News
AI SHIFT:

Indosat Bases AI Grid Plan On 80% Mid-Band 5G Rollout

Newsroom brief

Indosat says its Nokia 5G modernisation will add low-band 5G nationwide and mid-band 5G to about 80% of its network over three and a half years, but it has not disclosed site counts, capex or named AI Grid customers.

Verified against source materialEdited by SendTech Times Connectivity Desk
Indosat Bases AI Grid Plan On 80% Mid-Band 5G Rollout

Indosat Plans Low-Band 5G Across Its Nationwide Footprint

Indosat Ooredoo Hutchison is using its Indosat AI Grid 5G modernisation plan to turn a mobile-network upgrade into a distributed AI platform.

Vikram Sinha, the operator's president director and chief executive, said the Nokia programme will add low-band 5G across Indosat's full footprint, while mid-band 5G is expected to reach about 80% of the network within a three-and-a-half-year rollout window.

Sinha said the upgraded 5G RAN will provide a common architecture for connectivity and intelligence.

He described the AI Grid as a mix of centralised AI Factories and distributed AI computing capacity placed closer to users across Indonesia.

The stated ambition is broader than radio performance.

Sinha said the architecture is meant to let Indonesian-developed AI models and applications run in more places, including areas outside the largest urban markets.

Indosat is also using physical AI access points to make the plan visible before national AI-RAN deployment, including AI experience centres in Solo and Jayapura, Papua, an AI-RAN innovation centre in Surabaya and the Nvidia AI technology centre.

Those sites give the operator a training and testing layer while the wider radio upgrade proceeds.

Sahabat AI Gives The Rollout Its First Application Layer

Indosat is linking the network plan to Sahabat AI, its Indonesian-language AI model.

Sinha said Sahabat AI is already being used in government services, healthcare, education and agriculture, and he named public services, industry and digital services as sectors that could use the upgraded network.

The executive also named education, healthcare and enterprise functions as early application areas.

He said Sahabat AI is being used to support teachers and students, assist medical professionals as a digital co-pilot, and support business functions such as HR, finance and operations.

Those uses remain operator-described claims rather than independently measured adoption results.

Indosat did not publish customer counts, deployment volumes, service revenue or audited performance data for Sahabat AI in the material.

Nokia And Nvidia Remain Part Of The AI-RAN Roadmap

The 5G modernisation follows Nokia's announcement that it will supply new radio, baseband, Centralized RAN and automation technology for Indosat.

Nokia, Indosat and Nvidia are also collaborating on AI-RAN development.

Nokia's announcement said the companies planned field trials in Indonesia by the end of 2026 after demonstrating their first AI-RAN call at Mobile World Congress 2026.

Sinha said the modernisation supports future AI-RAN architecture and AI-enabled services with Nokia and Nvidia.

He said future uses could include Vision AI for cameras, sensors and drones, as well as Physical AI for robotics, automation and industrial operations.

The named examples keep the project centred on edge computing and network intelligence, not only national-language chatbot access.

Indosat Does Not Name Site Counts Or Commercial AI Grid Customers

The rollout still lacks several commercial details.

Sinha said Indosat is not disclosing the specific number of sites involved, and the material did not disclose capex, named enterprise AI Grid customers, service pricing, customer-level Sahabat AI adoption data or firm launch dates for the future Vision AI and Physical AI services.

Share this article
inXf

Related articles

More
Anterix And Lynk Test Satellite-To-Device Links For Utility Networks
Telco & Connectivity

Anterix And Lynk Test Satellite-To-Device Links For Utility Networks

Anterix and Lynk Global are testing whether 900 MHz private wireless spectrum can extend satellite-to-device connectivity into utilities and other critical infrastructure networks.

Oracle Warns AI Data Centre Buildout Depends On Power And Debt
Capital & Policy

Oracle Warns AI Data Centre Buildout Depends On Power And Debt

Oracle says its AI infrastructure expansion depends on data-centre capacity, power, components and customer payments. The company plans $70 billion in fiscal 2027 capex and about $40 billion in debt and equity, while renewal and power-cost risks remain public in its filing.

Heat Failures Put Data Centres, Telecoms Cabinets And Power Networks Under Strain
Cloud & Data Centers

Heat Failures Put Data Centres, Telecoms Cabinets And Power Networks Under Strain

Extreme heat is testing the physical systems behind digital services, from power transformers and telecoms cabinets to hospital data centres. BBC evidence from France and the UK shows outages at 40C, data-centre temperatures of 50.3C and rail cabinets that can exceed 70C, while no single national heat-proofing standard is named.

India's 8.33 GW Data-Center Pipeline Tests AI Infrastructure Readiness
Cloud & Data Centers

India's 8.33 GW Data-Center Pipeline Tests AI Infrastructure Readiness

Knight Frank India says the country's data-center development pipeline has reached 8.33 GW, with Mumbai, Hyderabad and Chennai showing how AI demand is turning capacity, power, fibre and subsea links into strategic constraints.

Hong Kong Offices Face An AI Infrastructure Test As Tenants Move Upmarket
Real Estate

Hong Kong Offices Face An AI Infrastructure Test As Tenants Move Upmarket

Knight Frank says AI adoption is adding energy, connectivity and training-space requirements to Hong Kong offices, putting older buildings under pressure as tenants favour better-equipped central assets.

Ciena's $50 Billion AI Network Target Puts Optical Capacity on the Hyperscaler Clock
Telco & Connectivity

Ciena's $50 Billion AI Network Target Puts Optical Capacity on the Hyperscaler Clock

Ciena says AI demand could roughly double its addressable market to about $50 billion by 2029 as hyperscalers and service providers invest in optical networking. It cited RLS Hyper Rail, DCOM, coherent modules and 400G/800G pluggable optics as demand areas while planning $250 million to $275 million in capex this year. The practical question is whether AI compute buildouts convert into durable network orders.

Keep Reading

More Stories

Latest
India AI Compute Buyers Face 36-To-52-Week GPU Lead TimesChips & SemiconductorsJul 6, 2026India AI Compute Buyers Face 36-To-52-Week GPU Lead TimesIndian AI infrastructure buyers are still reserving GPU capacity months ahead. NeevCloud cofounder Narendra Sen said next-generation enterprise AI GPU lead times now range between 36 and 52 weeks, while named Indian startups are shifting training and inference around scarce capacity.ESMA Warning Pushes Prediction Markets Toward Gambling And Derivatives RulesCapital & PolicyJul 6, 2026ESMA Warning Pushes Prediction Markets Toward Gambling And Derivatives RulesESMA warned that binary event contracts can fall under existing rules for binary options, derivatives or national gambling law. Kalshi, Polymarket and ADI Predictstreet show why prediction markets now face financial, gambling and insider-trading scrutiny.UK E-Scooter Ads Change After Road-Ban WarningsCapital & PolicyJul 6, 2026UK E-Scooter Ads Change After Road-Ban WarningsAmazon, Argos and Currys changed e-scooter wording after PA found commuter-style advertising that conflicted with UK rules limiting private e-scooters to private land.Anthropic Adds Nvidia BioNeMo Toolkit To Claude Science BetaAIJul 6, 2026Anthropic Adds Nvidia BioNeMo Toolkit To Claude Science BetaAnthropic has launched a Claude Science public beta that integrates Nvidia BioNeMo Agent Toolkit. AI News cited 18 of the top 20 drugmakers using BioNeMo, but the companies did not disclose beta-user counts, named Claude Science customers or independent benchmark methodology.Micron 9650 SSD Promises 28GB/Second Reads Before PCIe Gen6 Servers ShipChips & SemiconductorsJul 6, 2026Micron 9650 SSD Promises 28GB/Second Reads Before PCIe Gen6 Servers ShipMicron’s 9650 enterprise SSD is rated for up to 28GB/second reads and 5.5M random-read IOPS, with liquid cooling expected for the E1.S form factor. Public PCIe Gen6 server platforms were not yet available for a Computex demo.Takeda And Insilico Set US$600 Million AI Drug Discovery DealAIJul 6, 2026Takeda And Insilico Set US$600 Million AI Drug Discovery DealAI News reported that Takeda and Insilico Medicine signed an AI drug discovery collaboration with a potential value of about US$600 million. The companies did not disclose the disease targets, number of candidates or clinical handoff timetable.Socket Tracks 108 Malicious Packages In PolinRider Supply-Chain AttackCybersecurityJul 6, 2026Socket Tracks 108 Malicious Packages In PolinRider Supply-Chain AttackSocket reported 162 malicious release artefacts across 108 packages in the PolinRider supply-chain campaign. The report names Go, Packagist and Chrome extension exposure but does not identify victim companies.Digital Dubai Updates Data Manual For AI-Ready Government DataCapital & PolicyJul 5, 2026Digital Dubai Updates Data Manual For AI-Ready Government DataDubai Data and Statistics Establishment, part of Digital Dubai, has launched an updated Dubai Data Manual for government data governance, quality, sharing and compliance. The announcement links the manual to AI-ready public services, but does not disclose enforcement dates, penalties or adoption metrics for government entities.NYLIM Sees Tokenised Portfolios Beyond Stablecoin PaymentsFintech & Digital PaymentsJul 5, 2026NYLIM Sees Tokenised Portfolios Beyond Stablecoin PaymentsCoinDesk reported that New York Life Investment Management executive Thomas Sy sees tokenisation moving from stablecoin payments into personalised portfolios. The article cited $807 billion in NYLIM parent assets, $11 billion overseen by Sy’s team and missing infrastructure around collateral, clearing and prime brokerage.Microsoft Funds Frontier Company With $2.5 Billion For Enterprise AIAIJul 5, 2026Microsoft Funds Frontier Company With $2.5 Billion For Enterprise AIMicrosoft said it is putting $2.5 billion into Frontier Company, a new enterprise AI deployment unit with 6,000 specialists and early customers including Unilever and Novo Nordisk. Contract values, pricing and measured ROI remain undisclosed.Xiaomi Miloco 2.0 Connects Mijia Devices To Local Smart Home AI AgentAIJul 5, 2026Xiaomi Miloco 2.0 Connects Mijia Devices To Local Smart Home AI AgentZhidx reported that Xiaomi has released and open-sourced Xiaomi Miloco 2.0, a smart-home AI framework that connects Mijia devices, OpenClaw and household memory while keeping raw sensor data local and isolated from the agent.ASRock Rack Shows 1U Arm AGI Server Before CPU RampChips & SemiconductorsJul 5, 2026ASRock Rack Shows 1U Arm AGI Server Before CPU RampServeTheHome reported that ASRock Rack displayed a 1U4E1S-ARM server for Arm’s AGI CPU at Computex 2026, with one processor, 136 Neoverse V3 cores and PCIe Gen6 connectivity. The report did not list pricing, customer deployments or a commercial launch date for the 1U system.