#datacenters

The U.S. AI infrastructure race just sped up again. OpenAI says Stargate has already cleared the 10GW target it set for 2029, with more than 3GW added in the last 90 days alone.

Microsoft detailed a new MicroLED-based datacenter networking system on March 17, 2026. The project matters because it tackles one of the less visible constraints on AI scaling: the energy, distance and reliability limits of the links connecting servers and GPUs.

Thinking Machines Lab said it signed a multi-year strategic partnership with NVIDIA to deploy at least one gigawatt of next-generation Vera Rubin systems. The companies also plan to co-design training and serving systems and widen access to frontier AI and open models for enterprises, research institutions, and the scientific community.

Meta and Arm say they will co-develop multiple generations of AI-focused data center CPUs, starting with the Arm AGI CPU. Meta says the program is meant to raise performance per rack, improve efficiency, and extend its custom silicon stack beyond accelerators alone.

Meta said on January 9, 2026 that new agreements with Vistra, TerraPower, and Oklo could support up to 6.6 GW of new and existing clean power by 2035. The company tied the effort directly to the energy demands of its growing AI infrastructure, including the Prometheus supercluster in Ohio.

Meta announced a new 1GW data center campus in Lebanon, Indiana with more than $10 billion in combined infrastructure and community investment. The project includes major hiring, long-term local assistance funds, and sustainability commitments tied to energy and water stewardship.

The White House released a January 13, 2026 fact sheet outlining an executive-order framework to accelerate U.S. frontier AI infrastructure. The policy links federal site access to strict conditions on clean power buildout, domestic semiconductor sourcing, grid-upgrade responsibilities, and consumer electricity protections.