Nvidia Put $58.1 Billion Into Other Companies in Nine Months — 12 Times More Than a Year Ago
Nvidia doesn't just sell AI chips. It also invests in the companies that buy them. In the nine months to October 25, it spent $58.1 billion on stakes in other companies, including OpenAI and Anthropic — almost 12 times the $4.95 billion a year earlier. If AI demand slows, Nvidia loses twice: on chip sales and on the value of its stakes. Anyone renting GPUs should check what money keeps their providers afloat.
AI-processed from TNW; edited by Hamidun News
Nvidia sells AI accelerators to the whole world, but its latest quarterly report shows that the company increasingly hands money to the people who buy those accelerators. In the nine months ended October 25, it spent $58.1 billion on stakes in other companies.
A year earlier, the same period saw $4.95 billion, so the total grew almost 12 times. Picture a supplier that lends money to a customer, who then spends it on that supplier's products: sales grow, but part of the growth rests on the seller's own credit.
For developers, product managers, and founders, this is not abstract finance news. Compute prices, GPU availability, and the stability of AI startups that rent capacity all depend on this kind of arrangement.
Investments Up Almost 12 Times
The figure comes from the quarterly results Nvidia published on Tuesday. According to a comment from the chief financial officer quoted by Fortune, the money goes to "frontier labs, infrastructure financiers, and other private companies." The recipients include OpenAI and Anthropic.
For scale: over the same period Nvidia spent $61.2 billion on buying back its own shares ($36.3 billion a year earlier). In other words, its investments in other companies have almost caught up with the amount it returns to shareholders.
So far the result is positive: the stakes brought in $1.8 billion in net gains for the quarter. That figure does include a $1 billion loss on private holdings.
Guarantees With Conditions
Support for customers does not stop at equity. "Neoclouds" are companies that rent out computing capacity on Nvidia chips. According to interim CEO Colette Kress, Nvidia's future obligations to such companies came to $32 billion for the quarter. Credit guarantees reached $118 billion. Of that, $105 billion is tied to a data center in Ohio that OpenAI is leasing for 20 years.
Some of the guarantees take effect only once the customer already has financing in place. Kress explained that if construction is delayed or the borrower runs into problems, the guarantee stops applying. Put simply, this is not an unconditional check but insurance with conditions. Even so, that insurance ties Nvidia to its customers' fortunes more tightly than an ordinary hardware sale.
Kress called returning money to shareholders a "super priority." The buyback program has $227.3 billion left, and Nvidia plans to use it by the end of fiscal 2028. On Tuesday it also briefed analysts on future investments and guarantees, but the source material gives no details.
Why the Market Is Worried
Fortune puts the question bluntly: is Nvidia financing demand that then comes back to it as revenue? A month earlier, the IMF warned that such deals could amplify a market correction.
James Kisner of Water Tower Research described the risk this way: if AI spending slows, Nvidia loses twice — from weak chip sales and from the falling value of its stakes. In his words, the pain is amplified because the company is "doubly exposed."
What Practitioners Should Do
First, GPU prices and availability in the coming years depend not only on technology but also on the creditworthiness of OpenAI, Anthropic, and the neoclouds. If your product runs on rented capacity, check what money keeps your provider going and whether you depend on a single player.
Second, if you have long-term compute contracts, build in a slowdown scenario. A construction delay at your provider could hit you too.
Third, keep the limits of this news in mind. The numbers show the scale of the investments, but they do not prove those investments will lead to a crisis: the stakes are profitable so far, and some of the guarantees are conditional. Still, the growth was so sharp that Nvidia's financial reports deserve as much attention as new model releases.
- Nov 20, 2026Japanese Startup Open-Sources Haru-4B, a Voice AI That Listens While It Talks and Needs 12 GB of VRAM
- Nov 20, 2026Four Hours on an RTX 5090 Cost 348 ₽, or From 1 730 ₽ With a One-Day Minimum
- Nov 20, 2026Nvidia Is Close to Buying Inception for $500 Million to Get Diffusion LLMs
- Nov 20, 2026Nvidia quits selling H200 in China, and new US rules threaten $44–62 billion in lost revenue
Key Questions Answered
What are the key details on Investments Up Almost 12 Times?
The figure comes from the quarterly results Nvidia published on Tuesday. According to a comment from the chief financial officer quoted by Fortune, the money goes to "frontier labs, infrastructure financiers, and other private companies." The recipients include OpenAI and Anthropic. For scale: over the same period Nvidia spent $61.2 billion on buying back its own shares ($36.3 billion a year earlier).
What are the key details on Guarantees With Conditions?
Support for customers does not stop at equity. "Neoclouds" are companies that rent out computing capacity on Nvidia chips. According to interim CEO Colette Kress, Nvidia's future obligations to such companies came to $32 billion for the quarter. Credit guarantees reached $118 billion. Of that, $105 billion is tied to a data center in Ohio that OpenAI is leasing for 20 years. Some of the guarantees take effect only once the customer already has financing in place.
Why the Market Is Worried?
Fortune puts the question bluntly: is Nvidia financing demand that then comes back to it as revenue? A month earlier, the IMF warned that such deals could amplify a market correction. James Kisner of Water Tower Research described the risk this way: if AI spending slows, Nvidia loses twice — from weak chip sales and from the falling value of its stakes. In his words, the pain is amplified because the company is "doubly exposed."
Ship AI Faster with Founders, Engineers & Weekly Architecture Reviews
Inside Vibe Club (hamidun.com/club) — weekly live masterminds with Zhemal Khamidun, Tech Radar, full LMS access, and 370+ production Claude Code skills.
The AI world, distilled — once a week
Seven stories that actually mattered, hand-picked. No noise, no reposts, no press releases.
Done! Check your inbox for a confirmation.