Nvidia sold $2.4B of Hopper chips to China, then guided zero
Interim CEO Colette Kress says the licensed Hopper revenue is out of guidance entirely, and the question every board now faces is what replaces it.
Nvidia interim CEO Colette Kress said the company sold $2.4 billion of licensed Hopper chips to China last quarter and has none of that business in its guidance. For decision-makers, the takeaway is that export policy can convert booked China revenue into a guided zero, forcing boards to stress-test China exposure as a standalone line item rather than a blended growth story.
Nvidia sold $2.4 billion of licensed Hopper chips into China last quarter, and it is guiding investors to expect nothing from that business going forward. Interim CEO Colette Kress put both numbers on the table: a real, recent revenue line, and a forward view that assumes it disappears. That is the story in two data points, and it is the kind of disclosure that forces every board with China exposure to re-run its own math. For a company whose chips sit at the center of the AI buildout, that is not a rounding error.
The gap between those two figures is where the tension lives. The $2.4 billion is trailing revenue, money already booked. Zero in guidance is a forecast, a statement about what management believes it can still sell. When a company stops shipping a product line into a market and simultaneously removes it from the outlook, it is signaling that the change is not a timing issue. It is a structural one. Kress, running Nvidia as interim chief executive following Jensen Huang's exit, delivered that message without dressing it up.
Hopper is the architecture behind Nvidia's data center accelerators, the chips that train and run large AI models. For context, Washington has spent years tightening controls on exports of advanced AI silicon to China, and Nvidia has historically responded by designing parts that fit inside the rules rather than walking away from the market. That approach kept revenue flowing. Licenses are the mechanism: chips above certain performance thresholds cannot ship without authorization, and that authorization can be narrowed or pulled. The $2.4 billion Kress cited is the residue of that strategy, export-controlled chips that moved under license.
Now the residue is gone. Guidance is the number Wall Street actually underwrites, and a zero for China means every model built on Nvidia's forward revenue has to stand without it. Historically, China has been a meaningful slice of the company's data center business, and the strategy of designing parts that fit inside export rules was aimed at keeping a foot in the door. Kress's disclosure closes that door in the forecast, at least for the licensed Hopper line.
There is a governance layer here that boards should not skip past. A company guiding a major revenue line to zero is a market-moving event under any circumstances. Doing it with an interim CEO in the chair adds a second question: who owns the strategy that replaces the lost revenue, and on what timeline? Interim leaders can make hard calls, but they cannot make durable ones alone. Investors will want to know whether the China exit is a policy-driven pause or a permanent reallocation of Nvidia's capacity toward other buyers.
The second-order effects run wider than one company. Every chipmaker with China revenue, every hyperscaler buying accelerators, and every supplier in the AI supply chain now has a fresh data point about how quickly export policy can convert booked business into guided zero. The practical lesson is that China revenue is becoming a line item boards should stress-test separately rather than blend into a consolidated growth story. Companies that have already segmented it will look prescient. Companies that have not will be asked why.
The offsetting argument is demand everywhere else. AI infrastructure spending outside China has been the dominant driver of accelerator sales, and Nvidia's manufacturing capacity is finite. If the same chips that would have gone to China can be sold into other markets, the revenue hit is smaller than the headline number suggests. But that is an assumption, not a disclosure. What the disclosure does not include is an offsetting figure for the capacity freed up. Kress gave investors a number for what is leaving and no number for what replaces it.
That is the open question executives should carry into their next planning cycle: when a $2.4 billion quarterly line goes to zero in guidance, what is the replacement, and how confident is management in it? Watch the next set of results for two things. First, whether China reappears in guidance in any form. Second, whether the interim leadership structure is resolved, because the person who owns the answer will be the person setting Nvidia's strategy for the next several years. That answer will define the next phase of the AI trade.
This story's Key Insights and Take-aways are locked.
Create a free account to unlock Executive Actions for one credit.
Register to UnlockAlways free for Executives Club members. Join the Club
More in Business
S&P 500 Enters Correction, Down 10% From Its High on AI Worries
The index's slide past the 10 percent line is a symbolic verdict on investor mood, and the AI trade that powered the rally is now the thing testing it.
Dow and S&P 500 enter correction as Nasdaq logs worst week since 2020
All 11 S&P 500 sectors fell more than 1% in the same week for the first time since April 2025, a breadth signal that matters more than any single index level.
Chip stocks lose $2.3 trillion in a week, worst run in 25 years
A $2.3 trillion weekly wipeout in semiconductors is forcing every board with an AI story to ask whether it owns a thesis or just a crowded trade.



