All articles
Comparisons·February 18, 2026·5 min read

Hyperliquid vs dYdX for Market-Making Bots: Fee & Edge Analysis

Breaks down maker rebates, taker fees, funding rate mechanics, and API rate limits on Hyperliquid versus dYdX v4 to quantify which venue offers more edge per dollar of inventory deployed by an automated market maker.

If you're running a market-making bot and trying to decide between Hyperliquid and dYdX v4, the answer isn't simply "whichever has lower fees." The actual edge you extract depends on the full stack: maker rebates, taker spread costs, funding rate mechanics, order placement latency, and the API throughput ceiling your bot runs into at scale. This Hyperliquid vs dYdX market-making bot analysis breaks each of those down with the numbers that matter for automated flow.

Fee Structures: Maker Rebates and Taker Costs

Hyperliquid runs a tiered fee schedule based on 14-day notional volume. At the base tier (below $5M), makers receive a +0.002% rebate and takers pay 0.050%. Once you cross the first volume threshold ($5M–$25M), the maker rebate moves to +0.003% and taker drops to 0.045%. At the top tier (>$1B), you're earning +0.005% on every resting fill and paying nothing as a taker. The rebate is paid immediately on fill in the quote asset — no delayed settlement, no accrual lag.

dYdX v4 (the Cosmos chain version) uses a different model. There is no maker rebate in the traditional sense: the default maker fee is 0.000% (flat, no payment) and taker is 0.050% at base tier. Volume discounts exist — at >$100M 30-day volume, makers get a -0.001% rebate — but the thresholds are steep and the rebate amounts are smaller than Hyperliquid at comparable tiers.

In raw dollar terms per $1M notional filled as maker: Hyperliquid returns $20–$50 in rebates, dYdX returns $0–$10. If your bot turns over $10M/day in resting orders, that's a $200–$500/day difference in direct fee income before spread capture is counted.

Funding Rate Mechanics and Market-Maker Carry

Both venues use an 8-hour funding cycle, but the mechanics differ in ways that affect how you skew inventory.

Hyperliquid publishes its predicted funding rate continuously via websocket. The rate is derived from a mark/oracle price premium with a dampening function and caps at ±0.05% per period. Critically, the oracle price is sourced from a weighted median of validators running price feeds — meaning it's resistant to brief mark manipulations. For a market maker quoting around mid, this makes the expected funding for holding inventory reasonably predictable and exploitable: you can tilt your bid/ask asymmetrically to get filled on the side that pays you funding.

dYdX v4 uses a similar 8-hour epoch, also capped at ±0.05%, but the oracle is sourced from a different validator set and the premium calculation includes a 5-minute trailing TWAP rather than a real-time mark. The lag makes short-term funding prediction less accurate, which matters when you're sizing your inventory skew dynamically.

For a bot that factors funding into its quoting model — like our Hyperliquid market-making bot — the continuous Hyperliquid oracle feed is directly useful. The dYdX TWAP creates a ~5-minute blind spot where your carry estimate is stale.

API Rate Limits and Order Throughput

This is where the venues diverge most sharply in practice.

Hyperliquid exposes a REST endpoint and a websocket for order placement. REST is limited to ~1,200 requests/minute per IP, but the real bottleneck is the websocket: order placement via websocket is effectively unbounded within the L1 block rate, which processes every ~2 seconds. A well-tuned bot can place and cancel dozens of orders per cycle. The EIP-712 signing requirement means each order needs an on-chain signature, but this is parallelizable with pre-signing batches.

dYdX v4 runs on a Cosmos application chain with a block time of ~1 second and an order placement limit of ~80 orders per block per address enforced at the validator level. There is no secondary REST bypass — all orders route through the validator mempool. If you're running more than ~4,800 order operations per minute, you will hit this wall and orders will be dropped silently with no error response. For tight two-sided quoting in a volatile market where you're refreshing quotes on every price tick, this is a real constraint.

The practical ceiling for a dYdX v4 market maker is around 3–5 liquid markets simultaneously. Hyperliquid can realistically handle 10–20 markets with the same infrastructure footprint.

Liquidity Depth and Fill Quality

Edge on paper disappears if your quotes don't fill, or if fills come from informed flow that immediately moves against you.

Hyperliquid currently handles the majority of on-chain perpetuals volume — consistently above $1B/day across its most liquid pairs (BTC, ETH, SOL). The order book is tight enough that a market maker quoting 1–3 bps wide earns fills from real flow, not just bot-vs-bot games.

dYdX v4's liquidity has declined materially since the v3-to-v4 migration. Total daily volume sits in the $100M–$200M range on most days. Wider spreads are required to avoid adverse selection from a thinner, more sophisticated user base. For a market maker, thinner books mean your rebates are earned less frequently and your edge-per-fill needs to be wider just to break even on hedging costs.

Practical Trade-Offs and When to Choose Each

Factor Hyperliquid dYdX v4
Maker rebate (base tier) +0.002% 0.000%
Taker fee 0.050% 0.050%
Order throughput High (websocket, ~2s blocks) Capped (~80/block/address)
Funding oracle lag Near-real-time ~5-min TWAP
Daily volume >$1B ~$100–200M
Multi-market scaling 10–20 markets feasible 3–5 markets practical

Choose Hyperliquid if you need volume throughput, want positive maker rebates from day one, and are quoting more than 5 markets. The immediate rebate settlement and continuous oracle feed make inventory management substantially easier to implement correctly.

Choose dYdX if your strategy depends on Cosmos-native composability (IBC flows, on-chain settlement finality guarantees, validator-level order priority) or you have regulatory reasons to use a chain with decentralized governance. It is also marginally cheaper as a taker at high volume, which matters for delta-hedging legs.

For most market-making bots, the combination of higher rebates, faster oracle data, and higher order throughput makes Hyperliquid the stronger venue today. That calculus can change — dYdX's governance has proposed rebate programs several times — so it's worth re-running these numbers quarterly.


If you want a production market-making bot built for Hyperliquid or dYdX with inventory skew, funding-aware quoting, and hard risk limits baked in, talk to us — this is exactly what we build at TierZero.

Need a bot like this built?

We design, build and run trading bots on Solana, Hyperliquid and Polymarket.

Start a project
#Hyperliquid#dYdX#market-making#trading bots#perps#fee analysis#DeFi#comparisons