What happened
Chinese AI firm DeepSeek released its new open-source V4 model, claiming rivalry with OpenAI and Google DeepMind products. This follows DeepSeek's previous model (last year) delivering near industry-leading performance at a fraction of the price, alleging more efficient chip use than US counterparts, despite China's limited access to advanced chips.
Why it matters
DeepSeek's new V4 model, claiming rivalry with leading closed-source models, shifts AI development unit economics. For platform engineers and procurement teams, DeepSeek's established track record of delivering high performance at a fraction of the price with previous models, combined with the V4's open-source nature, offers a viable alternative. This follows Nvidia CEO Jensen Huang's observation that China's cheap energy allows firms to compensate for limited advanced chips with more low-efficiency hardware. This creates competitive pressure on global AI model pricing and accessibility, impacting strategic decisions for founders and investors.
