15 from social media accounts and full platform features. The rule puts age verification, child privacy, parental controls and platform compliance on a 12-month operating clock.]]></description>
  <enclosure url="https://stechtimes.com/_article-images/3ecadfd972e2733d-uae-sets-a-social-media-age-gate-enforcement-is-the-ha-1781780251626.jpg" type="image/jpeg" />
  <content:encoded><![CDATA[A Cabinet Rule Moves Child Safety Into Platform Design
The UAE Cabinet has set a national age gate for social media use, barring children under the age of 15 from creating, using or operating personal accounts. The restriction also reaches full platform features such as social interaction, publishing, commenting, sharing, public groups, open channels and other large interactive spaces.

The rule is not limited to one app category. It applies to social media services that let users create accounts or personal profiles, engage with other users, publish or share content, or rely on algorithmic systems to display, rank or recommend posts. It covers free and paid platforms whose services are available in the UAE or directed at users in the country.

That scope turns a child-safety decision into a platform-compliance problem. Social media companies will need technical and administrative controls that can block access for users below the age threshold while still handling privacy and data-protection duties.

The Middle Band Gets Protection, Not A Full Ban
The Cabinet resolution draws a second line for children between 15 and 16. They may use social media, but only with enhanced protective measures applied to their accounts. That makes the policy more granular than a simple access ban.

The listed safeguards cover content controls suited to age, limits on risky contact with unknown users, restrictions on time spent on the service and parental control tools. In practice, the platform account for a 15-year-old user is expected to behave differently from a standard adult account.

The resolution also treats recommendation systems as part of the compliance surface because it includes platforms that display, rank or recommend content through algorithms. That puts feed design, interaction features and account controls inside the same enforcement problem.

This is where implementation becomes difficult. A platform has to identify the user’s age, apply restrictions to the correct account, limit risky interactions and maintain parental tools without collecting more personal data than necessary. The resolution links safety controls directly to privacy obligations, not only content moderation.

Self-Declared Age Is No Longer Enough
The most operationally important line is the rejection of self-declared age as a valid verification method. Providers must use mechanisms that achieve a high level of accuracy in determining user age while meeting child privacy and personal data-protection standards.

The resolution also points to data minimisation, secure processing and limits on retention beyond the period strictly necessary. Those requirements narrow the room for blunt verification systems that gather excessive identity data simply to prove age.

Regular reviews and auditing are part of the mandate. That means compliance is not just a one-time settings change. Platforms will need evidence that their age-verification and protection systems continue to work after launch, after product changes and across services available to UAE users.

The 12-Month Clock Tests Compliance Capacity
The ruling gives platforms a compliance window of up to 12 months. Accounts created by children under 15 in breach of the resolution must be monitored, and platforms are expected to take immediate action to suspend or disable them.

For the UAE, the policy adds a clear digital-safety benchmark: under-15 access is barred, 15-to-16 accounts require extra safeguards, and self-declaration cannot carry the verification burden. For platforms, the unresolved test is practical: accurate age checks, limited data collection, parental controls and account enforcement all have to operate together before the 12-month deadline expires.]]></content:encoded>
</item>
</channel>
</rss>
