Nvidia has visibility into $500 billion in AI chip revenue through the end of 2026, said Colette Kress, CFO of Nvidia, in an analyst call.
“We currently have visibility to half a trillion dollars in Blackwell and Rubin revenue from the start of this year through the end of calendar year 2026,” she said in an analyst call today.
While all industry is transforming, it’s just the start.
“Yet, we are still in the early innings of these transitions that will impact our work,” she said.
OpenAI recently shared that its weekly user base is 800 million people. This past quarter, Nvidia announced AI factory and infrastructure projects amounting to an aggregate of five million GPUs, she said.
“Their earnings reinforce the never ending pursuit of more AI compute,” said Patrick Moorhead, chief analyst a Moor Insights & Strategy, in a message to GamesBeat. “There were doubts as the stock and those tied to it declined. What people need to incorporate are that the hyperscalers increased their CAPEX by 40%, Jensen announced $500B of revenue over 5-6 quarters at GTC DC, and that growth is primarily for new frontier models. What hasn’t been incorporated is the growth from enterprise and industrial AI in the mid-term. “
Jensen Huang, CEO of Nvidia, said in the analyst call, “There’s been a lot of talk about an AI bubble. From our vantage point, we don’t see it.”
He added, “Nvidia is unlike any other accelerator. We excel at every phase of AI, from pre-training and post-training to inference. And with our two-decade investment in CUDA acceleration libraries, we are also exceptional at science and engineering simulations, computer graphics, structured data processing, to classical machine learning.”
He noted Nvidia is addressing three big transformations: from CPU general-purpose computing to GPU computing as Moore’s Law slows. That’s a huge opportunity, with CPUs shifting to CUDA GPUs.
“Accelerated computing is reaching a tipping point,” he said.
On top of that AI is reaching a tipping point, with generative AI overtaking everything from search to generating images.
The new wave is Agentic AI, from Claude code to AI chaffeurs like Waymo. He said the fastest-growing companies in the world are pioneering agentic AI. Each of these fundamental changes is driving demand for AI chips Huang said.
But not all was ideal. Kress said sales of the H20, a stripped down model targeted at China, were approximately $50 million. That’s a small amount. She said sales didn’t happen in the quarter due to geopolitical issues and the competition in China.
“While we were disappointed in the current state that prevents us from shipping more competitive data center compute products to China, we are committed to continue engagement with the US and China governments and will continue to advocate for America’s ability to compete around the world,” Kress said.
She added, “To establish a sustainable leadership position in AI computing, America must win the support of every developer and be the platform of choice for every commercial business, including those in China.”
Huang said Nvidia has done a very good job planning its supply chain, which includes every major tech company in the world. He said the company has been planning for a good year.
“We are seeing all of these exponentials running at the same time,” he said.
One analyst asked how much vendor financing is playing a role in Nvidia sales. Since there are worries of a bubble, there is a lot of risk — reminiscent of the dotcom bubble — in loaning money to customers so they can place orders for products.
“Our customers’ financing is up to them,” Huang said. “We see growth for quite some time.”
He said multiple companies and most of industries and most countries are going “engage with generative AI.”
He also noted Nvidia has a very strong balance sheet.