Claude Code created daily developer surface
Why Sora failed
Claude Code showed that Anthropic could stop being just the model inside other companies products and become the product people open every day. That mattered because ChatGPT had already taught the market that the winning AI company is not just the one with the best model, but the one with the strongest habit loop. Claude Code gave Anthropic a daily developer surface, direct monetization, and a way to capture usage that had previously flowed through tools like Cursor and Bolt.1 2 3 4
- Before Claude Code, Anthropic was already the backbone of vibe coding through API demand from tools like Cursor, Bolt.new, and Vercel v0. Claude Code pulled that same coding demand into Anthropic owned workflow, where the user prompts, terminal actions, and subscription revenue sit closer to the lab.5 6 2
- The scale of the wedge was large enough to change strategy across the market. Claude Code reached an estimated $400M annualized revenue in July 2025, about 8% of Anthropic’s roughly $5B annual revenue at the time, which helps explain why OpenAI shifted GPUs and product focus toward coding and enterprise workflows.1 3
- The product shape also mattered. Claude Code was terminal native, which means developers could ask for a change, let the agent inspect files, write code, run commands, and debug inside the place they already work. That is much stickier than a pure chat window and harder for model rivals to dislodge without matching the workflow.7 8
This points toward AI labs becoming full stack software companies, not just model vendors. The next phase is a fight to own the recurring work surface for developers and then expand outward into adjacent white collar tasks, which is why coding became the clearest battleground between Anthropic and OpenAI.3 4 9