Nvidia's Private Stakes Reach $65.5 Billion as Its Commitments Grow

Nvidia has disclosed that its stakes in private companies are now worth $65.5 billion, up from the $22 billion reported in January. TechCrunch reported the figure on 18 November 2026, following the company's earnings call on Tuesday, 17 November.
The equity book has grown quickly across several measures. Nvidia put $16 billion into equities in the past quarter alone. Over the first nine months of its fiscal year, the company invested $58 billion in equities, covering both public and private companies. A year earlier, in November 2025, Nvidia had reported spending $5 billion on equity over the same nine-month span. That gap is large enough that the change in deployment pace is the central point of the disclosure.
According to PitchBook, Nvidia has invested in nearly 350 startups over nearly three years. Named holdings include Anthropic, OpenAI, Thinking Machines, Reflection, Lambda, Nscale and Crusoe. The spread covers frontier model developers (companies building the largest AI models), neocloud operators (providers renting GPU-based cloud capacity to AI developers) and infrastructure builders. Together they sit along the stack from compute supply to model training and inference, the step where a trained model produces answers.
Nvidia also reported $118 billion in guarantees. Of that total, $105 billion serves as a backstop for OpenAI's $500 billion data center in Ohio. A backstop is a promise to cover losses if the borrower cannot pay. A backstop of that size changes the risk profile of the relationship, which is no longer only an equity position. Nvidia's exposure now includes contingent obligations, meaning payments that arise only if certain events occur, tied to the financing of a single very large facility. Those obligations sit alongside the mark-to-market value of its minority stakes, the paper value of holdings revised to match the latest funding round prices.
Colette Kress, identified as Nvidia's interim CEO, addressed capital return on the Tuesday call. She said: "We intend to return all excess free cash flow net of strategic uses to shareholders during this last quarter and going forward." Free cash flow is the cash left after running the business and paying for capital spending. The statement is an intention about the current quarter and beyond. It does not describe a completed distribution, and the phrase "net of strategic uses" leaves the boundary between shareholder returns and investment in the portfolio to the company's own judgment.
TechCrunch's headline links the stake growth to Jensen Huang's exit and to uncertainty over future deals. The company's leadership arrangement is the context in which Kress spoke.
The broader context here is a familiar one for anyone who has tracked the infrastructure cycle. A supplier that finances its customers ties its hardware demand to their capital raises, much as a shop that lends money to its regulars depends on those regulars staying solvent. Equity stakes, purchase commitments and backstops each deepen that tie in a different way. For the startups on the list, Nvidia's balance sheet becomes part of their funding base. For Nvidia, the valuation gains on private holdings now sit alongside guarantees large enough to matter in their own right.
In my view, the most useful lens for the coming quarters is the gap between mark-ups and realized cash. A $65.5 billion carrying value measures what private rounds imply, not what Nvidia can sell at that price. Shareholders are being told that excess free cash flow will return to them, while the same company holds a large and growing set of positions whose liquidity depends on future financings and exits. Whether those two commitments fit together will depend on how the strategic-use qualifier is applied in practice.
For practitioners, the immediate effect is that a single supplier now sits on both sides of much of the frontier-model and AI cloud capital structure. That concentration cuts both ways. It can accelerate the buildout of data center capacity the industry needs, and it can make the health of that capacity more dependent on one company's balance sheet and appetite for risk.


