Harvey, a legal technology startup based in San Francisco, announced on Thursday the launch of its new model, Harvey Tenet. The company, which counts OpenAI, Sequoia Capital, and Andreessen Horowitz among its investors, confirmed that the model was post-trained using the Kimi K3 open-weight base model developed by the Chinese AI lab Moonshot AI.

This development marks a notable shift in the artificial intelligence industry. While many Western tech firms have historically relied on domestic infrastructure, the decision by Harvey to integrate a Chinese-developed model suggests that the high costs associated with training proprietary AI systems are driving companies to seek more cost-effective alternatives. By utilizing an open-weight model like Kimi K3, Harvey is able to leverage existing technology rather than building a foundational model from scratch.

The integration of a Chinese-developed model into a US-based legal tech platform is particularly significant given the current geopolitical climate surrounding technology and data security. While Harvey has framed the move as a strategic technical decision, the reliance on foreign open-weight systems by a firm backed by major Silicon Valley entities underscores the global nature of the current AI arms race. As development costs continue to soar, industry analysts suggest that more Western firms may look toward international open-weight systems to remain competitive in the legal automation space.