Open models hit 78.4% of token volume on Vercel AI Gateway; Chinese labs' spend tops OpenAI

burny_tech · x · 2026-09-21

Vercel CEO rauchg reports that open-weight models hit a record 78.4% of daily token volume on the Vercel AI Gateway, with closed models at just 21.6%.

By inference spend, Moonshot AI and DeepSeek rank #3 and #4, and combined with Z.ai their total spend surpasses OpenAI's (#2). Caveat: this measures inference spend across providers (mostly in the US), not revenue going to the open-weight labs. Nat Lambert notes absolute volume data would be even more informative than percentages.

Related event: Open-Source Models Hit Record 78.4% of Vercel Gateway Tokens(2 posts)→

Original post →

More from Venture

Venture channel →