On paper, it looked like ordinary engineering work: access to internal systems, knowledge of which tokens Robinhood was preparing to list and when the announcement would go out. US prosecutors say two employees turned that knowledge into a paycheck — and have charged them with commodities fraud and wire fraud.
According to the Justice Department, Hefu Chai, 36, and Huaisong Xiang, 30, repeatedly opened perpetual futures positions on the decentralized exchange Hyperliquid throughout 2025 and 2026 — tied to the exact tokens Robinhood was about to add to its app. They allegedly opened positions hours or days before the public listing announcement, then closed out once the price jumped. Each pocketed more than $50,000, prosecutors say — a modest sum by the standards of major insider cases, but the scheme stands out for its mechanics, not its size.
Crypto insider-trading cases have typically involved buying the token itself ahead of a listing. Here, if the allegations hold, the defendants went through a derivative instead — a perpetual futures contract — never technically touching the asset their employer was about to list. Lawyers following the case note it's the first time federal prosecutors have stretched classic insider-trading logic to cover this kind of derivative. US Attorney Jamie McDonald put it bluntly: swapping the underlying asset for a derivative isn't a loophole.
There's an irony here — the blockchain's own transparency is what caught them. Every position and timestamp on Hyperliquid is public, and independent analysts spotted a pattern too consistent to be coincidence months before charges were filed. Robinhood says it ran its own internal investigation and voluntarily reported the matter to authorities, calling insider trading a zero-tolerance issue.
The defendants face up to 10 years under the Commodity Exchange Act and up to 20 years for wire fraud. For the industry, the case is a signal: even where there's no traditional order book or KYC-gated exchange, just on-chain derivatives, regulators are now willing to dig — and to find wrongdoing using the same data anyone else could have looked at.
