Jensen Huang Says Nvidia Will Double AI Chip Sales Again in 2027 — Even As Investors Ask If the Boom Is Cooling
Speaking on the sidelines of a King Charles III AI summit in Scotland, Nvidia's CEO gave the market a number that cuts through months of "is the AI bubble bursting" speculation: expect Nvidia to sell twice as many chips next year as it does this year.
Here's what's behind that forecast, and why it's not just confidence talk:
→ Huang tied the doubling to broad, country-by-country demand — he says governments and companies in nearly every market where Nvidia operates want to invest in AI, not just the usual US hyperscalers
→ Nvidia just reported $96.2 billion in quarterly revenue, up 106% year-over-year, and has already guided to roughly 70% growth for the fiscal year ending January 2028 — a number that implies around $673 billion in annual sales if supply holds
→ Management has separately said unconstrained customer demand points even closer to a full doubling, with high-bandwidth memory remaining the binding constraint through fiscal 2028 — the bottleneck isn't orders, it's supply
→ Huang last disclosed shipping 6 million Blackwell GPUs over four quarters; the portfolio has since expanded to Blackwell and Rubin data-center chips, networking and optical silicon, CPUs, robotics processors, and even the chip inside Nintendo's Switch 2
→ The comments landed a day after Nvidia shares rose more than 2%, right as investors were actively weighing whether the AI infrastructure cycle is starting to cool
Huang used the same appearance to push back on calls for a slowdown in frontier-model development — arguing existing product-liability and safety laws are already enough, and that companies not confident in a system simply shouldn't ship it, rather than needing new regulation to force the issue.
Here's my read: when the most important supplier in an entire industry says the constraint is memory, not appetite, that's a very different story than "demand is slowing." Every AI startup, cloud provider, and enterprise buyer planning 2027 budgets now has to price in a supply-constrained market, not a cooling one — that changes negotiating leverage, timelines, and how much runway anyone should assume they have before getting the compute they need.
Do you think chip supply stays the bottleneck through 2027, or does something else — power, capital, or actual enterprise ROI — become the real ceiling first?
#Nvidia #AIChips #JensenHuang #TechNews #Semiconductors #AIInfrastructure
— 𝔖𝔞𝔫𝔡𝔢𝔢𝔭 ℜ𝔞𝔦𝔷𝔞