In what has become the most highly anticipated report of the tech earnings season, Nvidia Corp. couldn’t quite make investors happy.

The maker of graphics processing units that power the latest artificial intelligence boom posted revenue of $46.7 billion for the 2026 fiscal second quarter, up 56% year-over-year and ahead of the consensus analyst estimate of $46.1 billion.

Earnings per diluted share of $1.05 also beat estimates of $1.01. However, the company slightly missed its data center sales forecasts, sending the stock down more than 3% in early after-hours trading.

The Data Center division generated sales of $41.1 billion, up 56% from $26.2 billion a year earlier, but just short of the $41.3 billion analysts had expected. The company’s newest AI offering, the Blackwell Data Center platform, grew 17% sequentially over the previous quarter. Chief Executive Jensen Huang (pictured) said production of the Blackwell Ultra chip line is proceeding at full throttle and “demand is extraordinary.”

“They’re going to get criticism for missing the data center number, but I don’t think a slight miss on a huge number like that indicates any weakness in their business,” said Scott Bickley, an advisory fellow at Info-Tech Research Group Inc.

New product momentum

AI-optimized rack-based systems, which are a major product expansion area for Nvidia, are seeing widespread adoption, said Chief Financial Officer Colette Kress. “The current run rate is back at full speed, producing approximately 1,000 racks per week,” she said. “This output is expected to accelerate even further throughout the third quarter as additional capacity comes online.”

Another growth category, networking, delivered record revenue of $7.3 billion, up 46% sequentially and 98% from a year ago. Automotive revenue jumped 69% from a year ago, to $586 million, driven by strong adoption of new self-driving platforms.

“The company is making strategic moves across areas under their control, including continued development of state-of-the-art data center GPUs, innovation in new growth areas like robotics and physical AI, and promoting use cases that drive increased GPU purchases, such as reasoning AI and rack-scale computing,” said Forrester Research Inc. Senior Analyst Alvin Nguyen

The company’s gross margin of 72.7% was up sharply from 61% a year ago but down sequentially from 75.7% last quarter. Gaming revenue hit a record $4.3 billion, up 14% over last quarter increase and 49% over a year ago.

Nvidia forecast third-quarter revenue of around $54 billion, exceeding Wall Street expectations. The figure included no shipments to China.

Sales to China continue to hang over Wall Street’s expectations. Although the company booked no sales of its H20 chip to Chinese customers during the quarter because of government restrictions, it released $180 million worth of H20 inventory for sale outside China.

The H20 is a specialized, less powerful version of the company’s advanced AI chips, designed specifically to navigate U.S. export controls. Despite its lower power compared to the company’s top-tier chips, the H20’s utility in inference and its inclusion of high-bandwidth memory are important to Chinese tech firms as they develop their own AI capabilities.

Government stalemate

“U.S. government officials have expressed an expectation that the government will receive 15% of the revenue generated from licensed H20 sales, but to date, the USG has not published a regulation codifying such requirement,” Kress said. “We have not included H20 in our Q3 outlook as we continue to work through geopolitical issues.”

Executives said export restrictions weigh on revenues. “If geopolitical tensions subside, we anticipate $2 billion to $5 billion in H20 revenue in Q3,” Kress said.

Nvidia executives and industry watchers said Nvidia’s long-term outlook continues to dazzle. “Judging by the litany of partnerships that they listed on the press release, you can see them expanding into customized use cases and moving into other geographies outside of China,” said Info-Tech’s Bickley. “It seems like they’re securing their hold in this space.”

Huang said the transition to agentic AI and reasoning systems could be a major growth driver going forward.

“Where chat bots used to be one shot, AI now does research, thinks and creates a plan. It’s called long thinking,” he said. “The computation necessary for reasoning agentic AI models could be a hundred or a thousand times that of one-shot models. The amount of computation that has resulted from agentic AI has grown tremendously.”

Too big?

Although Nvidia faces no immediate headwinds, its $4.4 trillion valuation and reliance on a few large hyperscaler cloud providers has contributed to market jitters. With many people in the industry warning that AI is in a bubble, the question is how big the impact on Nvidia will be if and when it bursts.

“Its stock is 10% of the value of the Nasdaq,” Bickley said. “Once you get the numbers that big, at some point, they’re going to tip over.”

But Huang said the long-term opportunity remains huge. “Over the next five years are going to scale into effectively a $3 trillion to $4 trillion AI infrastructure opportunity,” he said, noting that capital expenditures by the top four cloud service providers alone have grown by about $600 billion over the last two years.

“We are in the beginning of this buildout in the technology advances have enabled AI to adapt and solve problems to many different industries,” he said.

Kevin Cook, senior stock strategist at Zacks Investment Research Inc., noted that Huang had characterized the opportunity at $1 trillion just a year ago.

“Jensen said ‘The AI race is on, and Blackwell is the platform at its center,'” he said. “Nvidia is not in a race with anyone because they own 90% market share.”

Photo: Robert Hof/SiliconANGLE Media