DeepSeek Flew Under the Radar

_{Sign up to receive Surface Area or follow me on Linkedin or Twitter. Book a meeting here to talk about ghostwriting from my agency, Principals Media}.

The craziest aspect of the DeepSeek AI story is how investors and the media reacted.

I don’t mean yesterday, when Nasdaq lost $1 trillion and my feed was flooded with articles.

I mean the previous seven days.

When DeepSeek announced on Jan. 20th that it had a new high-performing large language model that cost a fraction of OpenAI’s ChatGPT, few media outlets covered it and shares of Nvidia and other tech firms didn’t move.

I can’t think of a precedent for such a large market move coming a week after a corporate announcement.

It’s arguably the biggest dropped ball in memory on the part of financial journalists, traders and money managers.

I drew three lessons:

-Twitter matters, again

-Markets are herd animals

-The mainstream media (still) matters

Let’s unpack the sequence:

On Jan. 20th, DeepSeek tweeted it was unveiling a new, open source model.

A small, but not insignificant number of news outlets covered the story, including: AI News, Analytics India, VentureBeat, Techcrunch, Yahoo! Finance and SiliconAngle.

A number of individual influencers on social media quickly retweeted the DeepSeek post, including Louie Peters, Paul Riolo and, my friend, Alex Banks.

Debates ensued on X, and blogs like Stratechery and even at Davos.

By Wednesday, Jan. 22nd, however, people had tried the model and a consensus had emerged that it was the real deal.

Bloomberg podcaster Joe Weisenthal was the first person I saw frame the issue for investors.

He tweeted: “I’m going to ask what is probably a stupid question but if deepseek is as performant as it claims to be, and built on a fraction of the budget as competitors, does anyone change how they are valuing AI companies? Or the makers of AI related infrastructure.”

Joe is a singular figure in financial journalism in that he works for a large, established brand while still having both feet firmly in social media. 

That afternoon he published a newsletter with the headline : “I’m a DeepSeek AI Bro Now. ”

The market didn’t budge and Bloomberg wouldn’t write another article with DeepSeek in the headline for five days. And neither did the Wall Street Journal or many others. The New York Times published a piece Jan. 23rd.

Meanwhile, the debate continued on X.

What’s notable is that by that point virtually all the “angles” were explored. How much did DeepSeek really spend? Who built it? How was it done? 

There were deep dives by many AI experts. 

And yet, the market opened and closed on Thursday and Friday without much ado. 

By Sunday, the tide on social media changed. There was widespread expectation Nvidia was going to get crushed in the morning. 

On Monday, major media weighed in, accelerating the negative sentiment. Nvidia lost almost $600 billion in value, the largest loss by a company in a single day in history. 

When history is written, there may be Stan Druckenmiller-types who broke the bank on this trade. 

But for now, it’s more notably for how long it took so many to react.

BRIEF OBSERVATIONS

WORKING HARD: I doubt this is the perception.

WORKING HARDER: This seems about right.

WEIGHT LOSS: Crazy that weight loss drugs are also reducing supermarket sales.

BASKETBALL GREAT: This photo captures the man and the era.

THE 1980S: I actually feel this way about the 1970s. But the meme still works.