Skip to content
BiVelio joins the NVIDIA Inception Program NVIDIA Inception Program BiVelio is part of the ElevenLabs Grants Program ElevenLabs Grants

Article 8 · Capturing funding

How Hyperliquid funding is captured (and why it is not free)

Five ways to collect Hyperliquid funding, preregistered before testing them: after fees, none is an arbitrage.

Read the full article on Substack (in Spanish)

An educational tool, not investment advice. Historical figures from the article's versioned report, in-sample, in dollars and before taxes; past performance does not guarantee future results.

How much is left after fees if you collect the funding difference with another platform?

2.5%a year on capital, after costs
2.8%gross difference a year
0.30%difference needed to cover the costs
Other platform → a year on capital, after costs
A year on capital, after costs (BTC, $10,000 at 2x)

Short the BTC perpetual on Hyperliquid and long it on another platform, $10,000 of capital at 2x per leg, for the whole window: the difference minus fees, slippage and transfers.

How it's computed

From the article's report (hyperliquid/reports/captura_funding/tabla_14.csv), with data through Sep 28, 2026: the Modeler doesn't yet compute this way of capturing funding.

Main figure: matches the article From the article: With $10,000 at 2x for the whole window, BTC against Paradex left 2.5% a year net (interval, 1.0% to 4.2%). The article's versioned report gives the same figure (2.5%).

See the simple carry in the Modeler

Can you grab one hour's payment and get out?

99.99%of the hours, the payment didn't cover opening and closing at market
0.69%of the hours it covered two passive (maker) fees
0.16%of the hours it covered two taker fees, without slippage
Perpetuals → of the hours, the payment didn't cover opening and closing at market
Hours in which one hour's payment covered a round trip: two maker fees, two taker fees or at market (taker and slippage)

Get in just before the hourly payment and out just after: one hour's payment almost never covers opening and closing at market, from January 2024 to September 2026.

How it's computed

From the article's report (hyperliquid/reports/captura_funding/tabla_17.csv), with data through Sep 28, 2026: the Modeler doesn't yet compute this way of capturing funding.

Main figure: matches the article From the article: In 99.99% of 660,806 perpetual-hours, one hour's payment was smaller than the cost of opening and closing at market. The article's versioned report gives the same figure (99.99%).

See the simple carry in the Modeler