What happened: Nvidia CEO Jensen Huang said the company’s recent $30 billion investment in OpenAI may be the last major equity stake it takes before the AI lab potentially goes public, and suggested a much larger $100 billion investment is unlikely.

Why it matters: Nvidia’s influence over the AI stack increasingly runs through capacity deals and infrastructure commitments, not just venture-style bets — and Huang’s comments hint at a shift from “owning pieces of labs” toward selling (and allocating) scarce compute at scale.

Wider context: The report ties Nvidia’s OpenAI relationship to broader uncertainty around earlier, headline-grabbing infrastructure plans, and to an industry transition from training-heavy spending toward inference capacity — the part of AI that touches products and revenue sooner.

Nvidia CEO Huang says $30 billion OpenAI investment 'might be the last' — CNBC

Singularity Soup Take: The “AI boom” story is turning into an allocation problem: who gets how much compute, on what terms, and for which workloads. If Nvidia is stepping back from giant equity stakes, it’s partly because the real leverage now sits in the contracts, not the cap table.

Key Takeaways:

• IPO changes the math: Huang said OpenAI going public could make further Nvidia investment unlikely, implying that late-stage equity access — not enthusiasm for OpenAI — is what may be disappearing.
• From milestones to cash: The piece contrasts earlier multi-year, deployment-linked deal structures with a newer $30 billion investment that is not tied to buildout milestones, suggesting partnerships are being renegotiated under different risk assumptions.
• Inference is the new pressure point: CNBC notes demand is shifting from training to inference, with Nvidia reportedly building inference-optimized chips and OpenAI buying dedicated inference capacity — a signal that the next wave of competition will be about serving users cheaply, not just training bigger models.