Nvidia Scores Pentagon AI Deal as Hyperscaler Spending Hits $725 Billion — But the Clouds Are Gathering

Nvidia joins SpaceX, OpenAI in Pentagon AI contract as $725B infrastructure spending surge and in-house chip pivot challenge growth outlook.

The US Department of Defense signed contracts with seven technology companies on Friday to deploy AI tools across its classified networks, with Nvidia securing a place in the lineup just three weeks before its next earnings report. The agreement, which also includes SpaceX, OpenAI, Google, Microsoft, Amazon Web Services, and Reflection, grants access to the Pentagon’s Impact Level 6 and 7 networks. The stated goal: accelerate data synthesis, sharpen situational awareness, and improve decision-making in complex operational environments.

The deal fills a void left by Anthropic, which publicly refused to relax safety restrictions on its Claude models — particularly around autonomous weapons and domestic surveillance. The Pentagon has stressed it has no plans for mass surveillance of US citizens or fully autonomous weapons systems, but insists that all legally permissible government AI use must remain accessible. For Nvidia, this opens a demand channel that barely registered in previous forecasts.

The announcement came on a US holiday, meaning markets get their first chance to price in the news when trading resumes Monday.

Hyperscalers Pile In — But With a Twist

The Pentagon contract adds a fresh revenue stream to a core business already straining to keep pace with demand. Amazon and Alphabet both reported quarterly results this week, flagging surging demand for compute capacity tied to AI workloads — a direct tailwind for Nvidia’s data center franchise.

But the picture is more complicated than simple growth. The world’s largest tech companies have raised their combined 2026 AI infrastructure spending estimates to a staggering $725 billion. Amazon, Google, Meta, and Microsoft are pouring money into data centers, yet they are simultaneously accelerating their pivot to in-house processors — Google’s TPU, Microsoft’s Maia chips — reducing their reliance on Nvidia’s GPUs.

That strategic shift hit Nvidia’s stock hard on the last trading day of April, with shares sliding nearly 5% to close at $199.57. The decline pushed the company’s market capitalization back below the $5 trillion threshold. Market observers framed the selloff as a necessary cooling after the recent rally.

China Risks and Gray Market Fever

A separate pressure point is building in Asia. A crackdown on chip smuggling into China is rattling the secondary market for high-performance hardware, where extreme supply constraints have pushed the price of Nvidia’s B300 servers toward $1 million per unit in sanctioned regions. Analysts warn that such price extremes could drive Chinese customers toward domestic alternatives. Nvidia’s management has previously cautioned that revenue from China could vanish entirely over the long term due to trade restrictions. The company has already booked $4.5 billion in inventory write-downs linked to geopolitical risks.

Fundamentals Still Impress — But Valuation Is Cautious

Nvidia reported fourth-quarter fiscal 2026 revenue of $68.1 billion, up 73% year over year, with free cash flow of $34.9 billion in the quarter. Data center networking revenue surged 263%. Yet the stock trades around $208, well below the average analyst price target of $268 — or $274.38 according to a separate consensus estimate that implies more than 35% upside from recent levels.

Of 42 analysts covering the stock, 40 rate it a buy, one a hold, and one a sell. The cautious valuation reflects mounting competition from AMD and Intel, the hyperscaler shift to custom silicon, and the China overhang.

What’s Next: Earnings on May 20

Nvidia reports first-quarter fiscal 2027 results on May 20. Management has guided for 77% revenue growth, while the Wall Street consensus has crept up to a range of 79% to 81%. The company will need to convince investors that its next-generation architectures — Blackwell and Vera Rubin — can maintain the technological edge against a growing army of custom AI processors built by its own biggest customers.