Jensen Huang, founder, CEO, and NVIDIA’s tireless publicist, told attendees at the Goldman Sachs Communicope+ Technology Conference on Thursday why the company’s AI advantages and revenue will continue to grow at record rates through the end of next year.
He can see the future.
Whether Nvidia’s party is over is an endless debate as it faces increasing competition for GPUs and AI chips from all directions, including hyperscalers (Amazon, Microsoft, and Google, each developing their own) and AI labs (Anthropic and OpenAI, each developing their own), as well as startups such as newly public competitors Cerebras and Etched.
“Most people think that Nvidia makes chips, which means we need planes to ship what we make,” Huang said, adding that the company continues to fight that perception from its early days. Nvidia invented GPUs, which at the time were primarily sold to consumers to improve PC gaming. “A GPU today is not $399; it’s $8.5 million. It’s one GPU, all connected to NVLink, and it has 2 million parts. It’s 250,000 kilowatts. This is one GPU, and we’re shipping thousands of them.”
He added that orders for just one product, a computer system that combines 36 Grace CPUs and 72 Blackwell GPUs, are currently experiencing 27% month-over-month sales growth.
However, Huang’s bullish view was not limited to current sales figures. He used the opportunity to reiterate NVIDIA’s earnings outlook for next year. That’s the guidance the company provided last month when it reported yet another record quarter of earnings. It was then that he first said that next year’s sales could increase by 70%.
“I think we can grow 70% year-on-year. We are confident in that,” Huang said again on Thursday. Analysts expect the company to end the current fiscal year with sales of about $400 billion. So 70% growth means about $680 billion next year.
Huang explained why he is confident. Because his company is deeply embedded in all areas of AI, he believes he can see the future.
“Nvidia runs every model. Every lab can use us,” the CEO said, noting that this includes models from Anthropic, OpenAI, and Google, as well as promiscuous products. “We are a foundational platform for the AI ecosystem and a foundational platform for the AI industry.”
Nvidia has branched out into everything from suppliers such as memory chip makers to data center projects and startups.
“We track every gigawatt of land, power and shells in the world. Literally everything on the planet,” he said. (“Shell” refers to the shell of the data center building before it is equipped with computers).
“So, think about all your partners. How many neoclouds are reporting to us? How many OEMs are reporting to us? How many clouds are reporting to us? How many AI native companies are reporting to us? We work with all the companies, so we kind of know where everything is,” he said.
The comment led to inevitable questions about NVIDIA’s so-called circular trading, which invests in companies that buy its products. Plans like this famously contributed to the downfall of previous generations of Internet building suppliers like Lucent Technologies.
Mr. Hwang had a simple, if cheeky, response. “Well, you put in a little bit of money and you get a lot of money back, so it’s not cyclical,” he continued, joking, “If you look at the spreadsheet, you put $1 in and you get $100 back. Is that cyclical? If it’s cyclical, let’s do more.”
Jokes aside, he claimed that before these companies receive investment, Nvidia makes sure they have actual contracts that generate revenue from customers. Overall, he said he has seen $100 billion worth of such contracts and said, “I’m not taking any risks. … I need certainty.”
Only time will tell whether Nvidia’s AI stronghold is truly sustainable long-term. If there’s a golden rule in the tech industry, it’s that everything important is interrupted. For now, even Huang admits that much of the growth in AI is coming from AI-native startups that are raising huge amounts of funding and spending most of that cash on their own AI uses. As the AI industry matures, we expect companies to become more efficient in how they use their infrastructure and tokens.
But for now, NVIDIA is focusing on all areas, and we think we’ll have another bumper year ahead.
If you make a purchase through links in our articles, we may earn a small commission. This does not affect editorial independence.
