Nvidia pulls $2.4 billion in Hopper sales to China out of its forecasts

Nvidia is ending deliveries of licensed Hopper data centre chips to China, despite the Trump administration having approved H200 sales last year.

Nvidia headquarters in Santa Clara, California. Photographed by user Coolcaesar on August 4, 2018.
Photo: Coolcaesar, CC BY-SA 4.0 · source This adaptation is licensed under CC BY-SA 4.0.

Nvidia pulls $2.4 billion in Hopper sales to China out of its forecasts

Nvidia is ending deliveries of licensed Hopper data centre chips to China, despite the Trump administration having approved H200 sales last year. In the third quarter, China sales amounted to $2.4 billion – the amount has now been removed from the company's guidance, without the chief executive offering any rationale.

What was actually said

On Nvidia's quarterly earnings call on Tuesday, interim chief executive Colette Kress said the company is no longer shipping licensed Hopper products to China. According to Bloomberg, she stated that the company's forward guidance "does not include any licensed Hopper data centre products, and we are no longer shipping licensed Hopper data centre products to China" (Bloomberg via Yahoo Finance).

Kress gave no reason for the decision, according to AFP. For the third quarter, she reported that "sales of licensed Hopper data centre products to China was $2.4 billion" (AFP via MSN). This is, in other words, a revenue stream that existed in the quarter just reported – and is now gone from the forward picture the company itself is drawing.

It should be emphasised: everything we know about the statements comes from the news agencies' accounts of the earnings call. No transcript, press release or SEC filing exists in the available source material.

The timeline: from approved sales to a halt

The story has a clear reversal. Nvidia had long lobbied the US government for permission to sell its advanced Blackwell chips to China. The Trump administration rejected that request, but President Donald Trump approved sales of the Hopper processors known as the H200 – chips that, according to Bloomberg, sit two generations behind Nvidia's leading Rubin family (Bloomberg). That means Nvidia is now voluntarily (or at least without a stated mandate) ending a sale that was politically approved, but which concerned older technology.

The decision also came in a hectic regulatory week. On 16 November, the Commerce Department's Bureau of Industry and Security (BIS) issued an interim final rule regulating "remote access" to advanced AI semiconductors, with publication scheduled for 19 November (Locks ltd). According to Reuters, the Commerce Department's new rules additionally target overseas data centres and smuggling routes: the restrictions cover deliveries of advanced chips to Southeast Asian countries, including Malaysia, Singapore and Thailand – countries Chinese companies have used to gain access to computing power for training AI models (reporting citing Reuters).

The announcement thus came one day after the BIS rule and alongside the Southeast Asia tightening. The chronological proximity is a fact; whether it is a causal connection is not documented.

Why did Nvidia stop? The interpretations are contested

The actual reason has not been given, and there are at least two competing explanations in the sources – both should be read as interpretations, not findings.

The first points to US regulators. Dilin Wu, research strategist at Pepperstone, told AFP: "We don't know exactly what triggered the decision to halt shipments, but it certainly looks like Nvidia is trying to draw clearer boundaries around its China business and reassure regulators that compliance comes first" (AFP).

The second points to Beijing. Chandy Ye, a solicitor in Hong Kong and mainland China, argued that the practical effect is limited because Chinese authorities have already steered the technology sector away from the H200: "Chinese authorities have already actively steered the domestic tech sector away from procuring H200s since January 2026, meaning the local ecosystem has long since adapted to this decoupling," she told AFP. If that is correct, Nvidia's halt may partly reflect that Chinese demand for the H200 was already drying up – not merely a new US adjustment.

Both explanations could in principle be true at the same time: US compliance pressure from above and Chinese government pressure from below could both make an H200 revenue stream less worth defending. But the sources provide no documentation that allows us to weigh them against each other.

The finances: the China loss carries the guidance

The most telling financial signal is what the halt did to the outlook: nothing negative. Nvidia forecast fourth-quarter revenue of $130 billion, with a variance of 3 percent in either direction. That was higher than the average analyst estimate of $125.25 billion, according to figures compiled by LSEG (reporting citing LSEG).

In other words: a quarterly sale of $2.4 billion from licensed Hopper shipments to China disappears from the guidance, and yet the company's expectations sit above market consensus. This suggests – as analysis, not as stated fact – that China revenue from this product line has become small enough relative to total revenue that the company can remove it from the accounts without weakening the growth story. That makes the decision politically and regulatorily cheaper to take.

The leadership and the background noise

The decision falls in a period when Nvidia's leadership is in flux. Harvey Jones, who was appointed the company's first chair last week, said according to Bloomberg that the board recognises it "lives in a world" where "national security dynamics are making major changes in our marketplace" (Bloomberg).

AFP describes the company as "scandal-hit", with reference to an ongoing leadership crisis. There are media reports of ageing allegations related to leadership arrangements and possible Chinese espionage links, including Beijing's denials – but these are unproven allegations, and nothing in the source material documents that they influenced the Hopper decision. They belong in the story as background, not as explanation.

What remains to be answered

Three questions remain unanswered after Tuesday's announcement:

Why? Kress gave no rationale, and no primary source documents the company's reasoning. The Wu and Ye interpretations are the clearest available hypotheses, but neither is confirmed.

Is the halt permanent? No source indicates whether this is a lasting exit or a temporary pause pending regulatory clarity. That the guidance excludes the China products does not necessarily mean the door is closed for good.

How do the new rules interact? The BIS rule on remote access to advanced AI semiconductors is fresh, and the details of its thresholds cannot be verified against primary sources in the available material. How it – and the Southeast Asia tightening – will be enforced in practice remains to be seen.

What can be established is this: Nvidia has removed an existing China revenue stream of $2.4 billion per quarter from its own guidance, in the middle of a tightening round of export controls and a leadership crisis – and did so clearly, without the market's fourth-quarter expectations being shaken.

AIMag.no
AIMag.no
The AIMag.no editorial team covers artificial intelligence, tools, research, and regulation.

Get the best of AI MAG in your inbox

News, analysis, and ideas at the intersection of AI and society.