Capital only. We track vaults here and read no agent rail, so nothing on this page speaks to whether agents transact on this network. Counted 7 minutes ago.
The venues we read on this network, and what each holds. Agent-native means the protocol was built for agent-run strategies rather than having them arrive later.
| Protocol | Kind | Vaults | Capital | Share |
|---|---|---|---|---|
| Hyperliquidsite | perp vault | 306 | $424.61M | 95% |
| D2 Financesite | strategy vault | 4 | $17.27M | 4% |
| ether.fi Liquidsite | strategy vault | 1 | $3.09M | 1% |
| Upshiftsite | strategy vault | 1 | $958.3k | 0% |
| Vault | Capital | Risk |
|---|---|---|
| Hyperliquidity Provider (HLP) Hyperliquid | $183.39M | unscored |
| HLP Strategy X Hyperliquid | $100.08M | unscored |
| HLP Liquidator 2 Hyperliquid · agent-managed | $30.00M | 1 |
| [ Systemic Strategies ] L/S Grids Hyperliquid · agent-managed | $18.04M | 41 |
| drkmttr Hyperliquid · agent-managed | $16.05M | 31 |
| D2 Finance · D2HYPE D2 Finance | $11.47M | unscored |
| [ Systemic Strategies ] ♾️ HyperGrowth ♾️ Hyperliquid · agent-managed | $9.88M | 29 |
| Growi HF Hyperliquid · agent-managed | $9.42M | 31 |
Read from each vault's own contract, not from what a source called the pool. Where the two disagree the reported name is shown beneath — for a vault, an aggregator's pool symbol is usually the vault's own receipt token rather than the asset behind it.
| Asset | Vaults | Capital |
|---|---|---|
| USDC reported as LIQUIDHYPE | 307 | $427.70M |
| WHYPE reported as D2HYPE | 1 | $11.47M |
| UBTC reported as HBBTC | 2 | $4.06M |
| USD0 reported as HYPE-- | 1 | $1.97M |
| UETH reported as DGNUPSIDE | 1 | $730.1k |
The operators behind the vaults on this network, by capital under management here.
Capital is the total under vaults we track on Hyperliquid, not the network’s total value locked — we read agent-managed strategies and a comparison set, never the whole chain. Everything here is computed from records we hold; where a figure is missing this page says which kind of missing it is rather than printing a dash. See Methodology for how far back each reader reaches and whose limit it is.