DeepSeek, Stargate, and the new AI arms race

Key takeaways

DeepSeek's emergence has challenged the conventional wisdom of AI development costs, demonstrating that competitive models can be built without enormous investments.
The rivalry between open-source AI developments like DeepSeek and costly, closed-source projects like OpenAI's Stargate highlights broader geopolitical tensions.
As AI tools become cheaper and more accessible, businesses might shift away from relying solely on industry giants, leading to a more competitive landscape.
The podcast touches on the volatility of tech investments, suggesting that market corrections may stem from changes in how AI technology is produced and perceived.

Summary

The Disruption of AI Economics

The podcast explores the sudden rise of DeepSeek, a new AI model from a Chinese hedge fund that challenges the established norms around AI development costs and capabilities. DeepSeek's R1 model is reported to perform on par or better than leading models from OpenAI and others, yet costs significantly less to develop. The hosts discuss how this has led to investor panic and a severe drop in tech stocks, particularly impacting Nvidia.

A Clash of AI Philosophies: DeepSeek vs. Stargate

The episode contrasts DeepSeek's open-source model, which aims to democratize AI technology, with OpenAI's Stargate project, which seeks to invest an enormous $500 billion into AI infrastructure in the U.S. The discussion highlights the implications of this clash not only for technology but also for geopolitics, especially regarding U.S.-China relations.

The Future of AI Investment

Experts in the podcast contend that DeepSeek's approach questions the necessity of massive capital investments required by leading organizations. The hosts delve into the implications of cheaper AI models, exploring potential shifts in the market dynamics and investment strategies as companies might not need to spend exorbitant amounts on computing power to stay competitive.

Market Response and Predictions

The hosts cover the immediate impact on markets following DeepSeek’s emergence, discussing how investors are reevaluating the value of major tech stocks in light of the new competition. They also speculate on OpenAI's ability to sustain its growth and investment commitments amidst these changes.

Show notes

Today, we’re talking about DeepSeek, and how the open source AI model built by a Chinese startup has completely upended the conventional wisdom around chatbots, what they can do, and how much they should cost to develop.

We’re also talking about Stargate, OpenAI’s new $500 billion data center venture that’s supposed to supercharge domestic AI infrastructure. Both stand in stark contrast with one another — and represent a new, escalating front in the US-China relationship and the geopolitics of AI. Verge senior AI reporter Kylie Robison joins me to break it all down.

Links:

Why everyone is freaking out about DeepSeek | Verge

DeepSeek FAQ | Stratechery

DeepSeek: all the news about the startup that’s shaking up AI stocks | Verge

OpenAI and Softbank are starting a $500 billion AI data center company | Verge

The AI spending frenzy is just getting started | Command Line

After DeepSeek, VCs face questions about AI investments | NYT

Satya Nadella on Stargate: ‘All I know is I’m good for my $80 billion’ | Verge

OpenAI says it has evidence DeepSeek used its model to train competitor | FT

DeepSeek sparks global AI selloff, Nvidia loses about $593 billion of value | Reuters

Four big reasons to worry about DeepSeek (and four reasons to calm down) | Platformer

Credits:
Decoder is a production of The Verge and part of the Vox Media Podcast Network.
Our producers are Kate Cox and Nick Statt. Our editor is Ursa Wright.
The Decoder music is by Breakmaster Cylinder.
Learn more about your ad choices. Visit podcastchoices.com/adchoices