Indie Dev Asks How to Configure a Multi-Machine Local AI Fleet to Replace Frontier Subscriptions
jakethunderpants · reddit · 2026-09-21
An indie and AAA game developer wants to offset frontier model subscription costs with local hardware, keeping only ChatGPT/Codex and Grok. His fleet includes a Dell GB10 (128GB RAM), two 5070 machines (8GB VRAM), a 4090 laptop, an aging 8GB Mac Mini, and two idle 3080s. He is weighing how to allocate workloads and whether to keep an incoming Mac Studio M5 Ultra 256GB, aiming to offload as much coding as possible to local models.
More from Infra
- AI agents, not chatbots, are the real reason for the data center power buildout — GaryMarcus · 2026-09-21
- Matthew Green: assumptions about idle GPU capacity are probably wrong — matthew_d_green · 2026-09-21
- Compute lending panic overblown: $10B+ in data center deals priced in two months — Jsevillamol · 2026-09-21
- Meta spends 30 minutes at All-In Summit pitching data centres, gets heckled on stage — agstrait · 2026-09-21
- Cloudflare has been 'ngroking' for years, dev notes its tunnel daemon is nothing new — HankYeomans · 2026-09-21
- Why model-layer margins should be negatively correlated with CoreWeave, Oracle and Nvidia — GavinSBaker · 2026-09-21