Receive free Nvidia updates
We’ll send you a myFT Daily Digest email rounding up the latest Nvidia news every morning.
Shares in Nvidia hit a new all-time high on Thursday after the chipmaker’s revenue more than doubled in the latest quarter, as soaring demand for the processors needed to train the latest artificial intelligence models outstripped even heightened Wall Street estimates.
Reporting earnings on Wednesday after the close of New York markets, the Silicon Valley-based company also projected a bigger leap than expected in revenue in its current quarter, confirming that it was overcoming supply constraints more quickly than anticipated.
Shares in Nvidia opened 6 per cent higher at $502, capping a rally during which its stock market value more than tripled this year to more than $1.2tn. The group’s surging stock price has made it one of the biggest factors behind the AI-fuelled tech rally that has underpinned the broader US stock market this year.
At least 17 brokers raised their price targets for the stock following the earnings update. Profit-taking on the shares soon set in, and by late morning, the shares stood 1.3 per cent higher at $477.
According to published estimates, Wall Street analysts had been anticipating revenue of about $11.15bn for the latest quarter, with $12.5bn for the third quarter. But informal projections had raced ahead, with “whisper” estimates suggesting sales could hit more than $12bn in the second quarter followed by $14bn in the third.
In the event Nvidia did even better, reporting revenue of $13.51bn for the second quarter, up from $6.7bn in the same period last year, and predicting sales of $16bn in the current period, which ends in October. The third-quarter figure is close to the quarterly sales level that many analysts had not expected the company to achieve until next year.
Soaring demand for Nvidia’s GPUs, which dominate the market for training AI models, has made it the biggest winner from this year’s boom in the industry. Supply rather than demand has become the main constraint on the company’s growth in the short term.
Jensen Huang, chief executive, said the biggest cloud computing companies announced “massive infrastructures” in the latest quarter based on Nvidia’s newest AI chips, while other tech groups had struck partnerships with Nvidia to spread the latest AI technology to every industry. “The race is on to adopt generative AI,” he said.
Huang faced a barrage of questions from Wall Street on Wednesday about how durable the sales boom would be. Analysts probed whether investment in AI training would be followed by a wave of spending on AI inference, or running existing models, as well as whether software developers would come up with enough useful applications for generative AI to justify the heavy investment in new computing power that is currently under way.
The Nvidia CEO brushed off concerns, repeating recent claims that the leap in demand for his company’s chips represented a secular shift in data centre spending towards “accelerated computing and generative AI”.
“We have excellent visibility through the year and into next year,” he said. “I think this is not a near-term thing, this is a long-term industry transition. We’re seeing these two platform shifts happen at the same time.”
The second-quarter revenue surge — in which Nvidia’s sales to data centre customers jumped to $10.3bn, about $6bn higher than the previous three months — came despite a move by some of the biggest cloud computing companies to moderate their capital spending.
Huang said cloud companies were increasingly “diverting their capital investment” into his company’s chips and away from “general-purpose computing” as they expand their AI capabilities, suggesting that Nvidia’s success would come at the expense of chipmakers such as Intel.
The latest sales jump also came in the face of US regulations designed to prevent exports of the most advanced AI chips to China, one of Nvidia’s biggest markets. Nvidia has introduced less powerful versions of its top-of-the-line AI chips for China to comply with the rules.
Colette Kress, chief financial officer, said sales to Chinese customers accounted for about 20 per cent to 25 per cent of Nvidia’s data centre revenue in the latest quarter. She defended the continued high level of exports to China despite US efforts to slow that country’s AI advance, and said the company did not expect its success there to lead to any tightening of the US export controls. “We believe the current regulation is achieving the intended results,” she said.
The jump in sales lifted Nvidia’s after-tax profits to nearly $6.2bn, up from just over $2bn the year before. Pro forma earnings per share rose to $2.70, compared with Wall Street expectations of $2.02.
Additional reporting by Jennifer Hughes in New York