Nvidia CEO Jensen Huang says the artificial intelligence industry is moving from years of research into large-scale production. The shift comes as investors revisit whether AI valuations and capital spending have run too far: Nvidia raised its quarterly revenue guidance last month to $108 billion, up from $96.2 billion previously. Huang says the change reflects continued growth in demand for AI infrastructure, but investors still need more earnings data to determine whether that spending can translate into sustainable revenue and profit.
Huang says AI has entered a high-output phase
In an extended interview released on September 20, Huang said much of the AI industry’s work over the past 10 to 15 years had focused on research. This year, he said, the technology has increasingly been converted into usable, commercial products. Companies including OpenAI and Anthropic are moving from laboratory operations toward product businesses while expanding their computing capacity at a rapid pace.
Huang said the industry had effectively entered a “high-output phase” over roughly the previous six months. In his view, AI products are no longer merely technology demonstrations; some are already generating revenue. That development is a key reason Nvidia continues to maintain elevated revenue expectations.
Huang also rejected predictions that AI could cause human extinction before 2030, saying such claims lack a scientific basis. His comments did not settle broader market concerns about the speed of AI development, the cost of the required infrastructure or the timing of commercial returns.
Chip stocks face pressure as investors seek returns
Markets have not fully accepted the view that AI spending will continue to expand at its current pace. Nvidia shares fell 3.4% on September 14. Earlier comments from Anthropic CEO Dario Amodei, OpenAI CEO Sam Altman and Elon Musk expressing differing views on the speed of AI development or the risks facing the industry added to volatility in chip stocks and revived debate over a possible AI bubble.
Investor Michael Burry has also increased his short position in Nvidia, arguing that the company’s valuation is too high. At the same time, opposition to data-center construction has grown in some areas, with disputes focusing on power consumption, pressure on local infrastructure and the impact of projects on surrounding communities. Huang acknowledged that the AI industry has not handled these issues well.
Nvidia’s valuation is increasingly tied to compute demand
Nvidia currently has a market capitalization of about $5.3 trillion, making it one of the world’s largest listed companies by value. That scale means Huang’s assessment of AI spending matters not only to Nvidia’s orders and profits, but also to how investors value the wider AI infrastructure chain.
The central market question is no longer simply whether AI products have practical uses. Investors are also assessing how much capital companies must commit, when that spending can produce stable revenue, and whether cloud providers and model developers can sustain their current pace of compute purchases. OpenAI’s and Anthropic’s plans to expand computing resources support demand for Nvidia’s products, while also focusing attention on customers’ sources of funding and progress toward commercialization.
Nvidia’s next earnings report will provide an important test of whether the high-output phase is still accelerating. Quarterly revenue, growth in the data-center business and management’s comments on future orders may offer a clearer indication than broad industry assessments of whether AI compute demand is becoming a durable source of corporate revenue.