Key Takeaways
- Nvidia shares advanced 1.6% to $177.97 during Wednesday’s premarket session
- ARM Holdings unveiled the ARM AGI CPU for data centers, projecting $15B yearly revenue by 2031
- ARM’s new processor doesn’t directly challenge Nvidia’s GPU lineup — though it may overlap with Nvidia’s newly released Vera CPU series
- Amazon Web Services committed to purchasing 1 million Nvidia GPUs for AI inference workloads, a transaction valued above $50 billion
- Nvidia has restarted H200 chip manufacturing for China and is developing a Groq 3 variant compliant with export rules, representing approximately $32B in annual sales
Nvidia shares gained traction in Wednesday’s early trading session, brushing aside concerns about ARM Holdings’ debut in the AI processor arena. The stock movement coincided with two substantial revenue announcements that had received minimal attention from the market.
Arm revealed its inaugural data center processor on Tuesday evening — the ARM AGI CPU — identifying Meta Platforms and OpenAI as initial adopters. During an after-hours investor presentation, ARM outlined aggressive growth expectations, forecasting approximately $15 billion in yearly CPU revenue by 2031, contributing to an overall $25 billion revenue target.
ARM’s shares surged 12% in premarket activity following the disclosure.
However, industry experts were prompt to clarify that this new processor doesn’t directly compete with Nvidia’s primary GPU offerings.
According to Benchmark Research analyst Cody Acree, ARM’s strategy is “less about catching up to the accelerator wave and more about inserting itself deeper into the architecture that governs how AI infrastructure actually runs.”
Nvidia’s CEO Jensen Huang appeared in ARM’s promotional content, describing their nearly two-decade collaboration as creating “one seamless platform, from cloud to edge to AI factories.”
The competitive dynamics become more nuanced regarding Nvidia’s standalone Vera CPU lineup, which debuted at last week’s developer conference. J.P. Morgan’s Harlan Sur noted potential overlap between ARM’s chip and that product category. He also highlighted that Meta has already contracted with Nvidia for its ARM-based processors — creating an ambiguous competitive landscape.
Amazon Web Services Makes Massive Nvidia Commitment
Separately, Amazon Web Services revealed plans to acquire 1 million Nvidia GPUs dedicated to AI inference capabilities. The announcement caught many observers off guard — AWS had recently been characterized as hosting “the largest cluster of non-Nvidia chips in the world” following its Indiana facility’s October 2025 launch.
The arrangement also encompasses a “broad mix” of six other Nvidia processor variants, including the recently introduced Groq 3 inference chips, alongside Nvidia networking equipment. Industry estimates place the complete package’s value well beyond $50 billion, with completion expected by late 2027.
This singular agreement accounts for approximately 25% of Nvidia’s complete 2025 yearly revenue.
Chinese Market Operations Resume
CEO Jensen Huang confirmed last week that Nvidia is resuming manufacturing of its H200 processor — engineered to meet U.S. export control requirements — for Chinese customers. Industry sources also suggest a China-compliant Groq 3 chip variant is under development.
Nvidia had projected zero Chinese data center revenue in its Q1 outlook. Throughout 2025, these sales averaged approximately $8 billion quarterly — translating to roughly $32 billion on an annualized basis, representing about 15% of total 2025 revenue.
Together, the AWS agreement and China production restart account for more than $82 billion in revenue streams not incorporated into Nvidia’s current financial projections.
Nvidia shares traded 1.6% higher at $177.97 during Wednesday’s premarket session, bouncing back from a 0.3% decline in the previous trading day.
