US Tech Coalition Urges Open Approach to AI Models

Open-weight models make their trained parameters available for download, allowing companies, researchers and startups to customize and operate AI systems on their own infrastructure. Supporters say the approach lowers development and computing costs, reduces dependence on a handful of proprietary providers and speeds adoption across the economy. The issue has gained strategic urgency as increasingly capable Chinese models narrow the performance gap with leading US systems and attract developers worldwide.

On July 24, 2026, a coalition of 76 companies, including Nvidia, Microsoft, Meta, OpenAI, Google, Cisco and Andreessen Horowitz, urged Washington to avoid broad restrictions on open-weight AI. Three days later, 37 companies launched the Open Secure AI Alliance to promote transparent, collaborative development. The groups acknowledged that downloadable models can be modified for malicious purposes but argued that wider scrutiny can improve vulnerability detection, competition and cybersecurity resilience.