Pump.fun built crypto’s perfect casino: then the degens found better tables Written by James Tylee from Cyber.FM
How Pump.fun went from crypto’s most-used app to one more table in the casino
There is a particular kind of silence that settles over a casino at four in the morning, after the high rollers have left and only those who cannot walk away remain. Something similar has descended on Pump.fun. The machine that turned the “degen” from a crypto subculture into the mascot of a speculative generation is still running. Its lights remain on, tokens are still launched and substantial fees are still collected. But the frenzy has faded, automated traders have become more visible and Pump.fun has moved from cultural phenomenon to one more table in crypto’s expanding casino. This is not the story of a failed platform. It is the story of a platform that lost cultural dominance whilst remaining commercially viable.
The machine
Pump.fun launched on Solana in early 2024 with a brutally simple proposition: anyone could create a tradeable token in minutes, without coding, permission or the capital normally required to seed a market. The engine was a bonding curve. Rather than using a conventional order book, each token began inside a constant-product automated market maker. Every purchase moved the quoted price higher; every sale moved it lower and successful tokens eventually “graduated” into a deeper PumpSwap liquidity pool. Buying early provided a lower entry price but never guaranteed a profit. Returns still depended on subsequent demand and the ability to exit before liquidity disappeared. The model solved crypto’s cold-start problem. Creators obtained instant distribution, buyers received immediate price discovery and Pump.fun earned fees whenever users traded. Solana supplied the speed and low transaction costs needed to make millions of small speculative bets economically possible. At its January 2025 peak, Pump.fun generated approximately $149.8 million in monthly revenue and was responsible for a large proportion of new Solana token launches. In Q2 2026, Solan has generated over $250million of revenue whereby making it: “Top Blockchain Ranking by Daily Network Revenue”, proving that the platform had not disappeared even though monthly activity was far below its peak.
The real product was attention
Pump.fun did not need most tokens to succeed. It needed users to keep creating and trading them. Every failed coin could be disastrous for its holders yet commercially productive for the platform. The launchpad earned from activity, whilst traders absorbed the losses when attention moved elsewhere. Breakout tokens served as advertisements for the next launch, convincing new participants that a few dollars could become a fortune. This was Pump.fun’s true moat. The bonding curve was easy to copy. The crowd gathered around it was not. But when switching costs are negligible, market share built on incentives rather than trust is rented, not owned. Traders had little reason to remain loyal once competitors offered lower fees, stronger liquidity or a more attractive narrative.
The odds become visible
Memecoin culture treated losses as entertainment. The “degen” identity celebrated speed, risk and the willingness to laugh at financial self-destruction. For a time, enough traders won spectacularly to keep the lottery selling itself. The underlying probabilities were far less glamorous; Solidus Labs found that 98.6% of the Pump.fun tokens it examined had fallen below $1,000 of remaining liquidity and were effectively worthless under its methodology. That figure should not be described as proof that every such token was a deliberate rug pull. It includes outright fraud and pump-and-dump schemes but may also capture abandoned experiments and tokens that simply failed to attract demand. In July 2026, research preprint reviewed 832,941 token launches between May and June and found that fewer than 0.2% attracted enough demand to graduate within 24 hours. Although the findings have not yet undergone full academic peer review, they highlight the underlying problem: creating a token had become extremely easy, but finding enough genuine buyers to sustain it remained exceptionally difficult. The market was also worse than a zero-sum game. Traders did not simply exchange profits and losses among themselves; protocol fees, creator fees, transaction costs and poor execution continuously removed value, meaning participants collectively were likely to lose money even before failed tokens and fraudulent activity were considered. Pump.fun’s current bonding curve charges a total trading fee of 1.25%, divided between the creator and protocol. The house did not need to predict which token would win. It collected revenue from the attempt.
The crack and the exodus
Pump.fun’s official X account was compromised in February 2025 and used to promote fraudulent tokens, foreshadowing the trust problem that would follow. Revenue then fell sharply from its January peak. The platform subsequently attempted to turn its success into permanent capital. In July 2025, its public PUMP token sale raised approximately $600 million in just 12 minutes, alongside a reported $400 million private sale. Instead of securing loyalty, the token sharpened questions about whether the platform, token holders and traders shared the same economic interests. Competition exposed how little loyalty existed. In July 2025, Pump.fun’s share of tokens deployed through Solana launchpads briefly fell from 88% to 19%, whilst LetsBonk and other rivals attracted creators through different fees, communities and liquidity arrangements. Pump.fun later recovered much of its launch share through fee changes and incentives, but the episode revealed that its monopoly had been built on extraordinarily low switching costs. The wider memecoin market was also contracting. Speculators migrated toward perpetual-futures exchanges and prediction markets, where liquidity was deeper, trade sizes were larger and the gambling experience appeared more sophisticated. The casino did not close. The players found tables with better odds and higher limits.
Regulation enters the casino
Pump.fun now faces a regulatory challenge as fragmented as the market it helped create. In February 2025, SEC staff said that typical memecoins resembling collectibles generally do not constitute securities whilst stressing that the conclusion depends on the economic reality of each offering. Commissioner Caroline Crenshaw challenged that approach, arguing that broad labels cannot replace the fact-specific analysis required by securities law. In the UK, the FCA placed Pump.fun on its warning list in December 2024, stating that it might be providing or promoting financial services without authorization. UK users dealing with the platform would not normally have access to the Financial Ombudsman Service or Financial Services Compensation Scheme. Pump.fun and other Solana ecosystem participants also face pending US litigation containing securities and RICO allegations. These are contested claims, not judicial findings. The distinction matters. Pump.fun should be scrutinized rigorously, but criticism loses force when allegations are presented as verdicts.
The reckoning
Pump.fun has tried to realign its economics with token holders. In April 2026, it burned approximately $370 million of repurchased PUMP (around 36% of circulating supply) and committed 50% of revenue for one year to a programmatic buy-and-burn mechanism. Yet token engineering cannot recreate cultural momentum on its own; current data still show a functioning and profitable platform. DefiLlama reported approximately $18.8 million of revenue over the latest 30-day period when checked in July 2026. Pump.fun did not die. It became normalised. That may be the more important outcome. Pump.fun proved that blockchain can remove almost every technical and financial barrier to creating a market. It also proved that removing friction does not create value, trust or lasting demand. When issuance costs approach zero, markets do not automatically allocate capital more efficiently. They can manufacture disposable assets faster than human attention can absorb them. Pump.fun demonstrated that almost anyone can create a market. It did not demonstrate that every market deserves to exist.
The next generation of crypto platforms will be judged less by how many assets they create and more by whether users leave with something more durable than a transaction receipt. The degens did not leave because the machine stopped working. They left because it worked exactly as designed, and eventually they understood who the design rewarded.