DeepSeek Is Number One and the Market Is Having a Moment
A Chinese AI lab tops the App Store and the Nasdaq drops 600 points, and somehow people think these are related.
DeepSeek hit #1 on iOS. The market had its moment. Nvidia dropped. The analysis that "we can do more with the same chips" means you buy more chips, not fewer, was exactly right — Jevons paradox in real time.
DeepSeek is the top free app on iOS right now. A Chinese AI lab. Number one. Above everything.
The effects of this will be substantial — that part feels obvious. What's less obvious is which effects, running through which mechanisms, landing where.
The stock market read the WSJ piece and did its thing. Nvidia dropped. The narrative writes itself: if you can get frontier results with less compute, the chip buildout looks like overreach. Jensen Huang's net worth takes a hit. Everyone nods along.
Here's the thing though — the "less compute achieves more" framing doesn't actually cash out the way people are running with it. Even if DeepSeek's efficiency claims hold up completely, the conclusion isn't we need fewer chips. It's we can do more with the same chips, which tends to mean you want more chips, not fewer. Jevons paradox has entered the chat, as it always does, slightly too late to stop anyone from making the trade.
The funnier story is what's probably actually driving the dip. The US president spent the weekend going absolutely feral on Colombia over tariffs — threats, counter-threats, the full spectacle — and the broader geopolitical read is getting darker by the hour. That's a much more plausible explanation for a market move this size than "a chatbot app went viral."
But DeepSeek makes a better headline, so DeepSeek gets the credit.
The one thing that does feel genuinely real: local models are having their actual moment. Not the moment people kept announcing for two years, the one where you'd run a slightly worse version of GPT-3 on your laptop and call it liberation. This one. Where the gap between what fits on consumer hardware and what you'd actually want to use has closed to something you can squint past.
That's the story. Everything else is the market making up reasons after the fact.
Counterpoints
Push back, extend the argument, or sharpen it. New counterpoints go through review before they show up here.
No approved counterpoints yet.