The abrupt appearance of DeepSeek’s R1 advanced reasoning model at the start of the year was akin to the “shot heard ‘round the world” in AI circles. Major tech companies had spent recent years pouring billions into generative AI projects, products, and infrastructure. Meanwhile, Chinese startup DeepSeek created in just months a model as good as OpenAI’s then-most advanced product on industry-standard benchmarks—and claimed to have spent just $6 million to train it. DeepSeek’s model also used a fraction of the compute as its U.S. competitors—and it’s still free to use.
