Nvidia shares dropped 2% in Monday trading after a Wall Street Journal report highlighted internal doubts about the chipmaker's widely publicised $100 billion OpenAI investment.
CEO Jensen Huang told associates in Taipei over the weekend that the planned investment was "never a commitment" and would proceed "one step at a time"—language that contradicts the scale suggested when the deal was announced in September 2025.
What the deal actually is
The original agreement was a letter of intent for Nvidia to invest up to $100B progressively—$10B per gigawatt—to deploy 10 gigawatts of AI data centre capacity using Nvidia GPUs. The first 1GW deployment on Nvidia's Vera Rubin platform is scheduled for H2 2026. No definitive agreements have been signed, according to Nvidia's own filings.
OpenAI is separately raising up to $100B in a new funding round that could value it at $830B. Nvidia may participate with $20-30B in that round, but that's distinct from the infrastructure deal.
The circular funding problem
The arrangement raises familiar questions about circular AI economics. Nvidia invests in OpenAI, which buys Nvidia GPUs, which increases Nvidia's revenue, which supports its valuation. It's similar to Nvidia's $2B investment in CoreWeave, another major GPU customer.
Huang has reportedly expressed concerns about OpenAI's business discipline and competition from Google and Anthropic. Translation: he's watching whether OpenAI can sustain the GPU demand it's projecting.
What this means for enterprise
For CTOs tracking GPU availability, the non-binding nature of this deal is notable. If Nvidia can walk away from a $100B commitment to OpenAI, allocation priorities remain fluid. Enterprise customers should plan for continued GPU scarcity and keep alternative providers warm.
The first gigawatt deployment timeline—H2 2026—also suggests enterprise API costs from OpenAI won't materially improve before then. The batch API pricing advantages remain the main lever for cost optimisation in the near term.
Nvidia's stock reaction was modest because the market is adjusting to reality: the $100B figure was always more signal than commitment. The deal will happen in stages, if OpenAI's business holds up. We'll see.