American semiconductor company and maker of the world's leading artificial-intelligence chips. Nvidia accounts for about 90% of commercially available AI chips; AMD is its only serious American rival. In July 2025 Nvidia became the first company ever to reach a market value of $4trn. Its high-end processors are produced by TSMC in Taiwan. In late 2025 Jensen Huang said Nvidia had $500bn-worth of orders to deliver that year and the next. Jefferies, an investment bank, expects it to ship 6.3m AI chips in 2026, a 23% increase on 2025. Its software platform, CUDA, is by far the most widely used for programming AI chips; nearly all AI developers learn it, and it works only with Nvidia's hardware. None of Nvidia's 20 biggest suppliers is Chinese.
Before America's first export controls in October 2022, China accounted for about 22% of Nvidia's revenue. After those controls barred sales of its most powerful chips, Nvidia created the H800, a made-for-China model engineered to stay just under the limits. When America tightened the rules again in late 2023, banning any chip with too much computing power regardless of memory bandwidth, Nvidia responded with the H20, a chip that was inadvertently hobbled for training new AI models but well suited to inference—the process of running them. The Trump administration then, in effect, banned the H20 as well. Nvidia said the new rules would wipe $5.5bn from the value of its inventory.
Nvidia earns about a sixth of its revenue from China. On July 14th 2025 Nvidia announced it would again be permitted to sell H20 AI processors in China, after a ban that may have cost the company around $10bn in forgone Chinese revenue. The lifting could raise sales by $10bn-15bn in 2025 and net profit by $6bn-9bn. Nvidia's market value jumped $160bn to nearly $4.2trn the day after. However, if Nvidia lost China entirely it would benefit Huawei, which already offers AI chips more powerful than the H20.
On August 11th 2025 Trump confirmed the resumption of H20 sales to China, but imposed a 15% surcharge on the proceeds—in effect an export levy. Assuming H20 sales generate $20bn of revenue, the surcharge would net $3bn for the American government. AMD was also permitted to sell some of its AI chips to China. The export levy probably violates Article 1 of the constitution and may face legal challenge. Trump floated a similar surcharge arrangement for Nvidia's more advanced Blackwell chips. China has pressed Chinese firms to shun the H20s.
China's share of Nvidia's revenue has fallen to 13%. At the same time, sales to Singapore have more than doubled and now make up nearly 18% of the total, making it Nvidia's second-largest market after America. The company says many clients invoice through Singapore but ship to permitted destinations; fewer than 2% of chips sold there are delivered locally. In February 2025 Singaporean police arrested three men over the sale of $390m-worth of servers incorporating Nvidia chips, allegedly re-exported to Malaysia. Banned Nvidia chips now sell at a 30-50% mark-up through intermediaries on the grey market.
On March 19th 2026 federal prosecutors charged the co-founder of Supermicro, an American equipment-maker, with smuggling $2.5bn-worth of AI servers containing advanced Nvidia chips to Chinese customers—the largest such case to date. The value of those servers represents a tiny portion of the $194bn-worth of AI hardware Nvidia sold in the 12 months to January 2026. Nvidia insists that illicit diversion is a "losing proposition", since smuggled servers receive no service or support. In December 2025 Trump had relaxed an earlier ban on the sale of Nvidia's China-specific H200 chips. On March 17th 2026 Jensen Huang said orders were at last starting to flow again.
The Bureau of Industry and Security (BIS), the agency tasked with enforcing American export controls, has just one export-control officer responsible for all of South-East Asia and Australasia.
China's State Administration for Market Regulation (SAMR) is probing Nvidia alongside Alphabet. SAMR said in September 2025 that Nvidia had broken China's anti-monopoly laws in its 2020 acquisition of Mellanox, an Israeli supplier of computer networks—even though SAMR approved the deal at the time and did not specify which conditions Nvidia had breached. The antitrust agencies of the Trump administration are also investigating the company.
On December 8th 2025 Trump authorised the sale of Nvidia's H200 chips to "approved customers" in China, in exchange for a 25% cut for Uncle Sam. The H200, though lagging behind Nvidia's latest Blackwell chip and the forthcoming Rubin, is more than six times as powerful as the H20. Bloomberg Intelligence estimated the approval could net $10bn-15bn in annual revenue. Chinese regulators have urged local companies to avoid Nvidia altogether, citing security concerns. Huang admitted he had "no clue" whether Chinese authorities would permit the purchase of H200s; China's government was said to be considering allowing only limited access. David Sacks, Trump's AI adviser, argued against cutting China off entirely, a view that prevailed over that of China hawks in Trump's circle.
In late 2025 Alphabet, one of Nvidia's biggest customers, emerged as its fiercest competitor. Google launched Gemini 3, trained entirely on its own tensor-processing units (TPUs), and began selling them to others. Anthropic announced plans to use as many as 1m TPUs in a deal reportedly worth tens of billions of dollars. Reports that Meta was in talks to use Google's chips in its data centres by 2027 caused Nvidia to lose more than $100bn in market value on November 25th 2025, though it partly recovered. Bernstein estimates that Nvidia's GPUs account for over two-thirds of the cost of a typical AI server rack. Google's TPUs cost between a half and a tenth as much as an equivalent Nvidia chip. Jensen Huang described Google as "a very special case" given that it began developing chips long before the current AI wave, and dismissed other efforts as "super adorable and simple". Other tech giants including Amazon, Meta and Microsoft have been developing custom processors, and OpenAI announced a collaboration with Broadcom to develop its own silicon. Nvidia's edge lies partly in CUDA, which AI developers have become accustomed to; Google's TPU software has been created with its own products in mind, whereas CUDA caters to a wide range of applications.
Over the past 12 months Nvidia's gross, operating and net profit margins tightened by six, four and three percentage points as a share of sales, respectively, even as revenue continued to swell at an annual rate of around 60%.
On October 28th 2025 Nvidia held its first major jamboree in Washington, DC. Jensen Huang pledged to help Trump reindustrialise America. In mid-October 2025, nine months after Trump told Huang he wanted to reshore manufacturing, TSMC produced a wafer for Nvidia's most advanced GPU, Blackwell, at its Arizona fab. Huang said Blackwell's processors, memory and packaging would all soon be made in America. He also described how Foxconn, using AI-created simulations ("digital twins") built on Nvidia technology, was building fully robotic factories in America.
On October 29th 2025 Nvidia became the world's first $5trn company.
According to PitchBook, a data provider, Nvidia invested in 50 firms in 2025 alone; at the end of July it valued its ownership of private companies at just $4bn. It owns around 7% of CoreWeave, a highly leveraged neocloud expected to spend more than $10bn on Nvidia's chips in 2025. Nvidia recently signed a $6bn agreement to guarantee CoreWeave's revenue. It has also put $5bn into Intel and is providing some of the $20bn being raised by xAI. On November 18th 2025 Nvidia and Microsoft pledged to invest $15bn in Anthropic, which will for the first time train its models on Nvidia's chips.
In September 2025 Nvidia invested $5bn in Intel, as part of a deal to boost its beleaguered chipmaking rival. Days later it announced a proposed investment of up to $100bn in OpenAI, starting in the second half of 2026, to help the maker of ChatGPT buy 4m-5m of Nvidia's AI chips. The investment would increase in $10bn increments for every gigawatt of Nvidia-supported data-centre capacity that OpenAI builds, up to 10GW. Jensen Huang said the extra 5m GPUs would be roughly equal to all of Nvidia's GPU shipments in 2025. An unspoken benefit is that the deal would make OpenAI more reliant on Nvidia's chips, reducing the incentive to build its own. Pierre Ferragu of New Street Research calculated that Nvidia would invest $10bn for every $35bn of GPUs that OpenAI buys—meaning OpenAI pays 71% in cash and 29% in shares. Critics worry about the "circular dynamics" of Nvidia investing in companies it supplies with GPUs. Nvidia's share price climbed almost 4% after the announcement, raising its value to close to $4.5trn.
In 2025 Nvidia said it would unveil a fresh AI chip every year rather than every couple of years. Jensen Huang remarked that "when Blackwell starts shipping in volume, you couldn't give Hoppers away," referring to Nvidia's latest chips and their predecessors. This rapid obsolescence raises questions about how the big cloud companies depreciate their servers. Microsoft pushed server lifetimes from four to six years in 2022; Alphabet did the same in 2023; Amazon and Oracle changed from five to six in 2024; and Meta moved from five to five and a half years in January 2025. Amazon then reversed course for some kit, moving back to five years, noting this would cut operating profit in 2025 by $700m. Jim Chanos, a veteran short-seller, has argued that if the true economic lifespan of Meta's AI chips is two to three years, then "most of its 'profits' are materially overstated." A Barclays analysis estimated that higher depreciation costs would shave 5-10% from the earnings per share of Alphabet, Amazon and Meta. Nvidia's top three customers—widely believed to be Alphabet, Amazon and Microsoft—each spent some $15bn on its chips in 2024.
In August 2025 Nvidia reported quarterly revenue of close to $47bn, beating analyst forecasts of $46bn. Its latest Blackwell graphics-processing units (GPUs) and GB-series AI superchips (which combine two Blackwells with a general-purpose processor) accounted for almost 60% of total revenue. Nvidia probably sold over 600,000 Blackwells and nearly as many GBs in the quarter, nearly 20% more than the previous quarter. It is on track to sell 2.7m Blackwells and 2.4m GBs in the year. Between February 2024 and February 2026, analysts predict Nvidia will have sold some 6m Blackwells and 5.5m GBs.
A single Blackwell chip needs 1kW of power, three times more than its Hopper predecessor. Nvidia sells modules packed with 36 GB superchips (72 Blackwells and 36 general-purpose chips) designed to operate at 132kW; a secondary cooling system can add 160kW per rack. The extra power requirements are staggering: if half the chips sold end up in America and are operated at capacity, they would raise American power demand by 25 gigawatts. Analysts estimate a potential American power shortfall of 17-62GW by 2028-30. Nvidia plans to launch next-generation Rubin chips in 2026.
In May 2025 Jensen Huang, Nvidia's boss, warned that America is in danger of being left behind if American firms do not compete in China as it builds a "rich ecosystem" of AI applications. If they do not, Chinese technology and leadership "will diffuse all around the world", he said.
In late 2023 Jensen Huang began advocating that every country should have its own AI system—"sovereign AI." Governments are a lucrative new revenue source: Jefferies estimates sovereign initiatives could generate $200bn in cumulative revenue; Nvidia reckons spending could reach $1trn.
A paper published in June 2025 by Nvidia Research states that "small, rather than large, language models are the future of agentic AI." It argues that small language models (SLMs), with 40bn parameters or fewer, are sufficiently powerful to handle agentic tasks and more economical—a 7bn-parameter model can be ten to 30 times cheaper to run than a model up to 25 times bigger. It envisages a "Lego-like" approach to building agents with small, specialised models. The paper does not represent Nvidia's strategic thinking; the company maintains that business customers want models "of all shapes and sizes." Nvidia Nemotron Nano, a 9bn-parameter model, outperforms a Llama model released in April 2025 by Meta that is 40 times bigger.
Jensen Huang has more than 50 direct reports—an unusually wide "span of control" in management parlance. Mr Huang does not hold one-to-one meetings.
In 2022 Nvidia published the first results from FourCastNet, an AI weather program trained on decades of weather data. The company claimed it accurately predicted hurricanes and rainfall a week in advance with just two seconds of computing time—thousands of times less than what a numerical weather prediction model needs. The release kicked off a race among tech firms and weather agencies to build AI weather models of their own. See AI weather forecasting.
Nvidia has a generative-AI platform for drug discovery, and is signing deals to offer molecule-design services to pharmaceutical companies. In October 2025 it teamed up with Eli Lilly, the world's most valuable drugmaker, to build the pharma industry's most powerful supercomputer. On January 5th 2026 Jensen Huang declared that "the ChatGPT moment for robotics is here." During the Consumer Electronics Show that week, Nvidia unveiled a suite of chips and freely available AI models designed specifically for robots. See industrial robots.
Nvidia is carving out a role as an essential supplier for self-driving vehicles. It sells AI chips both to Waymo (to run simulations and sit inside its cars) and to carmakers developing autonomous systems of their own, including Mercedes-Benz and Stellantis. Tesla has spent billions training its AI systems on 100,000 Nvidia GPUs. Ali Kani, Nvidia's head of automotive, has suggested that once Waymo proves its safety case, it may be able to cut costs by shedding some hardware.
At its annual developer conference on March 16th 2026, Nvidia unveiled plans to become a "foundational company" on which the AI economy rests—selling different types of chips, bundling products into complete "AI factory" systems and embedding its technology more deeply into industries. In the year to January 2026 the firm generated $216bn in revenue, eight times what it made three years earlier. Its free cashflow is greater than those of other tech giants; it holds more than $62bn in cash, a third of it generated in the past year. Its GPUs account for over two-thirds of the total processing power available on the world's AI chips.
In December 2025 Nvidia paid $20bn to license technology and hire engineers from Groq, a startup specialising in inference chips. On March 16th 2026 it unveiled the Groq 3 LPX, a chip designed specifically for inference, with an architecture that departs from the traditional GPU. The chip has about 500 megabytes of on-chip SRAM—tiny compared with the B200's 192 gigabytes of off-chip memory—but relies on smart software to choreograph data movement and maximise computation and memory access. The chip is expected to reach the market later in 2026. It is pushing into CPUs using designs from Arm and in February 2026 struck a deal with Meta to supply CPU-only servers. Its networking equipment business generated $11bn in quarterly revenue.
Nvidia has released several families of open-source AI models aimed at specific industries, including Alpamayo for self-driving cars, GR00T for robotics and BioNeMo for biomedical research. Mercedes-Benz will soon ship vehicles equipped with Nvidia's self-driving systems.
Revenue from sovereign AI tripled in the last fiscal year to more than $30bn, about 15% of Nvidia's AI sales. Since 2020 the firm has made some 200 investments totalling over $65bn, including a $30bn bet on OpenAI. In March 2026 it invested more than $4bn in companies developing optical interconnects, which use light rather than wires to transfer data between chips.
Among conventional rivals, chip startups raised $17bn in 2025, more than in the previous two years combined, according to PitchBook. Bernstein says local Chinese suppliers such as Huawei, Cambricon and MetaX could grow from less than a fifth of China's AI-chip market in 2023 to more than nine-tenths by 2027.
Nvidia got its start in gaming but has been eclipsed by its AI business. In February 2025 it reported that quarterly revenue at its gaming division was down 11% year on year. Data-centre sales brought in $35.6bn in the same quarter, more than 90% of the total, up from just $3.6bn two years earlier. Short supply means its gaming products are frequently sold well above the recommended price. Power cables in some high-end cards have been melting during use, and in February 2025 Nvidia admitted that some cards had been sold with vital components missing.
"It is better for civilization to be going down the drain than to be coming up it."