Two months ago in San Jose, NVIDIA’s Jensen Huang and OpenAI’s Sam Altman posed for photos and handed reporters a headline: NVIDIA would invest up to $100 billion over several years as OpenAI built huge new AI data-centres. The promise sounded like a turning point — the chipmaker putting cash behind the company that powers ChatGPT. The press releases talked about millions of GPUs and at least 10 gigawatts of data-centre capacity, with an initial gigawatt expected to be deployed in the second half of 2026.
But the champagne-on-stage moment and legally binding contracts are two different things. In NVIDIA’s latest quarterly filing — the company’s Form 10-Q filed with the SEC — a blunt sentence landed in the section investors read closely: “There is no assurance that we will enter into definitive agreements with respect to the OpenAI opportunity or other potential investments, or that any investment will be completed on expected terms.” That line is in the company’s risk factors, the place where firms flag the gaps between optimistic statements and contractual reality.
This is not just legal caution. Big, staged announcements are increasingly common in AI: they help shape narratives, reassure customers and suppliers, and move markets. But turning an announcement into a multiyear, multibillion-dollar capital plan requires deal terms, closing conditions, regulatory approvals, financing and — often — mutually agreed milestones. NVIDIA’s 10-Q makes that procedural reality explicit for investors.
NVIDIA has been explicit elsewhere about its enthusiasm. Company statements and its earnings call stressed the strategic logic: NVIDIA wants its systems deployed in the largest AI factories, and OpenAI’s compute needs would be a huge, ongoing source of GPU demand. On the earnings call, NVIDIA’s CFO Colette Kress even cited OpenAI data — weekly users, enterprise customers and healthy gross margins — as part of the bullish case for the market and the partnership. CEO Jensen Huang, meanwhile, leaned into the point that “everything that OpenAI does runs on NVIDIA today.” Those public comments show optimism, but they aren’t contracts.
Why the language matters
Investors and analysts parse filings for these exact phrases. When a company warns there’s “no assurance,” it isn’t necessarily indicating deal collapse; it’s saying several steps remain before money changes hands. For a $100-billion program — a figure that will be disbursed progressively as gigawatts of capacity come online — milestones, proof-points and legal documentation matter. NVIDIA’s filing also notes other investments and agreements it’s pursuing, underlining this is part of an active — and sometimes contingent — capital strategy.
The broader financing scramble
NVIDIA is not sitting on the sidelines. Since the announcement, it has publicized a string of big strategic investments: a roughly $5 billion stake and partnership with Intel, and, separately, commitments around Anthropic (and now reported investments there as well). Meanwhile, OpenAI hasn’t been alone in lining up hardware suppliers; it signed a formal, signed supply agreement with AMD for large-scale deployments — a pact that included a warrant for up to 160 million AMD shares, a legal document filed with regulators. That AMD agreement is signed, public, and contains concrete vesting and exercise terms — the kind of paperwork that the NVIDIA-OpenAI headline did not yet display in the same way.
The industry has also seen recent moves into Anthropic: news reports show NVIDIA and Microsoft committing capital to Anthropic as that company expands. Those announcements underline how multiple large chipmakers and cloud providers are jockeying to secure relationships with frontier model builders.
OpenAI’s numbers — big ambition, big gaps
Sam Altman himself has been public about revenue goals and infrastructure plans. In a post on X, he wrote that OpenAI expects to end this year above a $20 billion annualized revenue run-rate, and that the company is looking at roughly $1.4 trillion in infrastructure commitments over the next several years — numbers that help explain why OpenAI needs multiple capital partners. Those are company projections, not audited guarantees; they show why both OpenAI and potential investors are thinking at an enormous scale, and why outside capital will likely be required.
So, will the $100 billion happen?
Short answer: maybe — and not all at once. The public documentation and reporting say:
- NVIDIA and OpenAI announced a framework describing up to $100 billion of progressive investment tied to deployment milestones and at least 10GW of NVIDIA systems. That announcement is real and public.
- NVIDIA’s own SEC filing explicitly warns there is “no assurance” the company will enter definitive agreements or that investments will happen on expected terms — a standard but important investor caution.
- OpenAI has also signed binding, signed deals with other suppliers (for example, the AMD supply agreement that includes a 160-million-share warrant, dated Oct. 5). Those signed agreements contrast with the more conditional language in NVIDIA’s filings.
What to watch next
If you’re watching this story, look for a few concrete signals:
- Definitive agreements or 8-K/10-Q disclosures from NVIDIA or OpenAI that move the September framework into signed, conditional contracts. The phrasing in SEC filings will show whether milestones, escrow, or other protections are in place.
- Cash flows and initial tranches — earlier reporting suggested an initial $10 billion tranche could be made available as early steps are completed; whether that money arrives and under what conditions will matter.
- Regulatory or antitrust scrutiny — given the size and circularity of some AI deals, regulators could probe whether certain structures distort competition. News wires and filings will flag if such scrutiny starts.
Staged photos and executive optimism are part of Silicon Valley’s playbook. But public companies also have to tell investors the sober truth: nothing is final until it’s signed, filed and funded. NVIDIA’s recent words to investors are a reminder of that small, crucial difference.
Discover more from GadgetBond
Subscribe to get the latest posts sent to your email.
