When crypto markets turn rocky, AI is a powerful tool — but not a miracle cure, says Nickel Digital When crypto prices plunged at the end of January, many trading firms turned to artificial intelligence to help triage losses and seize opportunities fast. That reliance has become widespread: a recent Nickel Digital Asset Management survey found 96% of executives at a group of trading firms — which together manage about $14 trillion in assets — say AI already plays a major role in core investment processes. But AI isn’t a panacea, Anatoly Crachilov, founding partner and CEO of Nickel, warns. “It’s a very tough market. AI will not save you; it’s not a savior,” he told us. While machine learning and predictive models are reshaping quantitative trading and go far beyond the consumer-facing large language models, they still struggle with bad or misleading data and can reach wrong conclusions if fed incorrect inputs. Where AI shines, Crachilov says, is risk management and sentiment-driven, data-based models that learn how to position portfolios under stress. It’s less effective at beating ultra-fast “sniper” bots that capitalize on low-liquidity crypto tokens; those require different tech and speed. Nickel — a London-based multimanager platform that allocates to more than 80 teams — nonetheless remains upbeat about the year despite the late-January slump. “Perhaps an achievement in its own right,” Crachilov noted. Nickel’s approach mixes heavy automation with disciplined human oversight. Every manager on the platform works within strict risk frameworks, including maximum drawdown limits that tighten during volatile periods. If managers breach those limits — whether their strategies are AI-driven or not — human teams will step in. “Sometimes you have to exercise discipline and stop those managers who break [max drawdown] limits,” Crachilov said. “Ultimately, there is a hard stop on how much pain we would allow in the portfolio.” Operationally, Nickel runs what Crachilov describes as a “military-style operation,” ingesting more than 100 million data points from the underlying book every 24 hours. Even so, that data flow “still requires human involvement,” he said — teams stay in contact with managers around the clock. The need for a human overlay is not theoretical: crypto exchanges can produce timeouts, wrong data, or transient “patches” of bad feeds. A human analyst can often spot that a position supposedly down 100% is a data glitch, whereas an automated system might mechanically trigger limits and shut down strategies unnecessarily. That concern about single points of failure underpins Nickel’s risk philosophy. “If there was one autonomous agent which is monitoring the whole portfolio, let's say something goes wrong with it, the risks could be potentially catastrophic,” Charles Adams, Nickel’s head of investor relations, said. Nickel’s defense is diversification: the fund is split across more than 80 managers and across hundreds — if not thousands — of subaccounts on exchanges, which helps reduce the risk that any single failure cascades through the portfolio. Bottom line: AI is accelerating and deepening how crypto trading teams analyze markets and manage risk, but in the fragile and noisy world of crypto data, human judgment remains a critical backstop. Read more AI-generated news on: undefined/news