DeepSeek Shocks AI World

In late January, DeepSeek released its R1 reasoning model, built on December’s V3. These models shocked the world, as overnight and from seemingly out of nowhere, they matched ChatGPT’s latest benchmarks at a fraction of the price. Markets responded instantly: Nvidia suffered a record one-day market-cap wipeout as investors feared efficiency, not brute-force scaling, could reset AI economics for both the U.S. and China. The releases, called AI’s “Sputnik moment,” invigorated AI as a geopolitical security dilemma, adding an arms-race element to an already combustible situation.