Chinese AI startup DeepSeek has emerged as a serious challenger to US dominance in AI, launching its open-source DeepSeek R1 model, which rivals OpenAI’s ChatGPT o1 in key benchmarks. The model’s cost-effective, high-performance approach highlights China’s innovative response to US export controls and sanctions to limit access to advanced chips. Founder Liang Wenfeng utilised a stockpile of Nvidia A100 GPUs and optimised training processes to develop R1 despite resource constraints.
DeepSeek’s rise highlights a broader shift within China’s AI ecosystem, where open-source models and collaboration among companies like Alibaba Cloud and 01.AI accelerate innovation.
As the US advances its $500 billion Stargate initiative and other AI policies under Trump’s administration, experts warn that sanctions may inadvertently push Chinese companies further ahead. With DeepSeek and others reshaping global AI dynamics, the US faces the challenge of maintaining its leadership in a rapidly evolving competitive landscape.
