Coinbase Says AI Now Helps Write More Than 95% of Its Code

Coinbase’s Rob Witoff said the company is increasingly relying on AI to improve execution, while emphasizing that high-agency employees remain essential for making strategic decisions and exercising sound judgment.

Artificial intelligence now generates more than 95% of Coinbase’s code, providing fresh insight into the crypto exchange’s AI-driven development strategy after the company reduced its workforce by 14% earlier this year.

Coinbase laid off 700 employees in May. In a message to staff, CEO Brian Armstrong said AI had “dramatically” accelerated the pace of work, adding that the company needed to “return to the speed and focus of our startup founding, with AI at our core.”

“Practically all of our employees use AI every day,” Coinbase’s Head of Platform, Rob Witoff, told . “And nearly all of our code — probably between 95% and 100% — is now written either by or with the help of large language models (LLMs).”

The latest figure is more than twice Coinbase’s estimate from February, when the company said AI assisted in writing about 40% of its code. The increase highlights the rapid acceleration of AI adoption across both the technology and cryptocurrency sectors.

Witoff said AI is used across a “wide spectrum,” with its role varying depending on the task. He explained that core cryptography still depends heavily on human expertise, while prototyping is almost entirely automated, and AI plays a more balanced role in the development of core systems.

“For example, when we’re writing core cryptography, we have industry-leading cryptographers that are meticulously researching and reviewing one line at a time.”

“We rely heavily on AI to test our code, verify that it functions as intended, identify potential vulnerabilities, and check the underlying math,” Witoff said. “However, that process still requires significant human oversight, while the development of internal prototypes is now almost entirely automated.”

Smaller Teams Empowered by Influential Tastemakers#

Witoff said the shift has enabled Coinbase to restructure into smaller, more experienced teams, with just two or three employees now able to complete work that previously required teams of 10 or more.

“There were many junior development positions affected,” Witoff said, referring to the May layoffs. He added that the workforce reductions also extended to marketing, legal, customer support, and compliance teams.

“For those smaller teams to work, for people to have the taste, the judgment, I think a lot about people having the battle scars so they know how to point agents in the right direction.”

Witoff said most Coinbase engineers now work alongside five to 10 AI agents at any given time, with those AI systems collectively handling coding tasks equivalent to the workload of roughly 1,200 employees.

Witoff said Coinbase expects AI agents to perform work equivalent to that of nearly 100,000 employees by 2030, reflecting the company’s long-term vision for AI-driven productivity.

Coinbase joined a growing number of companies that reduced their workforce this year as enterprise AI adoption accelerated. In March, crypto exchange Crypto.com laid off 12% of its employees, affecting roles the company said “do not adapt in our new world.”

In February, Block CEO Jack Dorsey announced plans to reduce the company’s workforce by 40% as part of a broader restructuring effort.

“We’re already seeing that the intelligence tools we’re creating and using, paired with smaller and flatter teams, are enabling a new way of working which fundamentally changes what it means to build and run a company,” Dorsey said on X.

Several other crypto firms, including Kraken, Gemini, Messari, and Dune, also attributed this year’s workforce reductions in part to efficiency gains driven by artificial intelligence.