Mon, May 25

Onchain Perps Solved Speed. The Next Question Is Institutional Plumbing

Market News

In 2026, the argument over whether onchain perpetuals could match centralized exchanges on execution has been settled in practice. Hyperliquid alone now accounts for roughly 44 percent of perp DEX volume, turns over more than $50 billion in derivatives a week, and routinely fills orders at sub-second latency on a Layer 1 designed specifically for trading. Critics who insisted, only two years ago, that decentralized infrastructure could not handle serious derivatives flow have mostly stopped insisting.

What still keeps the bulk of institutional capital out of these venues has less to do with engineering than with workflow. Quant desks and top-tier market makers run their order flow through FIX gateways tied into the same systems they use for Binance, OKX, the CME, and the legacy equities and FX venues. Maintenance margins on professional desks run well below the five-percent buffers most DeFi venues still apply to liquid pairs. Fee structures need to accommodate quotes that get cancelled and replaced at HFT cadence without a per-action network toll. Unrealized PnL has to be usable in real time, not parked until exit. None of these are new requirements. They are the basic operating assumptions of any sophisticated derivatives book, and onchain venues have, with limited exceptions, not met them.

Several perp-focused chains have begun building toward that gap. The most recent isAFX, a sovereign Layer 1 for derivatives that went live on mainnet on 18 May. Its launch is one of the clearer signals of where the next phase of competition is heading.

AFX is not pitching itself as faster than Hyperliquid. Its publicly cited figures, over 100,000 TPS and around 100ms median latency, sit in roughly the same performance band as the incumbent. The differences are in the layer beneath the order book, and they are aimed at a specific audience.

The headline difference is native FIX protocol support. FIX is what professional trading firms already use to connect to Binance, OKX, the CME, and the legacy equities and FX venues. Asking those firms to rewrite their order-management stack for a custom DEX API has been a consistent reason institutional flow has stayed inside centralized exchanges, and AFX claims to be the first decentralized derivatives exchange to offer FIX at protocol level. That claim will need pressure-testing once the platform matures, but the underlying observation is sound. The integration gap has been visible in the sector for years, and few perp DEXes have built specifically around it.

Margin requirements are another visible difference. AFX is launching with a 1.25 percent maintenance margin and what it describes as real-time re-utilization of unrealized profits. Mainstream perp venues, both decentralized and centralized retail-facing, typically run closer to five percent maintenance on liquid pairs. That is roughly a fourfold difference in capital efficiency at equivalent notional, which on a leveraged book is what determines how much risk a desk can carry at any one time.

Then there is the fee structure. Trading on AFX is gasless at the execution layer, which removes the cost penalty that has historically punished market makers who quote aggressively and cancel often. Hyperliquid solved this inside HyperCore. AFX takes the same idea and pairs it with a commitment to route the full network revenue back to participants. The protocol launched without venture allocation or private rounds, which puts its alignment structure close to Hyperliquid’s and reflects one of the more credible community-side signals in DeFi over the past 18 months.

None of this answers whether AFX will gain traction. The more useful questions sit further down the line.

One is liquidity. Hyperliquid’s lead is not really about consensus design; it is built on order books, market maker presence, and open interest that took eighteen months to accumulate. AFX is opening with BTC, ETH, gold, and crude oil perpetuals at up to 40x leverage, a sensible product mix for traders hedging across both crypto and traditional commodities, but order-book depth has to be built through real flow. The six-month metrics will tell more than the launch deck.

There is also the rest of the field. The perp DEX sectorpulled in more than $1.8 trillion in a single recent quarter, and the venues competing for that flow, including Aster, Lighter, edgeX, Paradex, and Pacifica, are moving quickly. Each is pulling on different levers, whether higher leverage, aggressive incentive programs, or niche product design, and the field shifts month to month. AFX’s combination of FIX support and tight margin requirements gives it a clear position, but the advantage only holds if those technical features translate into actual quant-desk integrations before someone else builds the same stack.

The sovereignty claim also deserves a careful read. Performance-focused L1s of this kind tend to launch with concentrated validator sets and ad-hoc intervention surfaces, concerns that have already followed Hyperliquid through 2025 and 2026. The fact that AFX launched without venture backing addresses one part of the alignment question, but it says nothing about validator decentralization, governance handover, or operational transparency. Those details deserve close reading before anyone equates “sovereign” with “trust-minimized.”

What the launch signals, more than the product itself, is where the sector’s attention is shifting. The first competitive phase of decentralized perpetuals was about catching up to centralized exchanges on the obvious metrics — latency, throughput, fee depth, interface quality. That work is largely done, and Hyperliquid did most of it. The next phase is about pulling in the trading flow that has stayed inside centralized venues for less obvious reasons, integration friction and margin inefficiency among them.

AFX is one answer to what that audience needs: native FIX, low maintenance margins, no general-purpose chain overhead, and a token structure that does not pre-allocate value to investors who will eventually exit. Whether the bet pays out will be readable from the order books over the next twelve months, not from launch coverage. But the question AFX is putting in front of the sector — what does an onchain derivatives venue need to add now that latency parity has been achieved — is the right one for this stage of the market.