The Honeypot That Exposed Nvidia's China Chip Problem

Good Revenue does not usually traffic in tabloid material, but this one is unavoidable. Nvidia CEO Jensen Huang has confirmed to employees that he had an affair with a woman US intelligence allegedly identified as a Chinese intelligence asset. Nvidia's board is now running an independent investigation. That part alone would be a one-day story. The part that makes it a genuine corporate and geopolitical problem is the timeline.

According to reporting from the Financial Times and the New York Times, US intelligence services concluded roughly a year ago that this woman, Alice Huang, a senior executive at the Singapore-based company Megaspeed, had cultivated a relationship with Jensen Huang on behalf of China's MSS. Nobody told Nvidia. Nobody told Jensen. The alleged reasoning floated in reporting is almost as uncomfortable as the affair itself: the administration may have held back to avoid disrupting chip sales to China and to avoid friction with a company that has been lobbying hard to keep that revenue flowing.

A Company That Was Already Under the Microscope

Here is where the plot thickens. Megaspeed was not some random contractor. It was an official Nvidia cloud partner that, through a Malaysian subsidiary called Speed Matrix, bought an estimated 4.8 billion dollars of Nvidia hardware between mid-2024 and early 2026. That same company is now under joint US-Singapore investigation for allegedly smuggling up to two billion dollars worth of restricted AI chips into China.

The mechanism is almost embarrassingly simple. US export controls target China directly, but they do not apply to Singapore or Malaysia. Chips move legally into those countries, then quietly make their way across the border. Neeta's point on this episode is that it is not an isolated workaround, it is a pattern repeating across Southeast Asia and, increasingly, the Middle East. Tens of billions of dollars in AI chips have landed in Malaysia and Singapore this year, with credible reports of similar activity in Qatar and the UAE, a jurisdiction the US recently upgraded in status despite long-standing objections from national security hawks in both parties.

Nvidia's Quiet Course Correction

The most telling detail may be Nvidia's own response. The Financial Times reported that Nvidia recently told global partners in the Middle East and Southeast Asia that at least 51 percent of their computing capacity must serve non-Chinese customers. Notably, this was not a written policy. It was passed down verbally to sales teams, the kind of move that suggests someone wanted as little paper trail as possible.

The timing invites an obvious question: did this directive arrive because Nvidia finally learned about the spy allegations and the exposure they created, or was it a separate response to mounting scrutiny of chip smuggling generally? We do not yet know, and that is precisely the kind of detail Congress is going to want nailed down.

Why This Matters Beyond the Gossip

Three Republican committee chairs have already asked Huang to brief them on November 18th, one day after Nvidia's third quarter earnings. With control of Congress shifting, incoming Democratic chairs are likely to want the same briefing, probably with a broader goal: understanding whether other American tech executives have been targeted with similar tactics, and whether export control loopholes are a systemic failure rather than a Megaspeed-specific one.

For Nvidia, the stakes are straightforward even if the politics are not. The board says it has no evidence Huang shared classified or confidential information. If that changes, his position becomes very hard to defend. For the broader AI chip trade, this episode is a reminder that export controls built around national borders are easy to route around when the real border that matters is corporate accountability, and right now that border looks porous on every side.