Same Bitcoin, Different Funding

01 / WHY THE KOREAN ANGLE

Korean perp demand lives offshore

Korean investors cannot trade crypto perpetuals at home. The country’s registered exchanges run spot markets, and Korean coverage this year has named overseas platforms that offer derivatives as one reason trading activity has moved abroad.[5] The demand shows up in traffic data. In August 2026, Korea was the third-largest country source of visits to hyperliquid.xyz, at 7.45%, after the United States and Singapore.[6]

Korean traders also follow one cross-venue gap closely: the kimchi premium, the difference between Bitcoin’s price on Korean exchanges and abroad. This year it has ranged from −3.70% to +4.05%, and it was negative on 123 of the 221 days CryptoQuant tracked to August.[5] Traders who read that gap every day tend to read funding the same way, treating a higher or lower rate on another venue as an opportunity in itself. The rest of this brief checks that habit against settled data.

The legal context matters for anyone studying this market. In June 2026, Korea’s Financial Intelligence Unit warned that foreign platforms serving Korean residents without registering are operating illegally, and that people who promote them for a fee may themselves face criminal charges. Korean-language pages, won payments and marketing aimed at Korean users all count as signs of doing business in Korea.[7] Separately, an amendment to the Foreign Exchange Transactions Act, promulgated on 2 June 2026, requires businesses that move crypto across the border to register and to report those transfers to the Bank of Korea from December 2026.[8]

TAKEAWAYKorean demand for perpetuals is real, but it sits offshore and under a legal line that is getting sharper.

02 / SAME BITCOIN, THREE CLOCKS

Different rules produced different signs

Funding is a payment between longs and shorts that pulls a perpetual’s price back toward spot. When the perpetual trades above spot, longs pay shorts; when it trades below, shorts pay longs. It is not a fee to the exchange, and whether it is a cost or an income depends on which side you hold. All three venues share this design, but they measure and settle it differently.[1][2][3]

Aster runs an oracle-priced 1001x mode and an order-book Pro mode. The funding rules and history used here are those of Pro mode.

To compare them fairly, the team used settled rates only, kept their signs and did not estimate or fill any missing values. Hyperliquid and dYdX settle every hour, so their eight hourly rates in each window were added up; Aster’s 8-hour settlement was used as published. That gives six 8-hour windows across the 48 hours.

It was not a quiet window. Trading volume ran at 4.34 times its own baseline on Hyperliquid, 14.90 times on dYdX and 4.38 times on Aster, where the baseline is the median of the 27 eight-hour windows from 10 to 19 August. Binance spot BTCUSDT rose 12.99% over the same period.[4] Each multiple compares a venue with itself, not with the others.

The totals split in sign: +5.94 bp on Hyperliquid, +4.18 bp on Aster and −10.36 bp on dYdX. A long paid funding on Hyperliquid and Aster and received it on dYdX. The widest gap at a single settlement was 3.865 bp over eight hours, between Hyperliquid and dYdX at 08:00 UTC on 21 August.

What the data cannot say yet is how much of the gap came from the rules and how much from market conditions on each venue, because a settled rate mixes both. Separating them means rebuilding each venue’s inputs and settings for the window, reproducing the published rates, and then changing one rule at a time. The team has built that procedure on synthetic data; it has not yet been run on the real window.

TAKEAWAYDifferent clocks gave different signs. How much the rules themselves contributed is still open.

03 / COST

The gap did not pay

A difference in funding is not a profit. To capture one, a trader holds opposite positions on two venues and pays to get in and out of both first.

AssumptionTake $10,000 of notional on each of Hyperliquid and Aster, with all four orders, two entries and two exits, paying taker fees. Under the base fee schedules that round trip costs about $17.[1][3]

MeasuredOver the same 48 hours, the funding gap between Hyperliquid and Aster was 1.76 bp, or $1.76 on $10,000. That covers about a tenth of the fees.

At that pace, about 0.037 bp an hour, earning back $17 would take more than 400 hours. That figure is still generous, because it leaves out slippage, price moves while the position is open, borrowing costs and the cost of moving collateral between venues.

TAKEAWAYA visible funding gap becomes a trade only when it beats the round-trip cost.

04 / WHAT TO CHECK

Compare total cost, then decide

For traders

1. Before comparing rates, check whether each number is predicted or settled, and hourly or 8-hourly. Numbers of a different kind do not belong side by side.
2. When another venue looks cheaper, turn the rate gap into money at your own position size and set it next to the round-trip fee of switching. If the money is smaller, the switch does not pay.
3. Fix the holding period first. For short holds, execution costs such as spread, depth and fees decide the answer. For holds of several days, cumulative funding and margin headroom take over.

For teams studying Korean demand

• Make funding easy to verify. Publish settled-rate history and the full method: sampling interval, averaging, settlement clock and the price used for payment. An audience used to comparing venues will check.
• Read the registration rules before any Korea-facing activity. Under the June 2026 guidance, Korean-language pages, won payments and Korea-targeted promotion all count toward doing business in Korea.[7]
• Expect cross-border flows to be visible. From December 2026, businesses that move crypto in and out of Korea report those transfers to the central bank.[8]

TAKEAWAYFix the size and holding period, add up every cost, and only then compare venues.