Juan Pedro Tomás
AI infrastructure push expands
The race to build next-generation AI infrastructure continues to intensify, with fresh moves across cloud compute, industrial platforms, and sovereign capacity. In the U.S., CoreWeave has expanded its partnership with OpenAI through a new $6.5 billion agreement, raising their total contract value to $22.4 billion and underscoring the company’s growing role as a key GPU provider for advanced model training. Meanwhile, Hitachi is leaning into “physical AI” with a distributed global AI factory built on Nvidia GPUs, integrating IoT, robotics, and automation to drive new efficiencies across rail, energy, and manufacturing. And in Canada, Telus has launched the country’s first sovereign AI factory in Quebec, a facility powered by Nvidia and HPE that ensures sensitive workloads stay within national borders. Together, these moves highlight how AI infrastructure is diversifying — spanning hyperscale partnerships, industrial transformation, and sovereign control. Let’s take a closer look.
Juan Pedro Tomas
Editor
RCRTech
AI Infrastructure Top 3
CoreWeave expands OpenAI deal: CoreWeave expands its partnership with OpenAI through a new deal worth up to $6.5 billion pushing total contracts to $22.4 billion while advancing global AI infrastructure and launching new growth initiatives.
Hitachi and Nvidia build AI factory: Hitachi is building a global AI cloud using Nvidia GPUs to deliver physical AI – industrial AI, plus domain-specific models – for rail, energy, factories, and infrastructure. It has agreed a deal to buy AI consultancy Synvert.
Telus opens sovereign AI hub: Telus has opened Canada’s first sovereign AI factory in Rimouski, offering secure domestic compute powered by Nvidia and HPE. The facility ensures data stays in-country while enabling healthcare, research, and enterprise AI.
AI-Powered Telecom Infrastructure
Supermicro, in collaboration with NVIDIA, delivers AI-powered infrastructure tailored for telcos, enhancing operational efficiency, network management, and customer experiences. Explore now
AI Today: What You Need to Know
SK Telecom launches AI unit: SK Telecom has launched an independent AI unit, pledging KRW5 trillion in investment to unify AI assets, drive company-wide transformation, and strengthen cybersecurity while targeting KRW5 trillion in annual revenues by 2030.
Bain warns shortfall: Consultancy projects $800 billion gap as AI demand outpaces global data center growth with trillions in revenue and capital needed to meet compute and power requirements by 2030.
Oracle debt raise: Company takes on $18 billion through bond sales to fund AI data center build-out including its multibillion-dollar partnership with OpenAI and expansion of Stargate capacity across the U.S.
IREN GPU expansion: AI cloud operator doubles fleet to 23k GPUs after $674 million procurement of Nvidia and AMD hardware targeting $500 million run-rate revenue by early 2026 as demand continues to surge.
CoreSite NY3 launch: New Jersey data center adds 138k square feet of AI-optimized capacity expanding CoreSite’s New York campus with direct cloud interconnections and liquid cooling for high-density workloads.
Coal plants extended: U.S. administration signals most coal-fired plants will delay retirement to meet AI power needs while boosting nuclear generation and invoking emergency powers to preserve grid capacity.
e&-Oracle alliance: UAE telco launches next-generation OneCloud powered by Oracle Alloy offering 200+ cloud services with sovereign hosting for AI and compliance-critical workloads inside national data centers.
Empower Semiconductor raise: U.S. chip power specialist secures $140 million Series D from Fidelity ADIA and others to scale breakthrough energy-saving solutions that target bottlenecks in AI processors and data centers.
Corintis cooling deal: Startup raises $24 million with Microsoft partnership to scale microfluidic cooling for AI chips demonstrating threefold heat removal efficiency compared with today’s leading data center technologies.
