Menlo Ventures Report: Chinese AI Models Dominate Global Enterprise Open-Source Spending
A report by Menlo Ventures reveals that Chinese AI models account for 89% of global enterprise spending on open-source models, led by Moonshot AI's Kimi, DeepSeek, and Zhipu AI's GLM. Specifically, Kimi captured 61% of the market share, followed by DeepSeek at 15% and GLM at 11%. This unexpected market dominance challenges the prevailing narrative of Western dominance in AI, showing that Chinese open-source models are capturing the vast majority of enterprise financial investments. It highlights a major shift in the global artificial intelligence landscape as enterprises increasingly adopt these competitive alternatives over Western models like Meta's Llama. In stark contrast to the Chinese models' combined 89% share, Meta Platforms' Llama series—traditionally considered the leading Western open-source model—holds only 7% of global enterprise spending in this category.
## BACKGROUND
Menlo Ventures is an American venture capital firm based in California that frequently analyzes technology and market trends across the artificial intelligence sector. Open-source models, or open-weights models, allow enterprises to freely access, modify, and deploy AI architectures on their own infrastructure.