Interestingly enough, traders heavily rely on AI when the market is in distress, but it doesn't always turn out to be the savior everyone hopes for.

According to recent research by Nickel Digital Asset Management, almost all (96%) top managers of trading firms indicate that AI now plays a central role in their investment processes. These companies manage approximately $14 trillion in assets combined. Sounds impressive, right? But here’s the interesting part: purely automated systems have serious blind spots.

Anatoly Crachilov, founder and CEO of Nickel, stated quite clearly in an interview: AI will not save you when the market is in trouble. It’s simply not a miracle cure. What AI does well is analyzing vast amounts of historical data and recognizing patterns. But as soon as you deal with faulty or incomplete information—which is common in crypto—things go awry. Machine learning models are poor at identifying bad data, leading to incorrect conclusions.

Nickel manages a multi-manager platform with more than 80 teams, and they remain optimistic about this year despite the recent crypto dip. How do they do that? By not losing human oversight. Each manager operates within strict risk frameworks with maximum drawdown limits. When markets are truly in distress—like at the end of January—sometimes the old-school approach must prevail. That means stopping managers who exceed their limits, regardless of whether their strategy is AI-driven.

Crachilov explained that Nickel functions like a military operation. They collect more than 100 million data points from order books every 24 hours. But even with so much data, human involvement remains essential. Sometimes they call managers in the middle of the night.

The big problem? Crypto exchanges are fragile. They can go offline for 15 minutes, produce incorrect data, or deliver bad feeds. An automated AI system could blindly enforce these limitations, while a human would realize it’s just a data stream issue. A position suddenly dropping 100%? Likely a technical glitch, not a real crisis.

Charles Adams, head of Investor Relations at Nickel, emphasizes their philosophy: no single failure point in the system. One autonomous agent monitoring the entire portfolio? Catastrophic risk. That’s why they diversify their assets across more than 80 managers and hundreds of subaccounts. That diversification is their safety net.

The conclusion? AI greatly assists with risk management and sentiment analysis, but when markets are truly in distress, human intelligence is necessary. The future is likely a hybrid approach: data-driven automation with human oversight as a safety mechanism.

By the way, the STRC market is showing interesting activity with nearly $200 million tokenized on Ethereum and record trading volumes of $1.6 billion. But that’s another story.