[DigitalToday intern reporter Seungah Yoo] Nvidia CEO Jensen Huang (Jensen Huang) reiterated a forecast that Nvidia's revenue next year could grow 70 percent from the previous year. He said growth would continue because Nvidia is deeply embedded across the AI ecosystem even as competition in the AI semiconductor market intensifies.

On Sept. 10, TechCrunch reported that Huang showed strong confidence at the Goldman Sachs Communacopia+Technology Conference in Nvidia's dominance in AI infrastructure and demand for data centers. Nvidia forecast at the event that growth in the AI industry would continue.

Huang pointed out that some still view Nvidia as a company that sells individual graphics processing units, or GPUs. "Most people think Nvidia makes chips," he said. "To ship what we make, you need an airplane," he added.

He stressed that Nvidia's business scale has changed significantly from when it sold consumer GPUs. Huang explained that a single GPU now costs $8.5 million rather than $399, and that a system connected with NVLink contains 2,000,000 parts and uses 250,000 kilowatts of power. He added that Nvidia ships thousands of such systems at a time.

He also cited rising demand for a specific product. Huang said the product is a computer system combining 36 Grace central processing units, or CPUs, and 72 Blackwell GPUs, and that its monthly sales growth rate is now 27 percent.

Huang reaffirmed his outlook for revenue growth next year. "I think we can grow 70 percent from a year earlier. I am confident about that," he said. If market expectations that Nvidia will finish the current fiscal year with about $400 billion in revenue are applied, a 70 percent increase would put next year's revenue at about $680 billion.

He cited Nvidia's provision of core infrastructure across the AI ecosystem as the basis for his confidence. "Nvidia runs every model. Every lab can use us," Huang said, mentioning Anthropic, OpenAI and Google's models as well as open-weight models. He also stressed that Nvidia is "the foundational platform of the AI ecosystem and the foundational platform of the AI industry."

He also highlighted the breadth of information Nvidia has secured. Huang said the company's connections span from suppliers such as memory chip makers to data center projects and startups. He said Nvidia tracks sites and power availability worldwide as well as the status of data center shell construction, and receives information from neo-cloud firms, original equipment manufacturers, cloud operators and AI-native companies.

Those remarks led to a question about so-called "circular transactions," in which companies Nvidia has invested in buy Nvidia products again. "It is not circular because if we invest a little, a lot of money comes back," he replied. "If you see on a spreadsheet that we put in $1 and $100 comes back, is that circular? If so, let's do more," he joked.

Huang added that Nvidia manages risk in its investment process. He said Nvidia checks before investing whether a company has secured contracts that generate actual customers and revenue. He said he has seen such contracts total $100 billion and added, "I do not take risks. I need deals that are certain."

Competition surrounding Nvidia is also intensifying. Hyperscalers such as Amazon, Microsoft and Google are developing their own AI chips, and AI companies such as Anthropic and OpenAI are also moving to develop in-house chips. Recently listed companies such as Cerebras and Etched are also competing with Nvidia.

It remains to be seen whether Nvidia's dominance in the AI market will hold over the long term. As the AI industry matures, companies may improve the efficiency of infrastructure and token usage and may not need the same level of computing resources. Huang also acknowledged that a significant portion of current AI growth is coming from spending by AI-native startups that have raised large amounts of funding.

Even so, Nvidia maintained its outlook that growth will continue next year, citing its connected position across the industry from AI models to data centers, supply chains and customers.