26 Crypto Layoffs, 9 Attributed to AI — Only 5 by the Companies Themselves

This article is based on Northpoint's "State of Crypto Marketing 2026" report. All data comes from public information, and the full report is free and requires no registration.

"AI is stealing jobs" is the easiest headline to write and the easiest to spread in 2026. But when Northpoint tallied the crypto marketing ecosystem, it didn't simply repeat that line. Instead, it did something more tedious but more useful: it categorized every public layoff one by one, then distinguished — which ones the companies themselves said were caused by AI, and which ones the media said were.

The Numbers: 26 Layoffs, Only 5 Attributed to AI by the Companies Themselves

The report covers the twelve months ending August 31, 2026, with a sample of 27 crypto companies (11 exchanges, 8 foundations, 5 wallets, 3 other licensed institutions) and 17 agencies. All data comes from public sources — no interviews, no anonymous sources.

The key figures are as follows:

• 26 layoffs on record.
• Of these, 9 were framed in some narrative as "AI-caused layoffs."
• Of those 9, only 5 were explained by the companies themselves using AI; the other 4 were frameworks the media constructed on their own.
• Not a single one attributed layoffs to the marketing department.

This data points to an easily overlooked distinction: in reporting on "AI replacing jobs," some of it is actually about headlines, not about the layoffs themselves. This doesn't mean AI hasn't changed how work is done — it means layoff announcements are one of the worst samples for measuring AI's impact.

A Stronger Signal: Senior Marketers Are Going to AI Companies

If you only look at layoff announcements, you'll miss talent movement in the other direction.

The report documents a fact that can be verified item by item through public records: seven senior Coinbase marketers, including the CMO, joined OpenAI and Anthropic. This isn't a vague "according to people familiar with the matter" — it's name-level movement traceable through public role changes and announcements.

This trend of talent flowing from crypto to AI is more persuasive than claims that "AI replaced a certain role." The layoff narrative is defensive and passive; talent actively moving to AI companies reflects practitioners' judgment about where skill value is shifting — marketers aren't staying put and learning AI tools, they're moving directly into building and promoting AI products.

What the Report Deliberately Doesn't Count

For a responsible report, what it refuses to calculate matters just as much. Northpoint explicitly lists three boundaries:

It doesn't count hiring numbers. The career pages of the largest exchanges can't be fully read by scanning tools, so the report ultimately gives no hiring figures at all. A position you can't find doesn't equal a vacant position — this principle avoids using "zero hiring" to infer "jobs disappearing."

Agency claims aren't contract facts. Agencies proactively list client names, but the companies named almost never confirm. The report therefore labels these as "claims," not facts.

A single frozen dataset. All numbers come from the same file set, reconcilable item by item with the appendix, and the claims ledger is also public in a companion repository. This means anyone can verify the source of every judgment.

Why career pages can't be scanned. Modern exchanges and large tech companies' career pages generally rely on JavaScript dynamic rendering, with job listings pulled asynchronously from backend APIs through frontend frameworks rather than written statically in HTML. This means conventional crawlers often capture only an empty shell page. In addition, many companies use third-party recruiting systems like Greenhouse, Lever, and Workday, and jobs are dynamically filtered by region, department, and permissions — what you see without logging in is only part of all positions. Getting complete data usually requires reverse-engineering their internal APIs or simulating login, both of which carry high cost and compliance risk. This is precisely the technical reason why "zero hiring" can't be directly equated with "jobs disappearing": not being able to scrape it doesn't mean it doesn't exist.

Implications for "Tracking AI's Employment Impact"

The report's methodology is more worth learning from than its conclusions. If a team wanted to build a system to track AI's impact on employment, the most tedious rules might be the most useful:

Record what companies say themselves and what others say about them separately, then print out the gap.

Of 26 layoffs, 9 were tagged with AI, but only 5 came from the companies' own mouths — that gap of 4 is the space between headline narrative and factual data. Most statistics on "AI stealing jobs" don't distinguish these two layers, flattening that space directly into certain causation.

Northpoint's full report is free, requires no registration, and lists 34 sources supporting its conclusions. For those working at crypto companies, the report page also offers an entry point: submit your company's site, and within a day you'll receive a hand-written one-page analysis comparing your company's position against the report's seven figures.

The attribution problem: correlation isn't causation. Layoffs are often the result of multiple overlapping factors — tightening financing conditions, business restructuring, M&A integration, rising compliance costs — all of which can occur at the same time as "introducing AI tools for efficiency." When a company simultaneously announces "embracing AI" and "optimizing workforce structure," outsiders easily mold two independent things into one causal chain. Rigorous attribution requires distinguishing three layers: the reasons companies themselves state, the net effect after excluding other variables, and the causal framework in media narratives. Northpoint extracts the first layer separately into a verifiable ledger, precisely so the latter two layers don't blend into the factual layer. This approach is a useful reference for anyone studying AI's employment impact: fix the evidence layers first, then talk about trends.

One-line Summary

Of 26 crypto layoffs, only 5 were attributed to AI by the companies themselves; 4 were media narratives — talent flowing to AI companies is the stronger signal.

Paragraph Summary

Northpoint's "State of Crypto Marketing 2026" report tallied public layoff data from 27 crypto companies and 17 agencies over the twelve months ending August 31, 2026, recording 26 layoffs, of which 9 were framed as "AI-caused" — but only 5 came from the companies themselves, while the other 4 were media-constructed frameworks, and none attributed the cause to the marketing department. The report argues that layoff announcements are one of the worst samples for measuring AI's impact. A more credible signal is talent movement: seven senior Coinbase marketers, including the CMO, joined OpenAI and Anthropic. The report deliberately doesn't count hiring numbers, nor treats agency client claims as facts; all data comes from a single frozen dataset and can be reconciled item by item. Its methodological takeaway: record separately what companies say themselves and what others say about them, and make the gap public.