coindesk.com
An institutional-grade exchange executes trades at predictable prices, enforces clear limits, withstands traffic spikes and protects traders from accidentally moving the market with large orders. However, a truly good exchange is defined by how it behaves when markets break. In other words, what makes an exchange good is its ability to perform when markets get messy, not how polished it looks on the surface.
Last year, CoinDesk released its November 2025 Exchange Benchmark, with 81 exchanges scored across governance, licensing, audits, KYC, proof of reserves, uptime and other metrics. It’s a helpful reference and shows progress across the board. The number of "Top-Tier" exchanges ticked up. More platforms are now sharing evidence that they actually hold the assets they claim, undergoing audits and standardized due diligence questionnaires. It’s certainly a step in the right direction.
But credentials alone don’t answer the deeper question: Does the exchange actually behave predictably when it counts?
A platform's onboarding polish, like claims about having better foreign exchange rates over another platform, should not be confused for its ability to execute trades and handle immense volumes. Our traders asked different operational questions: whether the system can process the enormous number of price updates modern trading generates, how fills are determined, and how market data is compiled and distributed. We also asked questions to help us plan for moments of extreme volatility: Can the system handle the massive volume of price updates modern markets generate? How are fills actually determined when liquidity disappears? How is market data compiled and distributed under stress?
While many have argued that the current market cycle is for institutions, most exchanges still run market structures designed for retail flows. This includes having matching engines that favor speed over fairness, the hiding of order routing information (which institutional traders rely on to understand how their orders are being handled), leverage that disappears during moments of volatility and clearing processes that rely on trust and Telegram messages. That’s not institutional. That’s retail, dressed up.
The recent liquidation cascade on October 10 was a reminder of how quickly stress builds in these markets. Nearly $20 billion in leveraged positions were wiped out across exchanges in under 24 hours. That kind of selloff can happen in any market, even traditional ones.
But in TradFi, those moments are governed by documented, enforceable playbooks. Controls such as circuit breakers, volatility bands, or formal bust rules are all enforced. In crypto, those rules are often vague or missing. Decisions can feel improvised. And that’s the difference: not that volatility doesn’t happen, but that responses aren’t always consistent or transparent.
What matters is how exchanges behaved during the event. Did users’ orders get filled at expected prices? Were margin calls executed fairly? Was the system stable? Did the venue follow its own rules or make new ones on the fly?
A good exchange isn't defined by how it markets leverage. It’s judged by how it handles the situation when things go wrong. Stress events shouldn’t invent new risks, but they will expose the ones that were already there. Retail traders aren't the problem here. The issue is the outdated market structure from crypto's early days that aren't built to handle the scale we're seeing today. There's a reason every institutional market values depth, and a lot of that depth actually comes from retail participation. Retail traders often bring directional flow, care less about getting the absolute best price on every trade, and stick around through different market conditions. Those qualities actually support tighter spreads and more efficient markets for everyone.
But not all participants benefit equally. Market makers and high-frequency traders can thrive on high-frequency retail flow. But institutions that need to move large amounts aren’t looking to trade in $1,500 chunks. They need real depth in the market, predictable pricing and clear rules.
So the issue isn’t who’s trading; it’s how the market is structured. When institutions have to use systems built for small, mobile-first trading, where it’s unclear how orders are lined up, prices flash in and out, and trades don’t always get filled consistently, they face extra costs and uneven risks. In mature markets, exchanges control how many price updates can be sent at once so that the prices you see at the top of the screen are actually tradeable, not just fleeting quotes to gain an advantage.
That’s not about limiting retail traders; it’s about keeping the system honest. Both retail traders and institutions can thrive together, but they need infrastructure built for both. That’s what makes an exchange genuinely institutional. The CoinDesk benchmark is a step forward. It puts data on the table. It raises the floor. And for market participants doing their diligence, it’s a useful starting point. But we can’t confuse scorecards with enhancements to underlying market structure.
We can’t call something institutional just because the users are. A good exchange is boring. Transparent. Battle-tested. Built for size. That’s the real benchmark. Most crypto exchanges want to be treated like institutional platforms. Some even aim to rival the credibility of the exchanges and prime brokers that dominate traditional finance. But structure is what earns that status, not branding.
The early tech lead and global reach crypto platforms have built is real. But it can’t be sustained without deeper investment in the infrastructure that institutional participants actually rely on: clean execution, observable credit and predictable behavior under stress. That’s what separates a trading venue from a good exchange.