The Delta-Neutral Book That Actually Survives: Pyramiding Funding Trades on Hyperliquid

Every desk that has run a serious perp book knows the seduction of the funding trade. It looks like free money: hold the spot, short the perp, collect the funding, sleep well. But the traders who actually keep that yield are not the ones who found the fattest funding rate — they are the ones who understood exactly where the position can break, and who scaled in with discipline instead of backing up the truck on day one. On Hyperliquid, where funding settles every hour and collateral is unified in USDC, both the upside and the failure modes are sharper than on a legacy 8-hour venue. This is a working note on how a Hyperliquid delta neutral strategy is actually constructed, why pyramiding into it is a risk decision and not a return decision, and the specific mechanics that decide whether your "hedged" book is really hedged.

None of this is a signal. It is a description of structure. If you cannot state your own liquidation price and your own funding-flip breakeven from memory, you are not running a delta-neutral book — you are running a directional book with extra steps.

Key points before you scale in

• Delta neutral ≠ risk neutral. Netting price exposure to roughly zero removes directional PnL, but leaves funding risk, liquidation risk on the levered leg, basis risk, and venue/collateral risk fully intact.

• The yield is a funding-rate capture, not an arbitrage guarantee. You are being paid because perp longs are structurally willing to pay to hold leverage — a flow that can reverse without warning.

• Hyperliquid pays funding hourly, computed from a premium index with a fixed interest component, and caps the rate at 4% per hour — mechanics that change your monitoring cadence versus 8-hour CEX venues.

• Pyramiding is position-sizing discipline, not alpha. Scaling in on scaled (decreasing) clips keeps average liquidation distance manageable; adding aggressively into a persistent-looking rate is how funding books die on the reversal.

• The perp leg is what liquidates. Your spot can be perfectly fine while a thin isolated short gets taken out on a wick — at which point you are naked long into the move you were trying to hedge.

What does "delta neutral" actually mean on Hyperliquid?

The construction is the standard cash-and-carry / basis trade, adapted to Hyperliquid's account model. You hold a long spot position in an asset and open an offsetting short perpetual of the same notional, so that a move in the underlying produces roughly equal and opposite PnL on the two legs. Net directional delta is close to zero; what remains is the funding stream the short perp collects when funding is positive (the common case in a market biased long). When funding is negative — perp trading below spot, shorts paying longs — the mirror construction is long perp against short spot, though borrowing spot to short is the harder leg to source for most desks.

On Hyperliquid the collateral picture is unusually clean: USDC is the universal collateral asset, and every perpetual position and HyperCore margin balance is denominated in USDC. That removes a layer of cross-collateral mismatch you would juggle on some CEXs. It does not remove the venue question for the spot leg. HyperCore has native spot markets — HYPE against USDC being the deepest — so a HYPE basis trade can live entirely on-venue with a single margin account. For assets without a deep native spot book, the long leg typically sits on a CEX or in a wallet while the short perp sits on Hyperliquid, which reintroduces cross-venue settlement, transfer latency, and two separate liquidation surfaces. That structural choice — one venue or two — is the first thing that decides how fragile the position is.

Where does the yield actually come from?

Funding is the mechanism that tethers the perp price to the underlying by periodically transferring payments between longs and shorts. On Hyperliquid the funding rate is defined as the average premium index plus a clamped interest term: F = premium index P + clamp(interest_rate − P, −0.0005, 0.0005). The premium is sampled every five seconds and averaged over the hour, and the interest component is fixed at 0.01% per 8 hours (about 0.00125% per hour). The realized funding payment at settlement is position size × oracle price × funding rate, exchanged peer-to-peer between traders with no protocol cut.

Two mechanics matter for a delta-neutral operator. First, Hyperliquid settles funding every hour, not every eight. A 0.01% hourly rate is roughly the economic equivalent of a 0.08% 8-hour rate on a legacy venue, but you receive it in hourly increments — which means faster compounding, faster feedback, and a shorter window before a regime change starts costing you. Second, the rate is capped at 4% per hour. That cap protects you from an unbounded funding bill on the wrong side, but it is also a reminder of how violent the tails can be: a rate anywhere near the cap signals a market so one-sided that your "boring" hedge has become a crowded trade.

Observation: reported net APRs on delta-neutral funding capture cluster around 3–12% on BTC and ETH and 20–60%+ on mid-cap and long-tail pairs, before fees. Interpretation: those long-tail numbers exist precisely because those markets are thin, newly listed, and prone to funding whipsaws — the yield is compensation for tail risk, not a free lunch. Treat headline APRs as regime-dependent and mean-reverting, not as a run-rate.

How do you build the position without leaving a gap?

The execution goal is to enter both legs at matched notional with minimal time between them, because any window where one leg is on and the other is not is naked directional exposure. In practice that means sizing the spot buy and the perp short to the same USD notional, accounting for the fact that the perp leg consumes margin while the spot leg is fully funded. A useful discipline is to leg in on limit or TWAP-style orders rather than crossing the spread with market orders on both sides — slippage on entry is a permanent cost that eats weeks of funding before you have earned anything.

Margin mode is the next decision. Cross margin generally suits a hedged pair on a single venue, because unrealized PnL on one position is available as margin, and a hedged book wants one leg's gains offsetting the other's drawdown inside the same account. Isolated margin ring-fences the perp leg's collateral — safer for the rest of your account if that one position blows up, but it also means the short can be liquidated while cash sits unused elsewhere in your account. For a genuinely delta-neutral single-venue book, cross is usually the more coherent choice; for a leg that is one of many uncorrelated positions, isolated contains the blast radius. There is no universally right answer — there is only the liquidation price each mode implies, which you should compute before you send the order.

Should you pyramid into a funding trade?

Pyramiding is the practice of scaling into a position in increments rather than committing full size upfront, adding as the thesis holds. Imported from trend trading, it is usually pitched as a way to "add to winners." In a delta-neutral funding book the framing has to change, because there is no price winner — the position is hedged. What you are actually scaling into is conviction that the funding regime persists and that the market has depth to absorb your growing short without moving the basis against you.

That reframing has a direct consequence for how you size the clips:

• Scaled (decreasing) pyramiding — each addition smaller than the last, e.g. 1.0 / 0.5 / 0.25 — is the conservative default. It caps how fast total notional (and therefore liquidation exposure on the perp leg) grows, and it means your average entry basis does not deteriorate too quickly if you are adding into a compressing rate.

• Fixed-size pyramiding — equal clips at each level — grows exposure linearly and is defensible only when the funding rate and book depth are both clearly stable.

• Aggressive pyramiding — increasing clips (1.0 / 1.5 / 2.0) — concentrates your largest size at the worst average basis and the highest total liquidation risk, right when the crowd is most one-sided. For a funding trade this is the failure pattern, not the pro move.

The reason scaled entries dominate here is asymmetry. When you pyramid into a funding position, every addition raises the total notional the perp leg carries, which moves your aggregate liquidation price closer to spot and increases the funding bill you will owe if the rate flips. The upside of adding is linear (a bit more funding). The downside is convex (a liquidation cascade or a funding reversal now hits a larger book). Scaled entries are simply the sizing rule that keeps that convexity from compounding against you. A reasonable operating discipline is to define the total target notional first, then divide it into decreasing clips with pre-set add levels — never to "feel out" additions after the position is already in profit-looking territory.

What actually blows up a delta-neutral book?

The comforting story is that a hedged position cannot lose. The honest story is that it has four live risks, and every real blowup traces to one of them.

Liquidation of the levered leg

This is the dominant failure mode and the one traders underweight. Your spot leg is unlevered and cannot be liquidated; your perp short is levered and can be. On Hyperliquid, maintenance margin is set to half the initial margin at max leverage, and depending on the asset's leverage tier ranges from roughly 1.25% (40x-eligible assets) up to 16.7% (3x assets) — for example, around 2.5% at a 20x tier. Cross positions are liquidated when account value including unrealized PnL falls below maintenance margin times open notional; isolated positions apply the same logic to only the isolated margin and that position's notional.

The trap: a short perp run at meaningful leverage can be liquidated on an adverse move long before your spot gain "rescues" it — and the two do not net automatically if they sit in different margin buckets or on different venues. If the short is liquidated on a wick, you are left holding only the long spot, fully directional, into exactly the move you were hedging against. The defense is unglamorous: run the perp leg at low effective leverage so the liquidation price sits far outside any plausible wick, keep spare USDC margin, and monitor because funding you pay erodes margin over time and drags the liquidation price toward spot even when price is flat.

Funding reversal

The yield is not contractual. A basis that pays you 20% annualized this week can flip to a 30% annualized cost next week if positioning rotates and the perp starts trading at a discount. Because Hyperliquid settles hourly, a reversal starts billing you fast, in small hourly bites that are easy to ignore until they add up. A delta-neutral book must define its exit on funding, not on price: a rule such as "close if the trailing funding rate turns negative and holds for N hours" is what separates a funding-capture trade from an accidental short-vol position that gives back a month of yield in a bad week.

Basis and venue risk

When the two legs sit on different venues, they have two oracles, two liquidation engines, and a transfer path between them. A gap in spot price between venues, a withdrawal delay, or a de-peg on the collateral leg can open a real directional exposure even though your position sheet says "neutral." Single-venue construction on HyperCore (native spot plus perp in one USDC account) collapses most of this, which is a structural argument for keeping the trade on-venue where a deep native spot book exists.

Auto-deleveraging

Hyperliquid uses auto-deleveraging to keep the platform solvent when a counterparty's account value goes negative faster than liquidation can close it: profitable, higher-leverage traders on the opposite side are ranked and can have positions reduced. If you are the profitable short in a violent move, ADL can partially close your hedge for you at the system's timing, not yours — re-opening directional exposure you did not choose. It is a tail event, but for a book that is supposed to be neutral it is exactly the kind of surprise that turns a "safe" strategy into a loss. Size and leverage discipline is what keeps you far enough down the ADL queue to not care.

Where to go next

A delta-neutral funding book is really three disciplines stacked: reading the funding regime, executing both legs cleanly, and sizing so neither leg can take you out. If any one is weak, the other two do not save you. Before scaling real size into one of these, it is worth being fluent in the underlying funding mechanics in the funding rate strategies guide, in the order tooling that lets you leg in without bleeding slippage in advanced orders: TWAP, trailing and conditional, and in the position-sizing and liquidation-distance framework in risk management core principles. The traders who compound this strategy are boring on purpose: small clips, low leverage on the perp leg, a hard funding-flip exit, and no interest in the fattest rate on the board.

Sources

• Hyperliquid Docs — Funding

• Hyperliquid Docs — Margining

• Hyperliquid Docs — Margin tiers

• Hyperliquid Docs — Liquidations

• Hyperliquid Docs — Auto-deleveraging

• Hyperliquid Docs — Clearinghouse