AI coding spend across 27 companies: a quant shop now pays $8M/month, devs ship 2x PRs
The Pragmatic Engineer · rss · 2026-10-21
The Pragmatic Engineer surveyed 27 companies on AI coding tool spend and PR throughput, plus aggregate data from CodeRabbit and Gitar.
Key findings:
- CodeRabbit users now ship 2x more PRs than a year ago (113-128% increase); Gitar sees "AI-native" teams hit 60+ PRs/dev/week.
- Company size doesn't predict spend: a 2-person .NET shop pays $3,000/dev/month while a 1,500-engineer public company caps devs at $150.
- A quant trading shop (2,000 devs) spends $8M/month, up 16x since January ($1,600/dev/month) — and the biggest spenders are traders, not engineers. Per-employee AI cap: $25,000.
- At Big Tech, Meta and NVIDIA push heavy AI use; Apple engineers are far more cautious.
- Spend peaked in June-July and has roughly halved since, driven by Opus 5.5 adoption and a shift to open models. Subsidized subscriptions ending could push some bills up 5-10x.
- Claude/Opus 5.5 dominate spend, ahead of Cursor, Codex, and Copilot. Almost no company can answer what the ROI actually is.
More from coding & agent
- Laude Institute launches Slingshots FOUR: 14 projects across 8 top institutions — a1zhang · 2026-10-21
- Research agent Épi sets 120 records, narrows n-queens constant interval 84x — dair_ai · 2026-10-21
- Compaction hooks in practice: keep last n messages verbatim, summarize the rest for full agent continuity — MoonL88537 · 2026-10-21
- The blank prompt box hides real operational burden: context, workflow and supervision fall on users — _AustinCalvert_ · 2026-10-21
- DeepMind paper: self-evolving agents forget as they learn — fastest learner GEPA also breaks old tasks most — rohanpaul_ai · 2026-10-21
- arXiv: Are Self-Evolving Agents Continual Learners? Plasticity comes at the cost of stability — rohanpaul_ai · 2026-10-21