Nvidia CEO Jensen Huang said at the Goldman Sachs Communacopia + Technology conference on September 10, 2026, that the company’s growth in artificial intelligence would continue. Huang highlighted the difference between Nvidia’s current products and the GPUs the company previously sold to consumers to enhance PC gaming. What is involved today is not a graphics card on its own, but a complex GPU system consisting of 2 million components and 250,000 kilowatts, connected to one another via NVLink. Huang said one of these systems is worth 8.5 million dollars and that Nvidia has shipped thousands of them; this does not mean that the price of the 399-dollar consumer GPU of the past has risen to 8.5 million dollars.
Huang also announced that sales of the computer system combining 36 Grace CPU and 72 Blackwell GPU have increased by 27% per month. Reiterating his belief that Nvidia could increase its revenue by 70% year over year next year, Huang said current fiscal-year revenue is expected to be approximately 400 billion dollars, while next year’s revenue is expected to be approximately 680 billion dollars.
The CEO argued that Nvidia is a foundational platform used in the models of all artificial intelligence laboratories, including Anthropic, OpenAI and Google. He said the company monitors many parts of the ecosystem, from memory chip manufacturers to data center projects, cloud services and artificial intelligence startups. Responding to criticism of “circular deals” involving companies in which Nvidia invests later purchasing Nvidia products, Huang said they require these companies to have genuine contracts generating revenue from customers before making investments. He put the total value of such contracts at 100 billion dollars.
Huang also acknowledged that artificial intelligence startups are driving growth, but said efficiency in infrastructure and token usage will increase as the industry matures. Nevertheless, Nvidia’s competitors include Amazon, Microsoft, Google, Anthropic, OpenAI, Cerebras and Etched.
Why it matters
For Nvidia, this picture shows that the company’s value is built less around individual hardware sales than around integrated infrastructure that connects processors and extends to data center projects. This means the development concerns not only artificial intelligence laboratories using Nvidia products, but also those assessing memory manufacturers, cloud services and startups within the same commercial ecosystem. Criticism that companies in which Nvidia invests later purchase products from Nvidia leaves open the question of how much growth is based on independent demand, despite the existence of customer contracts. Huang’s acknowledgment of efficiency gains, along with the mention of numerous competitors, also indicates that high revenue expectations should be considered together with issues such as the sustainability of infrastructure spending and the intensity of competition.
Background
Nvidia is not a new name in the FikirPilot archive: we have published 17 reports mentioning the name in the last 90 days; the most recent was dated 6 September 2026.