Cost math: Meta Muse would run $51 per user, making 'free for 4B users' a steep climb
bookwormengr · x · 2026-09-29
The author built an interactive GPU/CPU cost calculator for Meta Muse. TL;DR: at 15 min of daily use and 10M tokens/day (easy in an agentic world), it costs $51 per user, and Meta would need $70 per user to keep current margins — a steep task for a free product. Baselines come from SemiAnalysis (InferenceX, AgentX), DeepSeek DSec, AWS pricing and Meta results. The author predicted Muse has a small CPU component (4-8B) and a large GPU component (201B), and explains CPU oversubscription: each sandbox can have 2-4 cores but average utilization is under 5%, so one physical core can support 25 effective sandbox cores — the right formula is not 'per-sandbox power × concurrent users'. All parameters are adjustable in the calculator.
Related event: Meta Muse Inference Costs Estimated at $50 Per User Annually(3 posts)→
More from Infra
- Gemini 3.5 Flash now runs on air-gapped Google Distributed Cloud, offline up to 12 months — shashib · 2026-11-16
- Gemma.c: full Gemma 1/2/3 multimodal inference in a single C file, ~72 tok/s on CPU at INT8 — This-Purchase-3325 · 2026-11-16
- Sarvam AI Launches Token Factory: India-Hosted Inference for Open Models — itsOmSarraf_ · 2026-11-16
- BIS Issues Interim Final Rule Addressing Remote Access to Advanced AI Chips — apples_jimmy · 2026-11-16
- Jailbroken PS4 runs a 9B open-source model at up to 43 tok/s by talking directly to its 2013-era GPU — AIFlow_ML · 2026-11-16
- Mobile AI Factory: agents auto-build on-device SLMs with 19x less GPU time — pmttyji · 2026-11-16