Coinbase Now Writes 95–100% of Its Code with AI — Inside the Numbers, Layoffs, and the 100,000-Agent Goal by 2030

Coinbase has crossed a threshold that few publicly traded financial companies have even approached: between 95% and 100% of its code is now written or assisted by artificial intelligence. That figure — disclosed by Head of Platform Rob Witoff in a July 2026 Cointelegraph interview — represents one of the most aggressive AI adoptions in the tech industry, and it comes with a dramatic workforce restructuring, a bold 2030 target, and a growing debate about code quality and security .

Here is the verified data on the key numbers, timelines, and implications.

The AI-Assisted Code Surge: From 5.7% to Nearly 100% in 18 Months

The acceleration has been steep and well-documented. In Q1 2025, AI-generated code accounted for just 5.7% of Coinbase's total code . By September 2025, CEO Brian Armstrong announced that figure had reached roughly 40% . It stayed around that level through February 2026 .

Then, in a span of roughly five months, the percentage jumped from about 40% to 95–100%, as reported by Witoff in July 2026 . The jump represents an increase of roughly 55 percentage points in under half a year.

Coinbase executives say that nearly 100% of employees now use AI tools daily .

AI Agents Per Engineer: 5 to 10 Simultaneously

Witoff stated that most Coinbase engineers currently run 5 to 10 AI agents at the same time . These agents assist with coding, code review, testing, bug fixing, and other development tasks. The collective output of these agents is already equivalent to approximately 1,200 full-time developers .

The 2030 Target: 100,000 Full-Time Developers Worth of Agent Work

Coinbase's long-term goal is ambitious: by 2030, the company wants the collective work output of its AI agents to reach the equivalent of 100,000 full-time developers . Witoff made clear that certain critical areas — such as core cryptography — will still require human involvement, with AI primarily used for code testing, vulnerability checking, and prototyping .

Workforce Reduction: 700 Jobs Cut on May 5, 2026

On May 5, 2026, Coinbase announced it was cutting approximately 700 jobs, or 14% of its global workforce . CEO Brian Armstrong framed the cuts not as a typical cost-saving exercise, but as a fundamental restructuring to become a "lean, fast, AI-native" company . Specifically, the company eliminated "pure managers" and introduced "one-person teams" — a model where a single engineer, backed by AI agents, handles engineering, design, and product management .

The layoffs were widely covered by major outlets including Reuters, The Wall Street Journal, and TechCrunch . The company cited both "current market conditions" and the need to "optimize operations for the AI era" in its SEC filing .

Restructuring Cost: $50–60 Million

Coinbase expects to incur restructuring charges of $50 million to $60 million, nearly all of which will be cash-based severance and employee-related costs . The company said the process would be substantially complete by the end of Q2 2026 .

The Broader Industry Debate: Productivity Breakthrough or Security Risk?

The 95% AI-assisted code figure has intensified a debate that has been simmering across the tech industry. Critics have called the rapid ramp-up a "red flag," arguing that high volumes of AI-generated code can introduce subtle bugs, security vulnerabilities, and "hallucinated" logic that human reviewers struggle to catch at scale .

Coinbase's proponents frame it differently: they argue that AI-assisted development is a necessary competitive shift. The company is positioning itself as the payments rail for the coming "agentic economy," where autonomous AI agents will execute trades, payments, and contracts without human intervention . According to this view, the productivity gains — code shipped per developer is up 2x year-over-year, with top engineers pushing 100 pull requests a week — justify the accelerated AI adoption .

The broader question remains open: will other major financial and tech firms follow Coinbase's lead, or will regulatory scrutiny, code-quality concerns, and the risk of outages slow the trend? For now, Coinbase has made its bet clear.