Key Takeaways
- NVDA shares advanced 1.6% during Wednesday’s premarket session, trading at $177.97
- Arm Holdings unveiled the Arm AGI CPU for data centers, projecting $15B yearly revenue by 2031
- The new Arm processor targets different market segments than Nvidia’s GPUs, though potential overlap exists with Vera CPU line
- Amazon Web Services committed to purchasing 1 million Nvidia GPUs for AI inference workloads, valued above $50 billion
- H200 chip manufacturing for China restarts alongside planned Groq 3 variant, representing approximately $32B in annual sales
Nvidia shares showed strength during early Wednesday trading hours, maintaining momentum despite Arm Holdings’ debut in the AI chip sector. Two additional major revenue announcements accompanied this development, both receiving limited market attention previously.
Arm revealed its inaugural data center CPU on Tuesday — the Arm AGI CPU — with Meta Platforms and OpenAI signed on as initial adopters. During an after-market presentation, Arm outlined aggressive financial targets, forecasting approximately $15 billion in yearly CPU revenue by 2031, contributing to an overall $25 billion revenue objective.
Arm’s shares surged 12% in premarket activity after the reveal.
Analysts emphasized that this new processor addresses different market needs compared to Nvidia’s primary GPU offerings.
Benchmark Research analyst Cody Acree noted that Arm’s strategy focuses on “inserting itself deeper into the architecture that governs how AI infrastructure actually runs” rather than competing directly in the accelerator market.
Nvidia CEO Jensen Huang appeared in Arm’s promotional materials, highlighting nearly two decades of collaboration between the companies as creating “one seamless platform, from cloud to edge to AI factories.”
The competitive landscape becomes more nuanced when examining Nvidia’s recently introduced Vera CPUs, unveiled during last week’s developer conference. J.P. Morgan analyst Harlan Sur identified potential competition between Arm’s offering and Nvidia’s standalone CPU products. He noted Meta’s existing agreement with Nvidia for Arm-based processors adds complexity to market positioning.
Amazon Web Services Announces Major Nvidia GPU Purchase
Simultaneously, Amazon Web Services revealed plans to acquire 1 million Nvidia GPUs dedicated to AI inference capabilities. The announcement surprised industry observers — AWS had recently promoted its Indiana data center, operational since October 2025, as hosting “the largest cluster of non-Nvidia chips in the world.”
The agreement encompasses a “broad mix” of six additional Nvidia processor variants, including newly released Groq 3 inference chips, along with Nvidia networking infrastructure. Total contract value exceeds $50 billion, with completion scheduled for late 2027.
This individual contract accounts for roughly 25% of Nvidia’s complete 2025 yearly revenue.
Chinese Market Sales Resume
CEO Jensen Huang confirmed last week that Nvidia has resumed manufacturing its H200 processor — engineered to meet U.S. export compliance requirements — for Chinese customers. Industry reports suggest a China-compatible Groq 3 chip variant is under development.
Nvidia had projected zero Chinese data center revenue in its Q1 forecast. Throughout 2025, these sales generated approximately $8 billion quarterly — roughly $32 billion on an annual basis, representing about 15% of total 2025 revenue.
Together, the AWS agreement and China market re-entry account for more than $82 billion in revenue absent from Nvidia’s current financial projections.
Nvidia stock traded up 1.6% at $177.97 during Wednesday’s premarket session, rebounding from a 0.3% decline in the previous trading day.
