Insider selling is probably the most over-interpreted number in all of tech finance. Every quarter, a list comes out, people with famous names appear on it, and the internet decides it has found a secret message about where AI is headed. I’d argue the opposite: this particular list tells us almost nothing about AI, and that’s exactly why it’s worth understanding.
Here are the actual facts. Bloomberg reported on October 7 that Mark Stevens, a venture capitalist and a director at Nvidia, was the single largest insider seller of the third quarter of 2026. He sold 4,236,740 Nvidia shares, worth roughly $946.6 million. Other well-known names turned up further down: Arista Networks CEO Jayshree Ullal and Amazon founder Jeff Bezos. Not everyone on the list was in AI at all. Illumina director Keith Meister came in third with about $436 million in Illumina stock, and SharkNinja chair Xuning Wang also ranked. The data was compiled by the Washington Service, a firm that tracks insider buying and selling.
One more detail matters, and it’s the one that gets skipped: the same quarter saw Nvidia’s stock valuation approach its cheapest level in over a decade. Not its price, its valuation. Those are different things, and that difference is where most of the confusion lives.
What an insider sale actually is
If you’re newer to this world, “insider” sounds sinister. It isn’t a legal accusation. It’s a category. Directors, executives, and large shareholders at public companies are classified as insiders, which means their trades have to be disclosed publicly. That’s the whole system working as designed: these people file paperwork, the filings become public, and firms like the Washington Service add them up.
So when you read that someone sold nearly a billion dollars of stock, you are not reading a leak. You are reading a form that person was required to submit, voluntarily and on schedule.
Why would a director sell? Lots of reasons, and the honest answer is that the reporting does not identify them. The sources don’t give a definitive explanation for any of these sales. What’s generally true of people in this position is that their wealth sits overwhelmingly in one company’s shares, which is the kind of concentration any financial advisor would flag. Selling can be about taxes, diversification, charitable giving, pre-scheduled trading plans, or funding something else entirely. Stevens is a venture capitalist by trade, and venture capital requires capital.
Why the AI reading doesn’t hold up
The tempting story goes like this: AI insiders are cashing out, therefore AI insiders know something bad. I understand the appeal. It feels like you’ve caught someone in the act.
But look at the list again. Illumina is a genomics company. SharkNinja makes household appliances. Jeff Bezos sells Amazon stock on a regular, well-documented basis. If a list containing vacuum cleaners and DNA sequencers is evidence of an AI bubble, then it’s evidence of a bubble in absolutely everything, which is another way of saying it’s evidence of nothing in particular.
And the valuation point cuts directly against the panic reading. A company trading near its cheapest valuation in more than a decade is not a company the market is treating as absurdly overpriced. You can argue the market is wrong. You cannot argue the market is euphoric and simultaneously pricing shares at decade-low multiples.
What I’d watch instead
For readers here, who mostly want to know whether AI agents are real and useful rather than whether a stock is up or down, I’d suggest this filter. Stock sales are a fact about one person’s finances. They are not a fact about the technology.
The signals that actually tell you something about where AI agents are going look different:
- Are companies renewing the AI tools they bought last year, or quietly dropping them?
- Are agents being trusted with tasks that have consequences, or still stuck in demos?
- Is the cost of running these systems falling fast enough for small businesses to use them?
- Are the people building agents hiring, and for what roles?
Those questions are harder to answer than reading a list of dollar amounts. That’s precisely why the list gets so much attention. It’s a number, it’s large, and it has a famous name attached. Easy to share, easy to misread.
A more boring conclusion
My honest take on Q3 2026: a wealthy director sold a large amount of a stock he holds a lot of, during a quarter when that stock was not historically expensive, for reasons nobody has publicly confirmed. Several other wealthy people did similar things in unrelated industries.
That’s it. Not a warning, not a vote of confidence, just a disclosure form doing its job. The more useful skill isn’t decoding insider sales. It’s recognizing when a headline is offering you certainty that the underlying facts don’t support.
đź•’ Published: