Nvidia’s stock rose as much as 5.7% in after hours trading, after finishing the regular session up 3%.
In a conference call with investors on Wednesday, Nvidia executives stressed the underlying strength of the market, citing “visibility” into $500 billion in spending on its most advanced chips over the next 14 months, and a stunning $3 trillion to $4 trillion in annual spending industry-wide on AI infrastructure by the end of the decade.
“We are preparing for significant growth ahead,” said Nvidia CFO Colette Kress on the call.
During the third quarter, sales in the company’s datacenter unit, which accounts for the vast majority of Nvidia’s business, expanded 66% year-over-year to $51.2 billion, compared to the $49.7 billion expected by analysts. Overall revenue of $57 billion was above Nvidia’s own projections and topped the $55.5 billion expected by Wall Street.
Nvidia posted $31.9 billion of net income in the third quarter, or $1.30 per share, compared to the $1.25 EPS expected by analysts.
The company’s GPU chips are in high-demand among cloud providers and have are at the center of geopolitical issues like the U.S.-China trade war, with the U.S. having banned sales of the most advanced Blackwell chips to China. A modified H20 chip designed for China has been caught up in political uncertainty. The company said that data center chip sales to China were essentially nil in the third quarter, with no changes expected in the current quarter.
While input costs were rising, Nvidia said it expects to continue to be able to generate gross profit margins in the mid-70s percentage levels.
Huang touted his company’s longstanding role working with AI model makers, pointing to a relationship with OpenAI that dates back to 2016. “I delivered the first AI supercomputer ever made to OpenAI,” he said. He defended the billions of dollars in investments that Nvidia is making in some AI companies, pointing to the importance of ensuring that its silicon powers all the top AI models.
“We run OpenAI, we run Anthropic, we run xAI,” he said. “We run them all.”
Some of those deals have raised concerns over a “circular” business model that could overstate the true level of demand for AI products. In recent weeks, investors have been reassessing expectations, said Daniel Newman, analyst and CEO of the Futurum Group: “Has there been too much exuberance? Is this demand real?”
In response to an analyst question about the investments on Wednesday, Huang noted that the deal with Anthropic means the startup will use Nvidia chips for first time. “We’re expanding the reach of our ecosystem and we getting a share and investment in what will be a very successful company,” Huang said.
