Nvidia Data Center Sales Soar As Hyperscalers Adopt A100 GPU

Nvidia says the new 7-nanometer A100 data center GPU contributed 'meaningful' revenue in its first quarter, thanks to 'strong adoption' across leading hyperscalers. 'We think that's a true indication of their excitement about our platform and their excitement for A100,' CFO Collette Kress says.

Nvidia said sales of its data center GPUs soared 80 percent in the first quarter, thanks in part to "strong adoption" of the chipmaker's new A100 GPU across "leading hyperscalers."

The Santa Clara, Calif.-based company said its data center revenue for the first quarter of its 2021 fiscal year topped $1 billion for the first time, reaching $1.1 billion, an 18 percent increase from the previous quarter, driven by "broad-based demand across hyperscale and vertical industries."

[Related: Nvidia's 5 Biggest GTC 2020 Announcements: From A100 To SmartNICs]

That boost in data center revenue contributed to $3.08 billion in revenue for the first quarter, a 39 percent increase over the same period last year and $80 million higher than what Wall Street analysts were expecting. The company's earnings per share of $1.80 also beat expectations, by 12 cents.

Nvidia's stock price was nearly flat in after-hours trading to $351.01 per share.

The company said its first-quarter sales did not factor in its recently closed acquisition of Mellanox Technologies, whose first-quarter revenue was $429 million, a 40 percent increase over the same period last year. Mellanox's sales will begin factoring into Nvidia's revenue starting in the second quarter.

Jensen Huang, Nvidia's CEO and founder, said with the acquisition, Mellanox's high-speed interconnects give Nvidia end-to-end capabilities to drive "data center-scale computing."

"A CPU compute node was the unit of computing," he said on the first-quarter earnings call. "Going forward, the new unit of computing is an entire data center."

Collette Kress, Nvidia's CFO, said the A100, the company's first GPU based on its new 7-nanometer Ampere architecture, contributed "meaningful" revenue in the first quarter thanks in part to adoption by cloud service providers like Alibaba Cloud and Microsoft Azure.

The GPU powerhouse debuted the new A100 last week, saying the GPU unifies training and inference acceleration into one architecture that can outperform its V100 and T4 several times over. The new GPU also comes with the ability to be portioned into as many as seven distinct GPU instances.

The company's forecast for the second quarter is $3.65 billion, plus or minus 2 percent, which would represent a roughly 41 percent increase over the same period last year.

Kress said Nvidia continues to have "solid visibility" into the second quarter for data center demand, representing the company's continued confidence for the business.

"We think that's a true indication of their excitement about our platform and their excitement for A100," she said.

Nvidia's gaming segment saw a first-quarter revenue of $1.34 billion, a 27 percent year-over-year increase. Professional visualization was $307 million, a 15 percent year-over-year increase. Automotive, on the other hand, was $155 million, a 7 percent year-over-year decrease.