A semiconductor on a circuit board
Context image: computing hardware; not a photograph of an AI model. Photo: Infralist / Unsplash.

What the trackers showed

Ramp-linked reporting put GPT-6 Astra near 13% of tracked enterprise spending and Anthropic’s Fable near 8% in one recent comparison. Yet Anthropic still led broader Ramp adoption, 43.8% to OpenAI’s 39.8%. OpenRouter’s crossover was also segment-specific rather than proof of whole-market leadership.

Why This Matters

Enterprise AI is becoming a portfolio market. A company can lead in adoption, another in incremental spending and a third in developer traffic. Treating those denominators as interchangeable produces confident but false market-share claims.

Background and trade-offs

OpenAI’s advantage is distribution and a broad product surface. Anthropic has built trust around coding and enterprise controls. Buyers benefit from competition and falling switching costs, but procurement teams inherit the burden of evaluating safety, latency, price and vendor concentration rather than benchmarking one score.

What happens next

Astra’s early spending lead may persist, normalize or reflect launch experimentation. Watch retained usage after several billing cycles, workload-specific win rates and whether enterprises consolidate or keep routing tasks across models. The winner will be determined less by a launch-week chart than by repeatable economics in production.

Sources: Response Rift, Cybernews, Wccftech. Facts and figures are a fixed September 20, 2026 reporting snapshot and do not update live.

Technology / Enterprise AIBack to today’s edition