Coinbase Races Towards a Future Where AI Writes Over Half Its Code by October 2024

Coinbase CEO Brian Armstrong has set a bold target: by October 2024, more than half of the company’s daily code output should be generated by artificial intelligence. With nearly 40% already written by AI, this move signals just how rapidly the tech world is embracing automation in software development.

Key Takeaways

• Coinbase aims for 50% of its codebase to be AI-generated by October 2024

• Current AI-generated code hovers around 40%, more than double since April

• Armstrong has actively pushed engineers to embrace AI tools

• AI is being adopted beyond engineering, into design, product, and finance

• Debate continues over job impacts and responsible AI integration

AI’s Rapid Expansion Within Coinbase

The shift toward AI-generated code at Coinbase isn’t just a possibility—it’s happening at speed. The company has seen the portion of code written by AI climb from less than 20% to over 40% in a matter of months. Armstrong’s strategy has been both transformative and controversial: he’s required engineering teams to harness AI assistants like Copilot, Claude Code, and Cursor.

Notably, Armstrong enforced this by letting go of some engineers who resisted the new approach, an action he later admitted was a “heavy-handed” but clarifying move. The CEO’s vision is clear: Coinbase should be an “AI-Native” organization rather than one that simply replaces humans with machines.

AI Integration Across the Organization

Coinbase’s AI journey goes beyond backend engineering. Armstrong emphasizes that multiple departments—including finance, product management, and even decision-making teams—are utilizing AI. Some meetings now treat AI as an additional participant, with its insights factored into the collaborative process.

Such integration reflects broader trends in tech: recent research found that as of June 2025, 94% of tech firms were using AI code assistants. Most cite significant benefits in productivity and delivery speeds. Coinbase is also hiring aggressively, with nearly half of its open roles in technical or AI-related areas.

Challenges and Questions Around AI Adoption

While AI brings speed and scale, its integration has sparked internal and industry-wide debate. Some worry that heavy reliance on AI, especially without proper review, could lead to codebase knowledge gaps or introduce security vulnerabilities. At Coinbase, not all parts of the business are suitable for AI-generated code—front-end and standard operations see faster adoption than sensitive, low-level systems.

Concerns also remain over the data used to train AI models and whether those whose code or data trains these systems receive fair compensation. Ethics, transparency, and responsible use remain top priorities, even as the push for efficiency accelerates.

The Broader Impact: AI’s Role in Tech and Crypto

The company’s AI-first approach lands in the midst of an ongoing discussion about jobs. While some forecasts warn of eventual tech job declines, others, including Armstrong, argue that AI will empower workers, not replace them outright. With competitive pressure from industry giants and growing interest in AI across sectors, Coinbase’s strategy could set a new standard for how organizations embrace powerful new tools—responsibly and at speed.

Sources

• Brian Armstrong Wants Half of Coinbase Code To Be AI-Written By October, Yahoo Finance.

• Coinbase Aims to Make AI a Code-Making Co-Worker, AInvest.

• Brian Armstrong Says AI-Generated Code To 'Eclipse' Human Output at Coinbase by Year-End, CCN.com.

• Coinbase’s Brian Armstrong Says AI Writes 40% Of Its Code, Cointelegraph.

• Brian Armstrong Wants More Than Half Of Daily Code Written At Coinbase To Be AI-Generated: Should Use 'Responsibly As Much As We Possibly Can' - Amazon.com (NASDAQ:AMZN), Coinbase Global (NASDAQ:COIN), Benzinga.