Databricks co-founder argues US must go open source to beat China in AI
What Happened
VC Andy Konwinski argues that the U.S. is losing its AI research dominance to China.
Our Take
"Open source beats China" is VC bingo. The real fight's already lost on inference cost.
Open source didn't win against proprietary in any major infrastructure war. Linux won because it solved a problem (cost). Foundational model training? That's $100M+ compute. You can't crowdsource that. China's not winning because they're greedy with weights—they're winning because they'll spend capex that VC-funded US startups won't.
Open source isn't the strategy here; it's desperation reframed as principles.
What To Do
Stop waiting for open-weight models to catch up on performance—the capex asymmetry means proprietary wins on inference speed for 2+ years.
Cited By
React