The recordings surfaced first. A March meeting, co-founder Noah Tweedale explaining to staff that Pump Fun had "grown too quickly" and could no longer move "fast and rough." Layoffs followed in April. Then, in mid-June, the remaining employees signed a token agreement that scheduled a quarter of their PUMP allocation for unlock two months later. Then the firings continued. Over 40 staff, per one employee-backed account, cut in the last two months — several, the account claims, exactly one day before their vesting cliff. One employee. One day. The code spoke, but the logic was a lie.
Pump Fun is the memecoin launchpad that turned token creation into a slot machine. Cumulative revenue: over $1 billion. Peak headcount: roughly 100. Its PUMP token is down almost 76% from its September 2024 high. Its UK parent, Baton Corporation, has accounts overdue at Companies House by more than a month — a £375 oversight for a firm sitting on nine figures of revenue. It has been 365 days since Pump Fun promised an airdrop was "coming soon." Coming soon, in crypto, is a flexible interval. But the accumulation of deferred obligations — airdrop, vesting, filings — forms a pattern.
This is not an isolated event. Crypto firms have spent 2025 firing people. Coinbase cut 14% in May, blaming market conditions and AI. Gemini reduced staff by 25% in February, citing AI. Block fired about half its staff, citing AI. The industry has developed a standardized vocabulary for layoffs: AI, automation, macro headwinds. Pump Fun's explanation is different. "Grew too quickly" is not a market thesis; it is an admission of internal failure. It is also, on the surface, more honest than the AI euphemism. That surface is deceptive.
Let me present the sequence as a due diligence problem, because that is what this is. A company issues a token. It grants employees allocations with a vesting schedule. The schedule provides for a quarter unlock after a defined period. The company retains the right to terminate employment. The company exercises that right. The termination precedes the unlock. The employee forfeits.
In my audits of token distribution contracts, I have seen this design repeated across dozens of projects. The Solidity is clean: a cliff timestamp, a released mapping, a release() function that checks block.timestamp against the schedule. The code is rarely the problem. The problem is that the allocation is conditioned on an off-chain variable no smart contract can enforce — continued employment. Employers can fire. Vesting can forfeit. The contract between company and employee is, in practice, unilateral.
This is not an accident. It is the architecture of retention. Token vesting exists to align incentives, to make employees behave like long-term owners. The alignment holds only if the employer treats the schedule as a commitment. Pump Fun, per the recordings, wanted to move "fast and rough." That phrase is the tell. Speed and roughness are the opposite of commitment. Trust is a variable you cannot hardcode.
The mid-June token agreement matters. A quarter of the allocation unlocking two months later means those recipients were close to a liquid payout. The employee account claims the layoffs hit one day before that vesting period unlocked. One day. Whether that day was deliberate optimization or coincidence is irrelevant; the structural outcome is identical. The company captured the employee's labor, then canceled the consideration. The worker took on the liability without the asset.
In traditional equities, this sequence would trigger litigation. An employee terminated before vesting can sue, and discovery would expose whether the timing was engineered. Crypto has no such mechanism. Token agreements are drafted in Delaware or the British Virgin Islands, governed by arbitration clauses, and enforceable only if the employee can outspend a treasury backed by billions in cumulative fees. The asymmetry is not a bug in the code; it is the feature. Regulatory arbitrage extends to human resources.
Now, the overdue filings. Baton Corporation's accounts, dated to September 30, 2025, have not been submitted. The penalty is £375 beyond one month, £750 beyond three, £1,500 beyond six. These are rounding errors against $1 billion in cumulative revenue. The fine does not matter. The pattern does. A Companies House filing is a mechanical obligation — at most an afternoon of an accountant's time. Failing to do it while conducting layoffs indicates an operations team that is either underwater or indifferent. In my experience, that distinction does not matter either; the result is systemic sloppiness. Data does not lie, but it does not care.
The industry's other layoff narratives are worth contrasting. Coinbase said AI. Gemini said AI. Block said AI. These are forward-looking stories: we are replacing humans with software. Pump Fun said "grew too quickly." That is a backward-looking admission: we scaled, we broke, we are cutting. The AI narrative at least points to a future; the "fast and rough" narrative points to an internal failure of planning.
The deeper issue is what the layoffs reveal about Pump Fun's tokenomics. The PUMP token is down 76% from its high. The platform has over a billion in revenue. Taken together, these facts mean the token is priced as a claim on future speculation, not on past earnings. The employees' tokens were going to be sold. A wave of insider supply hitting a declining market is price-negative. Removing those employees — and their tokens — reduces future sell pressure. The firings, in that light, are not just cost-cutting. They are supply management. They built a palace on a fault line.
A broader lesson applies to every project that posts a vesting schedule in its documentation. The schedule is not a promise; it is a conditional instruction set. It says: if you remain employed, you receive tokens. It does not say: you receive tokens in consideration for work performed. The distinction is everything. A token allocation is not a wage; it is an option on the company's future generosity. And options, as every auditor knows, expire worthless when the counterparty decides they should.
To be fair to the bulls: the revenue is real. Pump Fun solved a genuine problem — reducing the cost and friction of token issuance to near zero. It generated a billion dollars in fees. That is not nothing; it is arguably one of the few functioning businesses in crypto. The layoffs, in isolation, could be rational management. Companies grow too fast. Corrections happen. Firing people one day before vesting, though, is a different category.
Also, the overdue filing could be a non-event. Companies House deadlines slip. A £375 penalty for a company with Pump Fun's cash flow is negligible. The token decline could be a repricing to fair value, not a signal of insolvency. The company may simply be tightening operations to survive — which is what companies are supposed to do. That is the strongest bull case. It is also the most dangerous one. It requires accepting that "fast and rough" is a management philosophy rather than a confession, and that the one-day-before-vesting termination was a scheduling coincidence. The code does not care about your beliefs.
The question is not whether Pump Fun will survive. It has revenue. It will likely survive. The question is who absorbs the risk next. Employees have been de-risked. Token holders are holding a 76% drawdown. Company filings are overdue. When a platform tells you it moves "fast and rough," believe it. Then verify who is standing in front of the train.