AI infrastructure spending intensifies as Anthropic, Amazon, and NVIDIA sign multi-billion-dollar compute and chip deals, while a federal court hands Anthropic a landmark win against a Pentagon blacklist.
Overview
Today's Highlights
Regulation & Policy
Federal judge rules Pentagon's "supply chain risk" designation of Anthropic was unlawful
Published: 2026-08-28
Corporate
NVIDIA reportedly nearing ~$12.9B acquisition of Hugging Face
Published: 2026-08-27
Corporate
Anthropic signs 6-year, $45B compute deal with UK's Nscale
Published: 2026-08-26
Corporate
Amazon expands NVIDIA chip order to 2 million units for AWS data centers
Published: 2026-08-26
Regulation & Policy
Judge Blocks Pentagon's "Supply Chain Risk" Label on Anthropic
Published: 2026-08-28
A federal judge in California, Rita Lin, struck down the Department of Defense's designation of Anthropic as a "supply chain risk," ruling it violated the First and Fifth Amendments.
The court issued a permanent injunction against the designation, which had followed a breakdown in contract negotiations between Anthropic and the Pentagon.
The ruling is one of the first judicial checks on how the U.S. government uses risk designations against AI vendors.
Why it matters: The case is an early precedent for AI companies asserting contract terms — such as limits on autonomous-weapons or surveillance use — against national security agencies.
Sources say NVIDIA has agreed to acquire open-source AI model-sharing platform Hugging Face for roughly $12.9 billion.
No formal agreement has been announced, and the deal could still fall through.
If completed, it would be NVIDIA's largest acquisition ever.
Why it matters: Bringing model-distribution infrastructure into NVIDIA's ecosystem raises questions about neutrality in open-source AI distribution and further consolidation of AI infrastructure.
Anthropic will buy AI cloud compute capacity from a Nscale data center under development in West Virginia over six years, worth $45 billion total.
The site is roughly 460 megawatts and is planned to run on NVIDIA's Vera Rubin chips.
The deal follows Microsoft's withdrawal from part of its contract at the same site.
Why it matters: As Anthropic looks toward a potential IPO, the race to lock in compute capacity underscores how fast AI development costs and power-infrastructure demand are rising.
Amazon Expands NVIDIA Chip Order to 2 Million Units
Published: 2026-08-26
2M Chips
additional NVIDIA chips ordered
Amazon and NVIDIA announced an expanded partnership to add roughly 2 million Blackwell Ultra, Rubin, and Rubin Ultra chips to AWS data centers between 2027 and 2028.
The plan also includes supplying 100,000 GPUs for U.S. government use.
The two companies are also expanding collaboration in robotics.
Why it matters: The intensifying AI infrastructure race among major cloud providers directly affects chip supply and procurement costs for businesses building on AI.
Anthropic, Amazon, and NVIDIA have each signed compute or chip deals worth tens of billions of dollars in quick succession, intensifying the race for AI infrastructure.
NVIDIA's reported (unconfirmed) move to acquire Hugging Face signals a deeper push into vertical integration, extending its reach into model distribution.
Friction between AI companies and the U.S. government took a turn in Anthropic's favor via the courts, putting a spotlight on how AI vendors negotiate and defend government contracts.
Wrap-up
Three Things to Remember
Key takeaways
Anthropic won a landmark legal victory, with a federal court blocking the Pentagon's "supply chain risk" designation as unlawful.
NVIDIA is reportedly closing in on its largest-ever acquisition — Hugging Face for about $12.9 billion — though the deal is not yet confirmed.
AI infrastructure spending accelerated sharply, with Anthropic-Nscale ($45B) and Amazon-NVIDIA (2M chips) deals signed within days of each other.
What to watch
Whether NVIDIA and Hugging Face confirm the reported acquisition, how the Pentagon responds to the court's injunction, and whether more cloud providers announce comparable multi-billion-dollar compute commitments.