Coinbase is cutting 14 per cent of its workforce, roughly 660 employees from a company of 4,700, two days before it reports the worst quarterly earnings in its history as a public company. CEO Brian Armstrong announced the layoffs on Tuesday in a letter to staff that mentioned the crypto market downturn in passing and spent most of its length describing how artificial intelligence had changed the nature of work at the company. The future of Coinbase, Armstrong wrote, is “intelligence, with humans around the edge.” The restructuring eliminates pure management roles, caps the organisation at five layers below the CEO and COO, and introduces what the company calls AI-native pods: small teams, some as small as a single person, that use AI tools to do what used to require a department. The framing is deliberate. Armstrong is not presenting this as a cost cut forced by a collapsing market. He is presenting it as a structural transformation that happens to coincide with one.

The numbers

The coincidence is hard to ignore. Wall Street expects Coinbase to report revenue of roughly 1.5 billion dollars for the first quarter of 2026, a 26 per cent decline from the same period last year. Earnings per share are forecast at 36 cents, down from 1.94 dollars a year ago. Consumer transaction revenue fell 45 per cent year on year to 734 million dollars as cryptocurrency prices collapsed and traders migrated to lower-fee tiers. Bitcoin recorded its worst first quarter since 2018, falling between 22 and 24 per cent. Ether dropped 41 per cent. Global cryptocurrency exchange volume fell nearly 48 per cent from its October 2025 peak to 4.3 trillion dollars in March, the lowest level since October 2024. Coinbase’s stock is trading 57 per cent below its 52-week high of 444.65 dollars. The company’s operating expenses rose 22 per cent year on year to 1.5 billion dollars, growing at more than twice the rate of revenue, driven in part by integration costs from its acquisition of the derivatives exchange Deribit. The restructuring will cost between 50 and 60 million dollars in severance.

The one bright spot is institutional trading. Revenue from institutional transactions grew 31 per cent year on year to 185 million dollars, supported by Deribit’s record quarter in derivatives. But derivatives cannot compensate for the collapse in consumer trading volume that has historically driven the majority of Coinbase’s revenue. The company guided subscription and services revenue to a 590 million dollar midpoint for the first quarter, missing Wall Street’s consensus of 761 million dollars by 22 per cent before the quarter had even finished. These are not the numbers of a company restructuring because AI made it possible. These are the numbers of a company restructuring because the market made it necessary.

The narrative

Armstrong’s letter is carefully constructed. He acknowledged the downturn but framed the layoffs as an acceleration of a transformation already under way. “Engineers use AI to ship in days what used to take a team weeks,” he wrote. The restructuring eliminates roles for “pure managers” and replaces them with player-coaches who both lead and build. AI-native pods are designed to operate with minimal coordination overhead, using AI tools for code generation, customer support automation, compliance monitoring, and internal operations. Armstrong told employees that Coinbase plans to hire only people with strong AI skills going forward, and that the company’s future depends on being “AI-first in everything.”

The language is familiar. Klarna froze hiring in 2025 citing AI productivity gains, claiming its AI assistant was doing the work of 700 customer service agents. Atlassian cut 1,600 jobs in March 2026 and replaced its CTO, framing the reduction as an adaptation to the AI era. Block, the payments company run by Jack Dorsey, cut 4,000 employees in February, nearly 40 per cent of its workforce, with Dorsey declaring that AI had made the company’s old structure obsolete. In each case, the announcement emphasised what AI could now do rather than what the market had done to revenue. In each case, the financial context suggested the market was the more immediate cause.

The pattern

The term for this is AI-washing: attributing layoffs to artificial intelligence without clear evidence that AI systems have replaced the workers being let go. A December 2025 survey found that 59 per cent of hiring managers admitted they emphasised AI in layoff announcements because it “plays better with stakeholders” than acknowledging financial constraints. Oxford Economics research found that AI-related job cuts accounted for just 4.5 per cent of total layoffs in the first eleven months of 2025, while standard market and economic conditions drove nearly four times as many. OpenAI’s Sam Altman described the practice as AI-washing in February, noting that fewer than one per cent of 2025 job losses could be directly attributed to artificial intelligence.

The distinction matters because the narrative determines the response. If Coinbase is restructuring because the crypto market crashed, the layoffs are cyclical and the company will rehire when trading volume recovers. If Coinbase is restructuring because AI has permanently reduced the number of humans needed to run a cryptocurrency exchange, the jobs are gone and the remaining employees are working in a fundamentally different organisation. Armstrong is betting on the second interpretation, or at least presenting it as the truth. The stock market tends to reward that framing. Block’s shares rose 18 per cent the day Dorsey announced his AI-driven cuts. Investors prefer structural transformation narratives to cyclical downturn admissions because the first implies higher future margins and the second implies the company is at the mercy of a market it cannot control.

The test

What makes the Coinbase restructuring different from Block’s or Klarna’s is the specificity of the organisational changes. Armstrong is not just cutting headcount and calling it AI. He is eliminating an entire layer of the management hierarchy, capping the organisation at five levels, and introducing a team structure, the AI-native pod, that is designed to function with AI tools as core infrastructure rather than productivity supplements. The single-person team is the logical endpoint of this model: one engineer, using AI for code generation, testing, deployment, monitoring, and customer interaction, producing the output that previously required a team of five or ten. If the model works, Coinbase becomes a case study in how a company with nearly five billion dollars in annual revenue can operate with a fraction of the workforce it had two years ago.

Mark Zuckerberg told Meta employees that the company’s layoffs were about capital expenditure reallocation, not AI productivity, a rare moment of candour in an industry where the AI narrative has become the default explanation for every reduction. Armstrong is not offering that candour. He is making a stronger claim: that AI has changed what is possible, and that the company’s structure must change to match. The claim may be true. AI coding tools have measurably increased developer productivity at companies that have adopted them. Customer support automation has reduced headcount at fintech companies that deployed it at scale. But the timing, two days before the company reports a quarter in which every major revenue line declined by double digits, makes the claim difficult to separate from the financial necessity that would have produced layoffs regardless of what AI could or could not do.

The context

The broader picture of AI’s impact on employment is more nuanced than any single company’s restructuring suggests. The tech sector has recorded more than 73,000 job cuts across 95 companies in the first four months of 2026, with projections that the full-year total will exceed the 124,000 eliminated in all of 2025. Roughly one in five of those cuts explicitly cite AI as a factor. But the crypto industry has its own cyclical pattern that predates artificial intelligence: Coinbase cut 18 per cent of its workforce in 2022 during the last crypto winter, when AI was not yet a plausible explanation for anything. The industry hires aggressively during bull markets and contracts during bear markets. That pattern has not changed. What has changed is the language companies use to describe it.

Coinbase’s remaining 4,000 employees will work in an organisation that Armstrong says is built for a world where AI does most of the routine work and humans focus on judgment, strategy, and the problems that machines cannot yet solve. The 660 people who lost their jobs will receive 16 weeks of base pay, two additional weeks per year of service, their next equity vest, and six months of healthcare coverage. The severance is generous by industry standards. The framing is not. Being told that a machine can do your job is a different experience from being told that the market turned against your employer, even if the second explanation is closer to the truth. Armstrong has chosen the first explanation because it positions Coinbase as a company that is ahead of a structural shift rather than behind a cyclical one. Whether the market believes him will be visible on Wednesday, when the earnings report lands and investors decide whether the AI-native pod is a genuine organisational innovation or a rebranding of a cost cut that the numbers demanded regardless.