Published Jan 30, 2025

DeepSeek, Stargate, and the new AI arms race

Explore the groundbreaking rise of DeepSeek, an open-source AI model reshaping the industry's cost dynamics and geopolitical landscape, as Nilay Patel and Kylie Robison delve into its impact compared to the controversial $500 billion Stargate project, questioning the future financial strategies in the AI arms race.
Episode Highlights
Decoder with Nilay Patel logo

Popular Clips

Questions from this episode

Episode Highlights

  • AI Costs

    DeepSeek's economic model is reshaping AI market expectations by offering high-performing models at a fraction of the cost of competitors like OpenAI. highlights that DeepSeek's models perform on par with or better than those from major players, yet are free to use 1. This stark contrast to OpenAI's costly GPT-4, which reportedly cost $100 million to train, has led to significant market upheaval. notes that the traditional AI cost structures are being challenged, as DeepSeek demonstrates that massive financial investments aren't the only path to AI advancement 2.

    The AI cost structures we've been operating under are going out the window.

    ---

    This shift is prompting a reevaluation of AI valuations and investment strategies, as companies realize that efficiency and innovation can rival sheer financial power.

       

    Market Impact

    The financial markets have reacted dramatically to DeepSeek's emergence, with significant impacts on AI company valuations. explains that the revelation of DeepSeek's capabilities led to a massive sell-off in tech stocks, particularly affecting Nvidia, which saw a record decline 3. This reaction underscores the market's sensitivity to shifts in AI technology and cost structures. The concept of Jevons Paradox is relevant here, as making AI more cost-effective could lead to increased consumption and demand 4.

    If we're able to do this so efficiently with lower cost, that means if we do it with way more chips, then it will create better models.

    ---

    This paradox highlights the potential for smaller startups to compete with established giants by leveraging cost efficiencies, potentially leveling the playing field in the AI industry.

       

    Open vs Closed

    DeepSeek's open-source model challenges the closed-source approaches of companies like OpenAI, sparking a new phase in the AI arms race. points out that DeepSeek's success demonstrates the potential of open-source models to rival proprietary ones, raising questions about the sustainability of closed-source strategies 5. This development is reshaping the AI landscape, as open-source models offer a cost-effective alternative that could disrupt traditional business models.

    DeepSeek and Stargate represent more than just two different paths towards superintelligence.

    ---

    The geopolitical implications are significant, as the US-China relationship and AI geopolitics become increasingly intertwined with these technological advancements 6.

Related Episodes