Sarvam launches Token Factory: India-hosted open-weight model inference with 99.95% SLA
itsOmSarraf_ · x · 2026-11-17
Indian AI company Sarvam AI launched Sarvam Token Factory, reportedly the first service hosting frontier open-weight models in India: it serves Sarvam's own models plus leading open models like GLM, DeepSeek and Gemma on Indian GPUs, bills in INR, targets regulated industries with optional zero data retention, and backs reliability with a 99.95% uptime SLA, auto-scaling capacity, fleet health monitoring and staged rollbacks. Commenters call it a first for India that could unlock many local enterprise use cases.
More from Infra
- Commerce and Justice issue joint statement on criminal probe of Singapore chip firm Megaspeed — dotey · 2026-11-17
- India's sovereign AI lab Sarvam hosts mostly Chinese models on subsidized H100s — ayushthakur0 · 2026-11-17
- Nvidia put Megaspeed on its 'no sell' list in June 2025 after US investigators requested sales records — Polymarket · 2026-11-17
- NVIDIA open-sources cuPotential, over 3x faster MD nonbonded interactions vs OpenMM — AllThingsApx · 2026-11-17
- Syzygy says compute shortage is a software problem, claiming 6% to 90% GPU utilization gains — teortaxesTex · 2026-11-17
- Replicating 4-bit forward-pass training on one B200: 38% faster at 3.3B with identical loss — OtherRaisin3426 · 2026-11-17