Inference Boom Runs Blind: H100 Hours Price from $2.20 to $8.00 with No Utilization Data
AccBalanced · x · 2026-10-14
The inference business is booming and demand still outpaces supply, yet buyers, lenders, and investors have little independent evidence of what they're getting.
- OpenRouter alone routes tokens at a 4.5-quadrillion annual run rate; CoreWeave reports $104B backlog and Nscale's IPO filing shows $103.4B active and contracted value—still no true fleet utilization or inference gross margin data.
- Pricing is opaque: H100 hours list at $2.20 on DeepInfra vs $8.00 on Fireworks for not-quite-the-same products; per Artificial Analysis, GPT-OSS-120B runs 36x faster on Cerebras than the slowest host, which is among the cheapest and scores higher on accuracy.
The market lacks an independent quality/price/utilization information layer.
More from Infra
- ECDSA.fail quantum challenge hits 893 qubits x 2.2M Toffolis, 25% cheaper than Google's best — jedisct1 · 2026-10-14
- Llamaforge v0.16: a control panel for upstream llama.cpp with local agents whose quotes are code-verified — Sleepybear2611 · 2026-10-14
- Questioning OpenAI's Decisions API: are cloud round-trip latency and privacy viable in production? — WhichCardiologist800 · 2026-10-14
- Kev-4B Decision Model Goes Live in Beta on Phala with NVIDIA Confidential Computing — bgmshana · 2026-10-14
- Zuckerberg: Meta committed hundreds of billions on AI infra before knowing it would work — GaryMarcus · 2026-10-14
- NVIDIA's 400G/lane optical spec puts tougher laser demands on CPO — jwt0625 · 2026-10-14