A new academic paper from Cornell Tech highlights inherent risks in collective crypto trading schemes known as CoinAlgs, which pool user funds for automated trades. These systems face trade-offs between transparency and profitability, leaving them open to losses and insider exploitation. The findings underscore challenges in making institutional-grade strategies accessible to retail investors.
Academics at Cornell Tech have published a paper on January 2 examining Collective Investment Algorithms, or CoinAlgs, which aggregate user funds in cryptocurrency trading and execute trades automatically. The study, co-authored by eight researchers, identifies fundamental trade-offs between profitability and economic fairness in these systems.
CoinAlgs promise to bring sophisticated, institutional-level investment approaches—similar to those used by firms like BlackRock and Renaissance Technologies—to everyday retail investors. However, in the crypto space, their rise has been fueled by advances in artificial intelligence, though they often operate without the safeguards of regulated financial intermediaries. Many focus on high-risk assets, such as volatile memecoins and emerging tokens.
The researchers analyzed historical data from Uniswap, the largest decentralized exchange, to model a profitable CoinAlg that predicts asset prices accurately. Their simulations revealed that private CoinAlgs risk information leakage, allowing arbitrageurs to profit from temporary price discrepancies. Insiders with privileged access could also engage in frontrunning or similar exploitative trades.
Transparent CoinAlgs, which make their models and data public to prevent insider issues, are equally vulnerable. Arbitrageurs can easily anticipate and profit from trades, and attempts to implement defenses significantly reduce overall profits. As the researchers noted, “CoinAlgs can either be transparent, and risk losing profits; or be private, and open the door for unfair value extraction by insiders.” They added that even seemingly benign setups might hide subtle backdoors benefiting adversaries.
Despite these vulnerabilities, CoinAlgs are expected to persist, attracting users with the allure of AI-driven returns. With ongoing investments in AI by companies like OpenAI, Anthropic, and Google, their popularity is likely to grow. The paper concludes, “CoinAlgs are an inevitable part of the financial landscape,” calling for future research into protective measures.