- Company name
- Nvidia
- Sector
- Semiconductors (AI chips)
- About
- Nvidia designs graphics processing units, called GPUs, and related hardware. Data centers use these chips to train and run AI models. Nvidia also sells networking equipment that links many chips together inside a data center. Cloud companies and AI labs buy this hardware. This is most of Nvidia's revenue.
- Thesis
- The market prices Nvidia at a discount because some of its huge revenue growth may not be fully organic. Large buyers use debt, or financing Nvidia itself provides, to pay for chips. This makes some demand look circular, meaning Nvidia's own money may be funding purchases of its own chips. The Department of Justice is also checking whether Nvidia structured a license deal with AI startup Groq to avoid normal antitrust review. Nvidia also cut its near-term gross margin outlook, to about 74% next quarter and 71% to 72% the quarter after, down from 75%. For the gap to close, revenue growth must keep coming from many customers, not just financed ones. The new Vera Rubin chip platform must also keep shipping widely. The Groq inquiry must end without a forced change to the deal, and gross margin must stop falling. The next test is Nvidia's fiscal Q3 earnings call, expected in mid to late November 2026. A Justice Department decision could take longer, likely into 2027. If demand and margins hold up over the next two to three quarters, the gap can close.
- Risks
- Nvidia first planned to invest up to $100 billion in OpenAI, to help fund OpenAI's purchase of Nvidia chips for a 10-gigawatt data center buildout. Its final investment in OpenAI's 2026 funding round was $30 billion. This kind of vendor financing ties Nvidia's future revenue to the financial health of one customer. It makes some reported demand harder to separate from organic demand.
Google, Amazon, and Microsoft each build their own AI chips. Custom AI chip shipments are growing faster than Nvidia's own GPU shipments. Analysts expect these custom chips to take a growing share of AI work that currently runs on Nvidia GPUs.
Nvidia lowered its near-term gross margin outlook, to about 74% next quarter and 71%-72% the quarter after, down from 75% in the quarter just reported. This shows rising costs are starting to weigh on profit even as revenue keeps growing fast.
Nvidia's near-term outlook excludes China data center revenue, after a year of shifting US export rules for its chips. The Department of Justice opened an antitrust probe in September 2026 into whether Nvidia structured its $20 billion license deal with AI startup Groq to avoid normal antitrust review. Both issues add regulatory uncertainty to Nvidia's growth path.
- Balance sheet
- Nvidia holds much more cash than debt. At July 26, 2026, it had $56.6 billion of cash, cash equivalents and marketable debt securities. It also had $42.8 billion of listed equity securities. Against this, it had $33.5 billion of senior unsecured notes at the parent and $5.5 billion of operating lease liabilities. The notes increased in June 2026, when Nvidia issued $25.0 billion in seven tranches. Only $1.0 billion falls due within one year, and $17.5 billion falls due after five years. In the first half of fiscal 2027, operating income of $117.3 billion covered interest expense of $329 million more than 300 times. The $25.0 billion commercial paper program was unused. Moody's rates the notes Aa1 with a positive outlook (affirmed August 2026), and S&P raised its rating to AA in June 2026. The balance sheet does not show the largest calls on cash. These are $279 billion of supply and capacity commitments and $51.2 billion of stakes in private companies. Nvidia also guarantees lease and power payments for an OpenAI data center campus in Ohio, capped at $105 billion and starting from fiscal 2029. The cash and cash flow give room to keep buying back shares ($39.8 billion in the first half of fiscal 2027) and to invest. But the guarantees tie part of its future risk to the credit of one customer.
As of 2026-07-26.
- Checkpoints
- Nov 17, 2026 (estimate): Fiscal Q3 2027 results: gross margin and customer breadth. Pending. Watch: Gross margin near the 74% guidance and data center growth from many customers pass; a margin below guidance or growth that depends on financed buyers fails.
Feb 2027 (estimate): Fiscal Q4 2027 results: margin floor and Vera Rubin ramp. Pending. Watch: Gross margin that holds at or above the 71% to 72% guidance while Vera Rubin ships widely passes; a further margin cut or a slow Rubin ramp fails.
Q2 2027 (estimate): Hugging Face purchase closes after regulatory review. Pending. Watch: A close in the first half of 2027 on the agreed terms passes; a block or a long delay by regulators fails.
2027 (estimate): Justice Department decision on the Groq license deal. Pending. Watch: An inquiry that ends with no forced change to the Groq deal passes; a lawsuit or an order to undo the deal fails.
- Recent events
- 2026-09-28: Nvidia raised its share repurchase authorization by $150 billion. The total remaining authorization is now $235 billion, to be used through fiscal 2028.
2026-09-09: The US Department of Justice began looking into whether Nvidia structured its $20 billion license deal with AI chip startup Groq to avoid antitrust review. The deal gave Nvidia a non-exclusive license to Groq's chip technology. Nvidia also hired several Groq staff, including founder Jonathan Ross.
2026-09-03: Nvidia agreed to buy AI model platform Hugging Face for about $11.9 billion, plus up to $1.0 billion of retention awards for staff. Nvidia expects the deal to close in the first half of 2027, after regulatory approvals.
2026-08-26: Nvidia reported Q2 fiscal 2027 results. Revenue was $96.2 billion, up 106% from last year. Data center revenue was $89.0 billion, up 117% from last year.
2026-08-26: On its Q2 fiscal 2027 earnings call, Nvidia guided next-quarter gross margin to about 74%. It guided the quarter after that to 71%-72%, down from 75% in the quarter just reported.
2025-09-22: Nvidia and OpenAI announced a partnership. Nvidia will invest up to $100 billion in OpenAI to help deploy at least 10 gigawatts of Nvidia systems.