LMSYS Launches SGLang Consortium with NVIDIA, AMD, Google, Meta and More

At the inaugural SGLang Summit, LMSYS officially announced the formation of the SGLang Consortium. Founding members include NVIDIA, AMD, Intel, Google, Meta, Cloudflare, Modal, and Radixark, all of whom have joined as maintainers to jointly shape the future of this open-source inference engine. It marks a milestone in SGLang's evolution from a community project to industry-led collaborative governance.

Confirmed

• LMSYS announced the SGLang Consortium during the opening keynote of the first-ever SGLang Summit
• The consortium has eight founding members: NVIDIA, AMD, Intel, Google, Meta, Cloudflare, Modal, and Radixark
• NVIDIA's official AI account endorsed SGLang and the Miles community at the summit
• NVIDIA's collaboration with SGLang spans five generations of accelerated computing architectures, from Ampere to Vera Rubin, backed by over two years of joint engineering
• NVIDIA also noted that the two sides have co-hosted 10+ related events (m4)

Unconfirmed

• The number of models with day-zero support varies across posts: m2 claims 100+ models, while m1/m4 say 60+. NVIDIA has not provided an official figure, so the original announcements should be treated as authoritative

Why it matters

The SGLang Consortium brings together the three major chip vendors—NVIDIA, AMD, and Intel—alongside major model providers like Google and Meta, signaling that the ecosystem race around open-source inference engines has entered a multi-vendor co-governance phase. NVIDIA's high-profile showcase of deep collaboration across five GPU generations also shows that leading hardware vendors are competing for dominance over the open-source inference stack and day-zero support capabilities.

2026-11-13 ~ 2026-11-14 · 6 related posts

• SGLang launches consortium with NVIDIA, AMD, Google, Meta, Intel and Cloudflare as founding members — xiaosun86 · 2026-11-13
• Nvidia, AMD, Google, Meta join new SGLang Consortium for open-source inference — ying11231 · 2026-11-13
• NVIDIA touts SGLang partnership: 5 GPU generations, 60+ day-zero model launches — NVIDIAAI · 2026-11-14
• [source] NVIDIA joins new SGLang Consortium after 5 GPU generations of collaboration — NVIDIAAI · 2026-11-14
• [source] NVIDIA details two-year SGLang collaboration spanning five GPU generations — NVIDIAAI · 2026-11-14

1 near-duplicate retellings: NVIDIAAI