Pump.fun Faces Scrutiny as Study Claims 90% of Top Traders May Be AI Bots

Study Suggests AI Bots Dominate Pump.fun Trading—But Not Everyone Agrees

A new wave of controversy has hit the meme coin world as fresh research suggests that nearly 90% of the top traders on Pump.fun may not be human at all.

According to findings from independent analyst Adam_Tehc, 93 of the top 100 accounts on Pump.fun and its sister platform PumpSwap exhibit bot-like behavior—remaining active for more than 18 hours a day. This revelation has raised concerns about the platform’s transparency, fairness, and the real nature of its skyrocketing trading activity.

Bots or Just Ultra-Committed Traders?

Pump.fun, a decentralized meme coin launchpad that recently confirmed rumors of its own token launch (PUMP), has seen explosive growth in both trading volume and user engagement. But this latest study casts a shadow over that momentum.

Adam’s methodology identifies a trader as a bot if their account shows signs of activity beyond 18 hours per day. He notes this was not a perfect science—acknowledging that a few high-performing human traders averaged close to 16 hours daily, blurring the line between human dedication and automation.

Take @Cupseyy, for example. This single account has reportedly traded more than $100 million on Pump.fun, with trading hours and frequency that raise eyebrows. But many believe such intense participation might still fall within the realm of human capability—especially in the 24/7 meme coin market.

Pump.fun’s Bot Problem Isn’t New

The new report isn’t the first red flag for Pump.fun. Blockchain surveillance firm Solidus Labs recently claimed that 98% of tokens on the platform show traits consistent with scams—adding another layer of concern to an already volatile environment.

Bot activity has long been suspected across the DeFi and meme coin ecosystems, but the scale claimed here is eye-opening. And with the highly anticipated PUMP token airdrop on the horizon, many fear that systematic bot trading could skew not only trading volume, but also airdrop distribution—raising questions about fairness and decentralization.

Skepticism Over the Data

Still, the methodology used in Adam’s report hasn’t gone unchallenged. Critics argue that defining a bot solely based on hours of activity oversimplifies a complex issue. Some traders might use automated tools for certain tasks while actively managing trades at other times. Others could be trading in shifts or across accounts.

The key question is: what defines "activity"? Is it clicking a button? Is it just having an active wallet executing trades via scripts? These distinctions matter when shaping public perception—and policy—for how platforms manage fair use and user rewards.