Why the rise of open source AI isn’t hurting Anthropic … yet
TL;DR
TechCrunch frames this less as direct cannibalization and more as a lifecycle split: open source models take over mature, cheaper production workloads while frontier models still power new, uncertain use cases. On Vercel’s AI Gateway, DeepSeek reportedly handled a little over one third of recent token volume; Anthropic still represented more than half of total AI spend on the platform.
Nauti's Take
This is an uncomfortable but realistic market structure: open source eats the routine work first, not necessarily the margin. Anthropic can live with that as long as teams still pay for Claude on difficult, risky, or undefined tasks.
The turning point comes when open models are not only cheaper, but good enough for the discovery phase too. Then coexistence becomes real pricing pressure.
Briefingshow
The key issue is not whether open source models are improving, but which layer of value they attack first. If companies only switch to cheaper models after validating a workflow, frontier labs remain the expensive discovery and experimentation layer. That protects Anthropic in the short term, but increases its dependence on a constant flow of new high-end use cases.
Sources