Anthropic Says AI Now Writes 90% of Its Code
CEO Dario Amodei tells Marc Benioff at Dreamforce that Claude handles most coding at Anthropic, lifting team productivity roughly tenfold while engineers focus on review and harder problems.
Topics
News
- Anthropic Says AI Now Writes 90% of Its Code
- Space42 Signs MoU With e& For V2X technologies
- Instagram Tightens Teen Controls Amid Safety Scrutiny
- AI Public Servants and Biotech Breakthroughs Make Waves at GITEX 2025
- From Oil to AI: UAE’s Stargate Campus Targets 200 MW Online by 2026
- GITEX 2025: Sam Altman Calls for Cheap, Plentiful AI to Close Global Gap
AI models now generate about 90% of code for most teams at AI startup Anthropic, CEO Dario Amodei said in a conversation with Salesforce CEO Marc Benioff at the Dreamforce conference.
Earlier this year, Amodei had said that AI would write 90% of the code within six months.
“Some people think that prediction is wrong, but within Anthropic and within a number of companies that we work with, that is absolutely true now,” he said, adding that teams are roughly 10x more productive and “rebalanced” around AI-assisted development.
Amodei argued the shift is boosting output rather than cutting jobs, with engineers focusing on review, complex problems, and supervision.
“If Claude is writing 90% of the code, what that means, usually, is you need just as many software engineers. You might need more, because they can then be more leveraged,” he said.
Optimism on AI Coding
Amodei isn’t alone in this optimism. Venture Capital firm Y Combinator’s CEO, Garry Tan, said roughly a quarter of its Winter 2025 startups have 95% or more of their code written by AI tools.
Microsoft CTO Kevin Scott has projected that 95% of all code will be AI-generated within five years.
Developers report tangible time savings. BairesDev’s Q3 2025 Dev Barometer found engineers save an average of 7.3 hours a week with AI assistance.
About 48% of them said 10–50% of their code is AI-assisted, averaging 23.5%. The survey covered 1,129 senior developers and 50 project managers across 63 countries, including 143 in India.
The market is expanding quickly. Grand View Research estimates the market size of AI coding tools at $4.86 billion in 2023, rising to $26.03 billion by 2030.
