Big crypto traders are quietly looking for a way off public blockchains — and a few firms are building one.
Why it’s happening
- In traditional markets, large traders hide orders in dark pools and off-exchange venues to avoid moving prices or broadcasting long-term strategies. As of January 2025, Bloomberg reported that more than half of U.S. equities trading occurred off public exchanges.
- Crypto has no true equivalent. Onchain trades and order books are visible to anyone, and data aggregators like DeFiLlama and Arkham make that visibility easy to analyze. That transparency creates a new problem: the bigger you are, the more obvious your moves — and the faster your “alpha” gets copied.
How this plays out
- Market makers providing liquidity on public decentralized exchanges say their strategies are frequently reverse-engineered. Denis Dariotis, co-founder of GoQuant (a crypto trading infrastructure firm backed by GSR), says one top market maker told them they must rotate strategies every three weeks to stay ahead of copycats. “That’s the alpha problem,” Dariotis said.
- Public onchain activity also exposes firms to reputational risk and regulatory scrutiny. High-profile tracing of institutional trades — for example discussions around Jane Street’s activity during the Terra/Luna collapse — can quickly become public narratives that firms must manage despite their trades being routine on private venues.
The proposed fix: private matching onchain
- GoQuant is launching GoDark, a DEX on Solana scheduled to start in May, that aims to hide trade details using zero-knowledge proofs. The platform is designed to conceal orders from other traders and even from the node operators running the matching engine — “a matching engine where nobody in the system can see what they’re matching,” as the company puts it.
- In internal tests, GoDark’s order matching runs in roughly 25–50 milliseconds. That’s faster than many current DEXes, which often see execution times in the hundreds of milliseconds, but it’s still an order of magnitude slower than the near-instant speeds available to firms co-located with centralized exchanges. For retail traders that latency gap may not matter; for the professional market makers GoDark hopes to attract, it could.
Liquidity and business-model challenges
- Privacy alone isn’t enough — a private exchange with no volume is just a “dark room.” GoDark plans to seed liquidity using a model similar to Hyperliquid’s HLP vault, where users deposit funds that are deployed as market-making liquidity and participants earn fees and early access to liquidations.
- That bootstrap approach worked for Hyperliquid, but many DEXes that relied on similar incentive programs saw volumes fall sharply when incentives ended. Attracting sustainable, long-term liquidity remains the harder commercial question.
Regulatory trade-offs
- Traditional dark pools conceal pre-trade details but still operate under post-trade reporting requirements and regulatory oversight. GoDark’s design, by contrast, is structurally incapable of producing a complete public audit trail. The team has added automated OFAC screening as a compliance measure, but that may be insufficient for regulators who, in recent years, have pushed crypto toward greater transparency rather than less.
- How regulators will respond — and whether truly private onchain trading will be limited to jurisdictions with lighter oversight — is an open question.
Product positioning
- The May launch is the retail-facing version of GoDark. The project is separate from GoQuant’s institutional product with the same name — a spot DEX built with Copper and GSR that enters production next month and targets a narrower institutional client base.
Bottom line
Crypto’s transparency, long touted as a feature, is becoming a bug for large liquidity providers. Projects like GoDark aim to recreate the privacy of TradFi dark pools onchain with zero-knowledge tech, but they face technical speed limits, liquidity economics, and likely regulatory scrutiny. Whether that’s enough to lure sustained institutional participation remains to be seen.
Read more AI-generated news on: undefined/news
