Nvidia forecasts decelerating growth after two-year AI boom

During a conference call with analysts Wednesday, the company’s leadership rejected the notion that interest in deploying AI infrastructure was flagging.

“The opportunity ahead is immense,” Chief Executive Officer Jensen Huang said. “We see $3 trillion to $4 trillion in AI infrastructure spend by the end of the decade.”

The company also approved an additional $60 billion in stock buybacks. Nvidia had $14.7 billion remaining under its previous repurchase plan at the end of the second quarter.

Sales in that period, which ended July 27, rose 56% to $46.7 billion. That compared with an average estimate of $46.2 billion. Though the gain added more than $16 billion in quarterly revenue from a year earlier, it was the smallest percentage increase in more than two years.

Second-quarter profit was $1.05 a share, minus certain items. Wall Street was looking for $1.01.

The data center unit, a division that’s now larger by itself than any other chipmaker, had sales of $41.1 billion. That compares with an average estimate of $41.3 billion. Gaming-related revenue — once Nvidia’s main source of income — was $4.29 billion. Analysts projected $3.8 billion on average. The automotive segment generated $586 million in sales, a bit shy of estimates.

The results showed hints that spending by giant data center operators “could tighten at the margins if near-term returns from AI applications remain difficult to quantify,” Emarketer analyst Jacob Bourne said in a note.

Nvidia is still dealing with the fallout from a growing US-China rivalry, where semiconductor technology has become a major flashpoint. In April, the Trump administration tightened restrictions on exports of data center processors to Chinese customers, effectively shutting Nvidia out of the market. Washington has subsequently rolled that back, saying that the US will allow some shipments in return for a 15% slice of the revenue.

At the same time, Beijing has encouraged a move away from using US technology in AI systems accessed by the Chinese government. The shifting policies have made it difficult for Wall Street to predict how much revenue Nvidia might be able to recover in the market. Some analysts have made projections in the billions of dollars, while others have refused to predict any China sales until the company makes the situation clearer.

Heading into the earnings report, Nvidia analysts had a roughly $15 billion gap between their highest and lowest estimates for third-quarter revenue — one of the largest such ranges in the history of the company.

Nvidia said it didn’t record any sales of its H20 AI chip to China-based customers in the second quarter, a decline of about $4 billion in revenue from the prior period. The third-quarter forecast excluded H20 sales as well.

Nvidia also noted that the US government hasn’t yet codified its plan to take a 15% cut of revenue from China AI chip sales. And it acknowledged risks to enacting the policy.

“Any request for a percentage of the revenue by the USG may subject us to litigation, increase our costs, and harm our competitive position and benefit competitors that are not subject to such arrangements,” Nvidia said in a filing.

Ultimately, $2 billion to $5 billion in H20 chips could be shipped to China in the current quarter, Nvidia said. The number depends on getting licenses from the US government, clearance that a “few” customers have received so far.

“If we had more orders, we can bill more,” Chief Financial Officer Colette Kress said during the conference call. She also said the company continues to urge the US government to approve a version of the more up-to-date Blackwell chip for sale in China.

“The opportunity for us to bring Blackwell to the China market is a real possibility,” Huang said. “We just have to keep advocating the sensibility of and the importance of American tech companies to be able to lead and win the AI race.”

If Nvidia was allowed to ship more capable products to the Asian nation, it would be able to take advantage of a $50 billion opportunity there, Huang said. Huge demand for AI systems in China means that market is set to grow at 50% a year, he said.

Under Huang, the 32-year-old chipmaker has suddenly become the biggest success story in the technology industry. Throughout most of its history, Nvidia lived in the shadow of larger rivals such as Intel Corp., carving out a modest living selling graphics processors to computer gamers.

Nvidia’s biggest breakthrough came when it adapted its graphics processing units, or GPUs, to run artificial intelligence software — creating something Huang calls accelerated computing.

As recently as 2022, Nvidia was a fraction of Intel’s size and booking less revenue in a year than it now generates in a quarter. These days, Nvidia is on course for annual sales of $200 billion — with the number estimated to eclipse $300 billion by 2028. That would give the company about a third of the chip industry’s total revenue.

But Nvidia is largely dependent on the spending plans of just a few companies. Microsoft Corp., Amazon.com Inc. and other giant data center operators account for about half of its sales. To diversify the business, Huang is pushing into new markets and providing a wider range of products. That includes offering complete computers, networking gear, software and services.

He’s determined to accelerate the adoption of AI across the economy, and he pushes his team to produce new hardware and software at a frenetic pace.

For now, the Santa Clara, California-based company is largely unchallenged in the market for its AI chips, known as accelerators. In-house efforts by companies such as Amazon and early-stage challenges from would-be rivals such as Advanced Micro Devices Inc. haven’t yet made a significant dent in its market share.

But the company faces other headaches. Aside from Nvidia’s struggles in China, the biggest impediment to growth has been the availability of supply. Like most chipmakers, Nvidia doesn’t own factories and relies on outsourced production, chiefly from Taiwan Semiconductor Manufacturing Co. Ramping up production of new technology remains an ongoing challenge.