NVIDIA CFO reveals $5 billion H20 revenue in China and confirms massive demand for next-generation Vera Rubin architecture.

At the Goldman Sachs Communacopia + Technology Conference, NVIDIA CFO Colette Kress attributed geopolitical tensions between the US and China as the primary obstacle to recognizing revenue from H20 AI GPU sales to the Asian market. Kress confirmed that NVIDIA has received H20 licenses from the Trump administration, and stated that H20 sales "could account for as much as $5 billion in revenue from H20 sales in the third quarter if tensions smooth out between the world's largest economies." Despite this headwind, NVIDIA's data center and networking revenues, excluding China-specific H20 GPUs, achieved 12% sequential growth in the second quarter. Looking ahead, Kress outlined that "what we're targeting right now for Q3 in our outlook is a 17% growth sequentially," signaling accelerating demand for the company's products.

NVIDIA has scaled production of its next-generation accelerators more rapidly than market expectations. According to Kress, the firm was "also at scale with our GB300 Ultra," noting that there was "a lot of discussion that said, I didn't know that would actually be a big part." She described the transition as seamless and emphasized that many observers underestimated "the amount of scale and volume we were actually able to put into market." Kress indicated that NVIDIA plans to further scale up GB200 and GB300 racks in the third quarter, aligning with reports suggesting that NVIDIA might experience a 300% sequential growth with GB300 in Q3.

Regarding the H20 situation in China, Kress confirmed that NVIDIA has "received licenses for several of our key customers in China" and wants the opportunity to "complete the shipments and ship the H20 architecture to its Chinese customers." However, she acknowledged that "there is still, this position right now, a little geopolitical, situation that we need work through between the two governments," with Chinese customers seeking assurance that the government supports receiving the H20 chips. While reports have indicated the Chinese government is urging local businesses to adopt domestic chips instead of NVIDIA's products, Kress asserted that the firm believes "there's a strong possibility that this [the China H20 shipments] will occur," estimating that the additional revenue from such shipments "could range between $2 billion to $5 billion."

Kress defended NVIDIA's emphasis on power efficiency amid industry debate over cost-effective chip alternatives following Broadcom's announcement of a $10 billion custom AI chip contract. She emphasized that "power is indispensable when it comes to AI computing," noting that "individuals generating more tokens and inferencing demands for reasoning models account for a large portion of reasoning model demand." Kress stressed that when building large AI systems, companies must consider power consumption over the system's lifetime—"four, six years or even further"—making it essential to balance "the most performant per watt and the most performant per dollars."

Regarding NVIDIA's next-generation Vera Rubin AI chips, Kress announced that "Rubin is on a path and that one-year cadence is going to be a journey we're ready to take on with Rubin. So Vera Rubin, six chips, all of them taped out." She revealed that the company has already secured significant customer demand, having "already had discussions to where we probably will see several gigawatts needs for our Vera Rubin" and has "pencilled that in so even way before it's even ready to go to market, we already are seeing gigawatts worth of needs as we go forward."