News
CAPACITY TEST:

Core42's 42MW Lake Mariner Expansion Turns US Power Capacity Into a Gulf AI Cloud Signal

Newsroom brief

Core42 expanded its AI cluster at TeraWulf's Lake Mariner site in Buffalo, raising compute capacity there by 42MW to 60MW. The UAE-headquartered G42 company's AI cloud platform has ten global sites operational, with additional deployments planned for 2026. The practical question is whether Core42 exercises further Lake Mariner capacity and turns the expanded US footprint into durable hyperscale and enterprise workloads.

Verified against source materialEdited by SendTech Times Cloud & Infrastructure Desk
Core42's 42MW Lake Mariner Expansion Turns US Power Capacity Into a Gulf AI Cloud Signal
Image source: Data Center Dynamics

Core42 Adds US AI Capacity at Lake Mariner

Core42 has expanded its AI cluster at TeraWulf's Lake Mariner site in Buffalo, New York, increasing compute capacity there by 42MW to 60MW.

The move extends the US footprint of the UAE-headquartered G42 company's AI cloud platform as demand for high-power AI workloads continues to shape data-center capacity decisions.

The Lake Mariner site gives Core42 a larger US base for hyperscale, AI-native and large enterprise workloads.

Talal M. Al Kaissi, CEO of Core42, said the company is scaling US infrastructure in line with long-term deployment programs and that the added Lake Mariner capacity extends its global AI infrastructure build-out.

Capacity Options Define the Next Phase

TeraWulf operates the Lake Mariner data center, where an earlier capacity agreement with Core42 was said to cover more than 70MW.

The site is targeting 750MW at full build-out.

Core42 also holds a time-limited option for 135MW of capacity that is scheduled for delivery in early 2026.

Core42 has deployed a combination of AMD Instinct MI300 GPUs and Nvidia infrastructure at the site.

The company also has US deployments in Dallas, Sunnyvale, Stockton and Minneapolis, including Condor Galaxy supercomputers developed in collaboration with Cerebras.

Its AI cloud platform has ten global sites operational, with additional deployments planned for 2026.

AI Cloud Demand Reshapes Former Mining Sites

Lake Mariner was previously used by TeraWulf for Bitcoin mining, but the campus has shifted toward AI and high-performance computing customers.

Six buildings are planned at the site, with Wulf Den at 2MW, CB1 at 16MW and the first phase of CB2 at 42MW live.

The second phase of CB2, CB3 at 42MW, CB4 and CB5 at 168MW are expected to be energized this year.

Fluidstack is another AI cloud customer tied to Lake Mariner through Google-backed agreements with TeraWulf.

Fluidstack plans to lease 360MW of critical IT load from the campus through two 10-year deals, while Google took a 14 percent stake in TeraWulf across those agreements.

The signal is that AI cloud expansion is turning power-backed data-center campuses into strategic infrastructure for both Gulf-backed platforms and US site operators.

The practical question is whether Core42 exercises further Lake Mariner capacity and converts its expanded US footprint into durable enterprise and hyperscale workloads.

Share this article
inXf

Related articles

More
Iren Plans 800MW Australia AI Data Center Campus as Power Becomes the Capacity Gate
Cloud & Data Centers

Iren Plans 800MW Australia AI Data Center Campus as Power Becomes the Capacity Gate

Iren signed a transmission connection agreement for a planned 800MW data center campus in Bundey, South Australia. The project is Iren's first Australian foray and is expected to be energized in 2028 as the company shifts more cash flow toward AI cloud infrastructure. The practical question is whether Iren can turn grid-connected power, financing and GPU capacity into energized AI cloud campuses on the announced timelines.

AI Compute Scarcity Is Redrawing The Infrastructure Map
AI

AI Compute Scarcity Is Redrawing The Infrastructure Map

AI infrastructure projects in India, Africa, Brazil and the UAE show how power, chip access, data location and inference demand are pushing compute beyond the traditional U.S. hyperscale cloud map.

Google Says AI Demand Is Outrunning Grid Decarbonisation
Cloud & Data Centers

Google Says AI Demand Is Outrunning Grid Decarbonisation

Google said its electricity demand climbed 37% in 2025 and has grown more than 250% since 2019 as AI and cloud infrastructure expanded. The company also cited 1 GW of demand-response capacity and more than 12 GW of new clean generation deals, but did not disclose what share of its total computing load can shift during grid stress.

Policloud’s AI Cloud Deal Places GPUs at Renewable Energy Sites
Cloud & Data Centers

Policloud’s AI Cloud Deal Places GPUs at Renewable Energy Sites

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.

AI Data-Centre Spending Turns Energy Costs Into an Inflation Test
Cloud & Data Centers

AI Data-Centre Spending Turns Energy Costs Into an Inflation Test

AI infrastructure spending is pushing data-centre power demand, construction costs and debt issuance into the inflation debate. Pew Research Centre counted more than 3,000 operational US data centres and about 1,500 more under construction or in early development. The practical question is whether productivity gains arrive before power, construction and financing costs spread further through the economy.

Phoenix Group Takes UAE AI Data-Center Capital Into Europe With DC Max Deal
Cloud & Data Centers

Phoenix Group Takes UAE AI Data-Center Capital Into Europe With DC Max Deal

Abu Dhabi-based Phoenix Group is partnering with France’s DC Max on an 18MW AI-ready data center in Lyon, using the project as the first step in a platform targeting more than 1GW of AI and HPC capacity across Europe and the GCC. The deal centers on land, permits, grid access and the speed required to meet AI compute demand.

Keep Reading

More Stories

Latest
Red Hat Maps AI-RAN Upgrade Path Around Operator EconomicsTelco & ConnectivityJul 19, 2026Red Hat Maps AI-RAN Upgrade Path Around Operator EconomicsRCR Wireless News reported that Red Hat expects AI-RAN adoption to start with operational gains on existing radio networks before operators test shared AI and RAN infrastructure towards 2030.Superhuman Adds Docs With AI Assistant And MCP LinksAIJul 19, 2026Superhuman Adds Docs With AI Assistant And MCP LinksNo Jitter reported that Superhuman has launched Docs, an AI-native collaboration product evolved from Coda, with Docs AI in beta and MCP connections to tools such as Claude, ChatGPT, Cursor, Jira and Salesforce.H2LooP Raises $2m For Embedded-System AI Coding ToolsAIJul 19, 2026H2LooP Raises $2m For Embedded-System AI Coding ToolsYourStory reported that Bengaluru startup H2LooP raised $2 million to build hardware-aware AI coding tools for firmware and embedded systems, with early deployments in semiconductor and automotive teams.France And Germany Name Arcadia As Starting Point For Sovereign Military AICapital & PolicyJul 19, 2026France And Germany Name Arcadia As Starting Point For Sovereign Military AITNW reported that France and Germany pledged to examine a European sovereign digital backbone for military AI, using France's Arcadia as the starting point after both countries moved intelligence services away from Palantir.Moonshot Sets 27 July Open-Source Release For Kimi K3 AI ModelAIJul 19, 2026Moonshot Sets 27 July Open-Source Release For Kimi K3 AI ModelBBC reported that Moonshot AI launched Kimi K3, a 2.8 trillion-parameter model scheduled for open-source release on 27 July. The Chinese startup says the model can rival OpenAI and Anthropic systems, but exact hardware requirements and enterprise customers remain undisclosed.Meta Smart Glasses Face Privacy Backlash As Courtroom Bans SpreadCapital & PolicyJul 19, 2026Meta Smart Glasses Face Privacy Backlash As Courtroom Bans SpreadYahoo Tech reported that backlash against AI-enabled smart glasses is growing as Meta expands camera-equipped eyewear, with courthouse bans, facial-recognition concerns and cloud-review warnings shaping the privacy debate.Nebius Raises $775M Secured Debt Facility For GPU Cloud BuildoutCapital & PolicyJul 19, 2026Nebius Raises $775M Secured Debt Facility For GPU Cloud BuildoutNebius said its first senior secured debt facility is backed by deployed GPU infrastructure and contracted cash flows, with the $775 million financing priced at SOFR + 2.50% and maturing on October 31, 2030.White House Pushes New Gatekeeping Role For Frontier AI Model AccessCapital & PolicyJul 18, 2026White House Pushes New Gatekeeping Role For Frontier AI Model AccessCNBC reported that the Trump administration is pressing deeper into access decisions for frontier AI models from OpenAI and Anthropic, while a White House official said release timing and scope remain company decisions.Apple And US DOJ Discuss Settlement In 2024 iPhone Antitrust SuitCapital & PolicyJul 18, 2026Apple And US DOJ Discuss Settlement In 2024 iPhone Antitrust SuitAppleInsider reported that Apple and the US Justice Department are in early settlement discussions over the 2024 iPhone antitrust lawsuit, while any agreement, trial path and final court order remain unresolved.NVIDIA Sets Jetson T3000 And T2000 Modules For Q1 2027 Robotics LaunchChips & SemiconductorsJul 18, 2026NVIDIA Sets Jetson T3000 And T2000 Modules For Q1 2027 Robotics LaunchNVIDIA introduced Jetson T3000 and T2000 modules for edge AI and robotics systems, with Q1 2027 availability planned and source-listed performance, memory and emulation details still tied to company-provided figures.Mozilla Sets Firefox Desktop And Android For Two-Week ReleasesDevices & Consumer TechJul 18, 2026Mozilla Sets Firefox Desktop And Android For Two-Week ReleasesThe Register reported that Mozilla plans to move Firefox Desktop and Android from a four-week release cadence to a two-week experiment in September 2026, while ESR releases remain annual and Firefox 153 is due on July 21.EY Adds Knowledge Graphs To Enterprise RAG FrameworkAIJul 18, 2026EY Adds Knowledge Graphs To Enterprise RAG FrameworkSiliconANGLE reported that EY has developed a multimodal RAG framework that retrieves text, images, charts and diagrams through separate indexes and a knowledge graph, while the white paper did not publish comparative benchmark results.