Perp DEX capital efficiency: active markets per $M TVL, live
Active perp markets divided by trading TVL in USD millions. Higher means more instruments supported per dollar of locked collateral. Synthetic oracle venues like Gains carry many markets on a shared pool; CLOB venues like Hyperliquid require dedicated market-maker inventory per pair.
TL;DR. As of , Ostium leads markets per $M TVL at 34 (24h at the average daily rate) across 12 ranked venues. Source: OpenChainBench, https://openchainbench.com/benchmarks/perp-capital-efficiency.
Read this carefully
TVL from DefiLlama /protocol/{slug} currentChainTvls, excluding staking and pool2 categories. Market count from each venue's native API (same source as the perp-active-markets bench). The ratio reflects architectural choices, not absolute liquidity: a high ratio means the venue can list many markets cheaply, not that each market has deep liquidity. Compare with perp-volume-share for the liquidity angle.
Every perp DEX comparison ranks by volume, fees or funding rates. Few ask the capital question: how many markets can a venue sustain per dollar it locks up? The answer reveals the structural gap between synthetic oracle venues and CLOB order book venues. A synthetic venue needs one shared collateral pool to back unlimited pairs. A CLOB like Hyperliquid needs market-maker inventory for every individual pair. The collateral requirements scale completely differently.
Ostium leads at 34 markets per $1M of trading TVL across 12 venues; Hyperliquid reads 0.020. The gap is architectural, not a performance difference. The bench does not judge which model is better: deep single-pair liquidity has real value. It measures a real dimension that existing leaderboards ignore entirely.
Methodology
The bench computes active_markets / (trading_tvl_usd / 1_000_000) for each tracked perp venue. Both inputs are updated every 5 minutes. Active market count comes from each venue's native API via the perp-cohort-stats harness (same source as the perp-active-markets bench). Trading TVL comes from DefiLlama /protocol/{slug} currentChainTvls with staking and pool2 keys excluded, also polled every 5 minutes. The ratio is recomputed on every Prometheus scrape. TVL fluctuates with collateral deposits and crypto prices, so the 24h average smooths intraday noise.
- Cadence: every 5 minutes. Both active_markets and TVL are re-read on the same sweep cycle.
- Active markets: native venue API per the perp-active-markets bench methodology. Gains: its Arbitrum trading variables minus delisted pairs. Hyperliquid: info metaAndAssetCtxs. dYdX: indexer perpetualMarkets. Others: native endpoints.
- TVL: DefiLlama /protocol/{slug} currentChainTvls. Staking and pool2 keys excluded. Multi-chain TVL summed across all deployment chains.
- Gains TVL: GToken vault across Arbitrum, Base, Polygon and ApeChain. This is the shared liquidity pool that backs all synthetic pairs simultaneously.
- Hyperliquid TVL: total user deposits (USDC). Funds serve as both margin and implicit market-maker collateral; every listed pair draws from this pool for maker fills.
- Ratio formula: avg_over_time(markets[24h]) / avg_over_time(tvl_usd[24h]) * 1_000_000.
- Floor. A venue ranks only with at least $1M of trading TVL. The metric divides by TVL, so a small vault inflates the ratio: one market on $120k of backing reads 8.4 and ranks fourth, above four venues carrying 66 to 334 markets, which inverts what a high number means here. The threshold sits in a gap in the data: on 2026-10-22 the smallest ranked vault was $3.95M, so no row in the field moves.
- Interpretation: a ratio of 40 means the venue can sustain 40 active markets per $1M locked. Higher is more capital-efficient in terms of market breadth.
- Cohort: every venue of the perp-cohort-stats harness that publishes both an active market count and a DefiLlama TVL, 12 venues as of 2026-09-22 (Gains, Hyperliquid, dYdX v4, Ostium, Aster, Extended, Lighter, Paradex, edgeX, Pacifica, Aevo, GRVT). Ondo Perps, Polymarket and Vertex have a market count but no TVL series; GMX v2 has a TVL but no native market count. The four are left out.
Results: 12 venues ranked by Markets per $M TVL (24h at the average daily rate)
| № | Venues | Total | Success |
|---|---|---|---|
| 1 | Ostium | 34 | 99.97% (n=2880) |
| 2 | Gains | 27 | 99.97% (n=2880) |
| 3 | Aevo | 16 | 83.02% (n=2880) |
| 4 | GRVT | 6 | 83.33% (n=2880) |
| 5 | Extended | 3 | 99.88% (n=2880) |
| 6 | edgeX | 3 | 83.17% (n=2880) |
| 7 | Paradex | 3 | 99.99% (n=2880) |
| 8 | Aster | 0.824 | 99.99% (n=2880) |
| 9 | dYdX v4 | 0.736 | 99.98% (n=2880) |
| 10 | Lighter | 0.268 | 99.91% (n=2880) |
| 11 | Papertrade | 0.077 | 80.00% (n=2880) |
| 12 | Hyperliquid | 0.020 | 99.99% (n=2880) |
Frequently asked
What does capital efficiency mean here?
It measures how many distinct perpetual markets a venue can sustain per $1M of trading collateral locked. Higher is more efficient: you get more instruments per dollar of capital. It does not measure depth per market.
Why do oracle venues score so high on this ratio?
Gains uses synthetic oracles backed by a single shared liquidity pool (the GToken vault). Adding a new pair costs near zero additional capital. CLOB venues like Hyperliquid or GMX need per-pair market-maker inventory or per-pool collateral, so capital scales with pair count.
Is capital efficiency always desirable?
Not necessarily. Deep per-pair liquidity has value for large traders. A venue with hundreds of pairs on a few million dollars of TVL may have thin liquidity per pair. Pair this bench with perp-volume-share for the depth angle.
Why is TVL from DefiLlama and not native APIs?
DefiLlama aggregates TVL across all deployment chains and normalizes the methodology across venues. Native TVL endpoints (where they exist) differ in scope: some include user margin, others only LP collateral.
Does TVL change with crypto prices?
Yes. Venues whose TVL includes crypto-denominated assets (LP tokens, staked tokens) see their USD TVL move with prices. The 24h average smooths intraday moves, but the ratio can shift in volatile markets.
Source code github.com/ChainBench/OpenChainBench/tree/main/harnesses/perp-cohort-stats