Stanford study using ADP payroll data shows workers 22-25 in AI-exposed crypto roles lost 16% relative employment. Entry-level cuts at Coinbase, Crypto.com follow the pattern.
Generative AI is cutting the first rung of the career ladder in crypto. A working paper from the Stanford Digital Economy Lab, using high-frequency payroll data from ADP (the largest U.S. payroll processor), finds that workers aged 22 to 25 in AI-exposed occupations experienced a 16% relative decline in employment after generative AI went mainstream. The effect is concentrated in roles that historically served as career on-ramps – junior software development and customer service. Older workers in the same jobs held steady or grew.
The pattern is already visible inside the digital asset industry. Two of the sector’s largest employers, Coinbase and Crypto.com, have announced AI-driven restructurings that removed whole cohorts of early-career staff. Meanwhile, data from a16z shows roughly 1,000 jobs have shifted from crypto firms to AI startups since late 2022, and entry-level crypto hiring contracted sharply in early 2026.
If the entry-level pipeline dries up, the next generation of crypto developers, security researchers, and protocol engineers has no clear on-ramp. A 16% decline in early-career hiring sustained over five to ten years compounds into a structural talent shortage.
The study, titled Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence, was authored by Erik Brynjolfsson, Bharat Chandar, and Ruyu Chen. It relies on ADP’s payroll records covering millions of U.S. workers – not survey data or LinkedIn scraping. The researchers controlled for industry mix, geographic differences, and macroeconomic trends. The 16% decline persisted.
Jobs with high AI exposure – roles where generative AI automates tasks rather than augmenting them – saw the largest relative reduction. Junior software development and customer service positions, which had been standard entry points for career progression, were hit hardest. Companies are not cutting pay for entry-level workers. They are simply hiring fewer of them.
Workers 26 and older in the same occupations saw either no change or slight employment growth. The mechanism: generative AI automates repetitive, structured tasks, exactly the kind that junior employees historically performed while learning the basics. Senior workers handle complex, judgment-heavy work that current AI models struggle with.
Key insight: The entry-level rung is being removed, not shifted. The result is a narrower funnel into the careers that depend on those early years.
Two large crypto employers have publicly attributed layoffs to AI restructuring, matching the study’s pattern.
Coinbase cut approximately 14% of its workforce – about 700 positions – in May 2026. The company explicitly cited AI-driven restructuring as the reason. The cuts targeted roles where automation could replace entry-level functions, particularly in customer support and junior engineering.
Crypto.com implemented a 12% staff reduction in March 2026 under comparable rationale. The firm stated that AI tools allowed it to operate with fewer junior employees while retaining senior talent for oversight and complex problem-solving.
| Metric | Value | Source / Entity |
|---|---|---|
| Employment decline for workers 22-25 in AI-exposed roles | 16% | Stanford Digital Economy Lab (Nov 2025) |
| Coinbase workforce reduction | ~14% (~700 positions) | Company announcement (May 2026) |
| Crypto.com workforce reduction | 12% | Company announcement (Mar 2026) |
| Jobs shifted from crypto firms to AI startups since late 2022 | ~1,000 | a16z data |
a16z data offers a direct look at the movement. Since late 2022, roughly 1,000 jobs have moved out of crypto-focused companies into AI startups. Many of the displaced workers are not switching sectors voluntarily; crypto firms simply are not filling the entry-level roles they once did. Independent hiring data shows entry-level openings contracted sharply in the first quarter of 2026, exactly when the Stanford study predicted the largest AI displacement effect, given the lag between generative AI’s mainstream adoption (late 2022) and when companies fully restructured workflows.
The read-through for crypto is not theoretical. The sector, which already struggled to attract and retain talent during the 2022–2023 bear market, now faces a structural reduction in the pool of workers who enter the industry at the ground level.
If the entry-level pipeline dries up, the question becomes where the next generation of crypto developers, security researchers, and protocol engineers will come from. The senior talent that companies retain today started somewhere. They learned by doing the grunt work – debugging, triaging tickets, writing tests – that AI now handles.
A 16% year-over-year decline in early-career hiring does not sound catastrophic in isolation. Sustained over five to ten years, it means an entire cohort of workers who never developed the foundational skills that make senior talent valuable. The effect compounds: fewer junior workers today means fewer mid-level workers available to become senior tomorrow.
The digital asset industry depends on specialized technical talent. It is not a sector where generalized AI models can replace domain expertise required for smart contract auditing, consensus protocol development, or on-chain forensics. If the talent pipeline narrows, firms may face higher compensation costs for existing senior workers, slower output, and greater concentration risk as a handful of individuals hold critical knowledge.
ADP – the payroll provider whose data made the study possible – carries an Alpha Score of 54/100 (Mixed) in the Industrials sector, per the stock page at /stocks/adp. That score reflects a neutral read on a company whose business model depends on payroll processing for millions of workers. If hiring patterns change permanently, the data itself will reflect that change.
The next concrete marker for analysts watching the sector is whether crypto firms begin running apprenticeship or internship programs to rebuild the entry-level pipeline artificially. If they do not, the talent bottleneck will show up first in the quality and frequency of protocol upgrades and later in audit delays – both measurable indicators that the Stanford study’s finding is playing out in real time.
Prepared with AlphaScala research tooling and grounded in primary market data: live prices, fundamentals, SEC filings, hedge-fund holdings, and insider activity. Each story is checked against AlphaScala publishing rules before release. Educational coverage, not personalized advice.