Nvidia is using buybacks to pay off non-believers

Something is artificial alright… Buying back stock at the peak of the hype cycle will not age well.

Travis Webb
3 min readAug 25


Media Distortions

I went to dinner with a friend two nights ago, and along with an excellent asada enchilada and a margarita, we had a discussion that went roughly like this:

Friend: AI is going to kill us all. It’s going to take our jobs. Did you see that lawyer who used ChatGPT to cite legal cases that turned out to be fake?

Me: Yes, that lawyer is the perfect example of someone who watches a lot of television news and starts to believe the hype that this thing is some all-knowing god-like Deep Thought computer.

Friend: ok but everyone in the media is saying the same thing, so why are you right and they are all wrong?

I am not right or wrong, because I am not making any claim whatsoever, nor do I have any motives to make a claim one way or the other. I may benefit at the margin somewhat in my normal job if GenAI turns out to be a deity worth worshipping, but that isn’t enough for me to proclaim the end of the world in order to make a few bucks.

One the other side, many people have quite strong and transparent motives to make public claims that money and fame and leaders will be made and unmade because of GenAI.

  • The television (and to a lesser degree, print) media thrives by keeping people amped up about some crisis or other.
  • Salespeople at the Cloud providers and AI companies are doing their job — making claims about the very thing they are selling
  • Nvidia’s Jensen Huang has Gone Full Masa, claiming a permanent shift in aggregate demand due to today’s AI hype.
  • Optimism is rarely punished on Wall St. The average analyst price target in 2021 missed today’s price (~$450) by more than half, and was around a third of the current target of $625. Clearly, the un-prescient targets of 2021 are being rapidly revised to “correct” a previously incorrect forecast.

Believing your own marketing

Ground truth is difficult to ascertain in the middle of a hype cycle. Buying back stock, however, is an unmistakable signal that you really want the market to believe what you believe. You’re saying to the world: “I believe so thoroughly in our products and strategy that I’m willing reinvest excess profits back into the company. Not only that, I’m doing so at all-time highs”. If the CEO believes it, why shouldn’t I?

But stock repurchases can accomplish many things at once beyond the simplistic “reducing the share count” arithmetic. Even if the share price is at an all-time high, An important effect of a repurchase program is that you are effectively buying out non-believers. Even the biggest skeptic is at least somewhat mollified if they just made a lot of money from the thing they are skeptical of.

Something’s artificial alright…

Normally when your company is wildly overvalued, you want to sell shares into the market to raise cash, especially as real rates are rising and capital becomes more expensive. Repurchasing shares makes sense when 1) your stock is undervalued (or at least reasonably valued), and 2) you have excess cash that you can’t use elsewhere in the business. Nvidia arguably could meet criterion #2, but certainly not #1.

A repurchase authorization of course does not obligate the company to buy its own shares, and I suspect that even the board isn’t crazy enough buy NVDA shares at these levels. The GenAI hype could be totally legit, but it still can’t possibly make sense to repurchase shares at a time of unprecedented market demand for your stock. The authorization feels more like signal than action, and doesn’t at all feel like an expression of strength; rather it feels like $25B marketing campaign to force the hype train to keep rolling when it will no longer do so under its own momentum.



Travis Webb

Solutions Architect at Google