(Bloomberg) — Nvidia Corp. has assured investors that its new product line will continue to fuel artificial intelligence-led growth, while also signaling that the rush to release chips is proving more costly than expected.
Most read on Bloomberg
Listen to the Bloomberg Daybreak Europe podcast on Apple, Spotify or wherever you listen.
Speaking after releasing quarterly results, CEO Jensen Huang said Nvidia’s highly anticipated Blackwell products would ship this quarter amid “very strong” demand. But chip production and engineering costs will weigh on profit margins, and Nvidia’s sales forecast for the current period doesn’t match some of Wall Street’s most optimistic projections.
That sparked a lukewarm reaction from investors, who had bid up Nvidia’s shares nearly 200% this year before the earnings report was released. After this dizzying rebound, which made the microchip maker the most valuable company in the world, anything but an explosive quarter was bound to be a disappointment.
Shares fell as much as 3.6% on Thursday before rebounding in the afternoon. They closed up 0.5% at $146.67.
Nvidia forecast fourth-quarter revenue of around $37.5 billion. While the average analyst estimate was $37.1 billion, projections ranged as high as $41 billion.
“Forecasts seem to show weaker growth, but that could be because Nvidia is conservative,” said Alvin Nguyen, an analyst at Forrester Research Inc. “In the short term, there is no concern about demand for AI. Nvidia is doing everything it should be doing.
The company’s biggest source of revenue is its accelerator chip, which helps develop artificial intelligence models by bombarding them with data. Since the launch of OpenAI’s ChatGPT chatbot in 2022, a frenzy of AI services has created insatiable demand for the product.
All About Nvidia Chips, AI Hype, and What’s Next: QuickTake
Wall Street is closely watching the launch of Blackwell, the latest in this category, which is faster and has an improved ability to interface with other semiconductors. Manufacturing issues have slowed the rollout, and Nvidia warned again Wednesday of supply constraints. Demand for these products is expected to exceed supply for several quarters.
“Critical questions regarding Blackwell’s production ramp and customer concentration remain major concerns,” Jacob Bourne, an analyst at Emarketer, said in a note. “There is little room for missteps in execution in 2025.”
Huang said Blackwell was now in “full production” and there was still an appetite for Hopper, the previous model. “Blackwell is now in the hands of all of our major partners,” he said on the conference call.
But the move to Blackwell took its toll on profitability. The company’s gross margin, which measures the percentage of sales remaining after deducting the cost of production, will fall to 73% this quarter from 75% in the previous period. This figure is expected to rebound when new products are produced on a larger scale and the economic situation is more favorable.
When asked if Nvidia’s gross margin could return to the mid-70s by the middle of next year, Chief Financial Officer Colette Kress said that was a reasonable assumption. Nvidia remains well above its peers in this category: its closest rival, Advanced Micro Devices Inc., has a gross margin 20 percentage points lower. That of Intel Corp. doesn’t even represent half of Nvidia’s total.
Nvidia’s growth over the past two years has been meteoric. Its sales are on pace to double for the second year in a row, and the company is now making more profit than it previously generated in terms of total revenue.
Nvidia’s revenue increased 94% to $35.1 billion in the fiscal third quarter, ended October 27. Excluding certain items, profit was 81 cents per share. Analysts had forecast revenue of about $33.25 billion and earnings of 74 cents per share.
Nvidia’s largest division, the data center unit, saw revenue double from the previous year to $30.8 billion. That beat Wall Street estimates.
But networking revenue within that unit has declined sequentially, and the company is more dependent than ever on a small group of customers: cloud service providers. This cohort, which includes companies such as Microsoft Corp. and Amazon.com Inc.’s AWS, accounted for 50% of data center revenue, up from 45% in the previous period.
Investors want this figure to fall, to show that the use of AI is spreading throughout the economy.
Other recent earnings reports have given strong signals in favor of AI. Nvidia’s customers, including Microsoft, Amazon and Meta Platforms Inc., have reaffirmed their commitment to spending heavily on AI infrastructure.
WATCH: The Surprising Origin Story of Nvidia’s AI
Nvidia has only missed analysts’ estimates for quarterly revenue once in the last five years. And it has exceeded expectations by as much as 20% in recent periods, setting the bar high for its performance.
Its data center division alone now generates more revenue than rivals Intel and AMD combined. This year’s net profit is on pace to surpass revenue at Intel, a company that has been the chip industry’s largest company for decades.
Nvidia made its name selling graphics processors, but discovered that the technology also had applications for AI. Its chips help software models during the training process, when they learn to recognize and respond to real-world input. Nvidia’s components are also used in systems that then run the software, a step known as inference, and help power services like ChatGPT.
The Santa Clara, Calif.-based company has rapidly expanded its product line to include networking, software and services, and fully integrated computing systems. Huang travels the world to lobby for wider adoption of his technology and attempts to expand its use to businesses and government agencies.
“The AI era is upon us and it is vast and diverse,” Huang said.
