The real costs of copy trading on Hyperliquid (beyond the fee)

The fee is the smallest line on the bill

Ask most people what copy trading costs and they will quote the platform fee. It is the number on the landing page, so it is the number that gets compared.

It is also, for any system mirroring active perp traders, usually not the dominant cost. Exchange fees, spread and funding are paid on every unit of turnover, and a high-turnover leader can generate several times their account value in notional volume per month.

This note lays out the whole stack, in the order it accrues, and ends with the questions that let you price any copy system — this one included — before committing capital to it.

Visible fees

Start with the costs that are explicitly quoted somewhere, because they are the easiest to compare and the easiest to over-weight.

Exchange taker and maker fees

Hyperliquid charges its own trading fees, which differ for liquidity-taking and liquidity-providing orders and vary with volume tier.

Mirrored orders are overwhelmingly taker orders. A copy system that waits to be filled passively is a copy system that misses the leader's move, so it crosses the spread instead. That choice is defensible, but it means you should assume taker rates on essentially all mirrored volume.

Fees apply to opens and closes alike. A round trip is two charges, and the leader's decision to scale in and out in tranches multiplies the count.

Builder fees on mirrored notional

Hyperliquid supports a native builder fee: a rate a builder attaches to the orders it submits, approved in advance by the account owner up to a maximum. Here the rate is 0.1% of mirrored notional volume, applied at execution on opens and closes.

The structural property that matters is what it is charged on. A volume fee accrues whenever the system trades, in profitable and unprofitable periods alike, and stops entirely when it does not trade.

Profit shares and subscriptions behave differently

A profit share costs nothing in a flat period and a large amount in a strong one, with high-water-mark accounting so the same gain is not charged twice. It is insensitive to turnover, which makes it cheaper for very active strategies and more expensive for successful ones.

A subscription costs the same whether the system trades or not, which makes it cheap at large size and punitive at small size.

There is no universally cheaper model. The correct comparison is your expected turnover and your expected return against each fee structure, not the rates side by side.

Slippage and fill divergence

Spread is a certain cost. On every entry and exit you pay approximately half the spread relative to mid, twice per round trip, in every sleeve.

Slippage is the additional cost of size: your order walks into the book beyond the best level. It grows with order size relative to available depth, which means it is worst in exactly the markets where a leader's edge is most likely to be real — smaller, less efficient perps.

Copy trading adds a specific extra: you fill after the leader. Their order has already consumed the top of the book, and the price you see is the price their trade helped create. The mechanics of this are covered in the tracking-error note; here it is only necessary to register that it is a cost, paid on every mirrored fill, and that it is invisible in any fee schedule.

Funding rate impact

Perpetual futures pay or receive funding periodically depending on which side you hold and how the perp trades relative to spot. Holding the crowded side of a strongly positioned market is a continuous cost that has nothing to do with whether the position is right.

For copy trading there is a timing wrinkle. Funding accrues to whoever holds the position at the funding stamp. Entering minutes after the leader can change whether you receive or pay a particular payment, and exiting minutes later can add one they avoided.

For short-horizon leaders this nets out to noise. For leaders who hold directional positions across many funding periods, funding can be the largest single line in the cost stack, larger than fees and slippage combined.

Capital efficiency and opportunity cost

Isolated sub-accounts contain failures. They also fragment margin: each sleeve must independently hold enough collateral to open its positions and absorb normal adverse movement, and that collateral cannot be lent to a sleeve that needs it more.

The result is that a diversified basket holds more total margin for the same exposure than a single cross-margined account would. The excess is not lost, but it is not working either, and its opportunity cost is real.

There is a second, less obvious efficiency cost. Capital committed to sleeves is capital you cannot deploy elsewhere on short notice, and unwinding a sleeve to free it means paying the exit costs of positions you did not want to close.

This is a genuine trade-off, not a hidden charge. Containment costs capital efficiency; the question is whether the containment is worth it for your account, and at small size it frequently is not.

Tracking error as an economic cost

Tracking error is normally discussed as a fidelity problem. It is more usefully treated as a cost, because it has an expected sign once fees and timing are included.

Your fills are worse than the leader's on average, not symmetrically distributed around them: you trade after them, into the liquidity they left. Add taker fees on both legs and the expected divergence over many trades is negative even when individual trades come out ahead.

It is also not constant. It scales with the leader's turnover, with their preference for illiquid markets, and with how quickly their edge decays after their trade becomes public. A leader whose signal is still valuable ten minutes later costs far less to copy than one whose entire edge is in the first few seconds.

Treating tracking error as a cost changes leader selection: two leaders with identical returns are not equally copyable, and the slower one is cheaper.

How costs compound

Almost every line above is multiplied by the same variable: turnover. Exchange fees, builder fees, spread, slippage and a large part of tracking error all scale with how much notional gets traded, not with how much capital you have.

That makes turnover the single most important cost input, and it is chosen by the leaders, not by you. A basket of leaders who each turn the book over daily costs an order of magnitude more to run than a basket holding positions for days, at the same fee rate and the same account size.

Compounding runs the other way too. Costs are paid from the capital that would otherwise compound, so a persistent drag reduces the base for every subsequent period. This is why small, recurring costs matter more than occasional large ones of the same total size.

Any arithmetic here would be illustrative only, because the inputs — turnover, spread, funding — are properties of the leaders and markets in a given period, not constants that can be quoted in advance.

• Turnover multiplies fees, spread, slippage and tracking error together.
• Funding scales with holding time, not turnover — the one exception.
• Idle margin scales with sleeve count, not with trading at all.
• Costs are paid from compounding capital, so drag persists.

How to evaluate the true cost of any copy system

The useful exercise is not comparing headline rates. It is reconstructing the whole stack for the specific leaders a system would run for you.

A system that answers these questions clearly is not necessarily cheap, but it is measurable. One that cannot answer them is asking you to accept an unpriced cost.

• What is the platform fee charged on — volume, profits or time?
• What is the expected monthly mirrored notional relative to account size?
• Are mirrored orders taker or maker, and what is the assumed fill rate?
• Which markets do the leaders trade, and how deep are those books?
• What is the average holding period, and how many funding stamps does it cross?
• How much margin sits idle across sleeves at a typical moment?
• Is realised tracking error published against the model, or only the model?
• Are the exit costs of a leader replacement disclosed?

Limitations of any cost estimate

Every input above is variable. Spread widens in stress, funding flips sign, turnover changes when leaders change, and slippage depends on depth at the moment of the fill rather than on an average.

A cost estimate is therefore a range and a set of sensitivities, not a number. Anyone presenting a precise total cost figure for a copy system in advance is quoting an assumption, not a measurement.

What can be stated honestly is which costs exist, what drives each of them, and which ones the system's design bounds. That is what this note has tried to do.

Past performance is not indicative of future results. Perpetual futures are leveraged instruments and carry a substantial risk of loss, including the loss of your entire position.

Conclusion

Price the stack, not the headline. Exchange fees, builder fees, spread, slippage, funding, idle margin and tracking error together determine what a copy system actually costs you, and turnover is the multiplier on most of them.

The fee structure used here is set out in the builder-fee note, the HyperMirror-specific version of this cost breakdown is in the running-cost note, and the residual risks are stated plainly in the risk disclosure.