Nvidia CEO Jensen Huang says he's found a 'brand new' $200B market for the company
Nvidia CEO Jensen Huang, in a keynote speech at the Computex 2026 conference in Taiwan, announced a 'brand new' $200 billion market opportunity ahead of the company. Huang spoke of this new market as 'physical AI' -- meaning robotics, autonomous vehicles and industrial-automation responses. According to TechCrunch, Huang stated in his presentation that Nvidia's $250 billion annual data-centre revenue could be matched by an independent market structure of comparable scale.
Huang's concept of 'physical AI' refers to the capacity of AI systems not only to produce digital text and images but to act on the physical world -- through robots, vehicle drivetrains and factory equipment. The vision is not new; Nvidia has been entering this market since 2024 with its Isaac robotics platform and Drive autonomous-vehicle stack. But what Huang did at Computex 2026 was to encode these efforts as a separate 'market segment' and present them to investors as a growth axis.
In his speech, Huang announced two new product launches by Nvidia. The first product is Cosmos, a humanoid-robot foundation model -- a 14-billion-parameter AI model developed by Nvidia. Cosmos is designed to give robot companies a base intelligence capacity 'like an operating system'; developer companies will be able to build their own applications on top of Cosmos. The second product is Drive Thor, a next-generation autonomous-vehicle chip -- offered to top-tier auto-makers with 1,000 TOPS processing capacity.
Nvidia's real competition in this market is complex. On one side is the startup ecosystem (humanoid-robot-focused entities such as Figure AI, Skild AI, Physical Intelligence), on another side autonomous-vehicle makers (Tesla, Waymo) and industrial giants (Siemens, ABB, Fanuc). Huang positioned Nvidia's strategy as 'the neutral infrastructure provider that supplies its chip to every player'; this is a strategy in parallel with Nvidia's position in the cloud market.
Huang gave two data points as the basis of the assumption of a $200 billion 'physical AI' market: (1) the global robotics market is projected to reach $250-280 billion by 2030, and (2) the autonomous-vehicle sector exceeds $100 billion per year. Nvidia's target of capturing 30-40% chip share in each of these markets points directly to an annual revenue potential of $80-110 billion. That is roughly one-third of current data-centre revenue -- the mathematical basis for how the phrase 'brand new' should be interpreted.
Financial analysts reacted largely positively to Huang's announcement. Goldman Sachs analyst Toshiya Hari published a client note: 'Nvidia's expansion into the physical AI market should be evaluated alongside the company's maturation in the AI data-centre market. This will help maintain the share-price multiple at a high level.' The share price rose 3.2% after Huang's keynote.
Among direct competitors, AMD has yet to make a direct entry into the physical AI segment; the company is focused on taking share in the AI data-centre market with Intel. Tesla is continuing its Optimus humanoid-robot programme -- but Tesla's strategy lies on the side of developing its own chips. Intel has not demonstrated proven capability in robotic chips; it lags clearly behind Nvidia in the data-centre side. Chinese rivals Huawei and Cambricon have limited direct market access due to US export restrictions.
The regulatory framework is an important unknown in the development of physical AI. In the United States, the Federal Trade Commission has begun reviewing safety standards for autonomous-vehicle AIs; under the European Union's AI Act, AI systems in the 'high risk' category must pass a mandatory conformity assessment. Nvidia's chips are mostly outside regulation as a 'fundamental component', but the AI Act's 'foundation model' section still has an indirect impact on the end products in which the chips are sold.
Another notable subject of Huang's keynote was Nvidia's investment in India. The company announced it will open a $2 billion physical-AI R&D centre in Bangalore; the centre is planned to employ 1,500 engineers by the start of 2027. This represents an expansion of Nvidia's global engineering capacity beyond Taiwan, the United States and Israel; the preference for India specifically for software-automation integration suggests that physical AI will develop at the industrial-software intersection.
This article is in the nature of commercial news and corporate strategy analysis; it does not constitute or substitute for individual investment advice. Nvidia's new strategic announcements may create share-price volatility; investment decisions should be made together with licensed financial advice.