Founder predicts most inference moves to variable or async pricing, fixed price stays scarce and pricey
evanjconrad · x · 2026-10-18
evanjconrad predicts that almost all inference will shift to either variable pricing or async, with only a small share remaining at fixed prices — and that slice will be very expensive.
In the follow-up thread, JoshPurtell and lukedavis note this makes life harder for application developers: building apps that control end-user latency under fluctuating API cost/latency is already difficult, and even without strict latency constraints, balancing cost vs. quality remains hard. lukedavis adds that the human-interactive part of applications will matter less over time as most inference becomes part of massive, unknowable async workloads.
More from Infra
- OpenJet claims Qwen3.8-27B coding runs 1.9x faster than llama.cpp on one RTX 3090 — JeremyCMorgan · 2026-10-23
- Jensen Huang lays out NVIDIA's full Vera Rubin agentic system: NVL72, MGX, CMX and RTX — rohanpaul_ai · 2026-10-23
- Jensen Huang details Vera Rubin agentic stack: NVL72, MGX, CMX and RTX — rohanpaul_ai · 2026-10-23
- ONNX to Transformers.js: running models directly in your browser — JosephJacks_ · 2026-10-23
- MiniMax M3.1 Flash free for 3 days on 0G, with 1M context and 512K output — MiniMax_AI · 2026-10-23
- "We bet on Cloudflare and it paid off": R2 and D1 at billion scale — dinasaur_404 · 2026-10-23