Collecting Kalshi perp funding for 30 days: Bitcoin passed, Ether did not
Own the coin, short the perp, keep the funding. Bitcoin made 6.3% a year, Ether lost money, and fees decided it.
Kalshi's perpetual futures pay a small amount every 8 hours between buyers and sellers, called funding. When more people want to be long, longs pay shorts. So there is a classic trade: own the coin, short the same amount of the perp, and the price moves cancel. What is left is the funding. We ran it on paper for 30 days on Bitcoin and Ether, with the pass mark written first.
The setup
- Per coin, $1,500: $1,000 of the coin bought elsewhere, and $500 of margin behind a $1,000 short on Kalshi's perp.
- Entered once on 14 October at the live price, and held.
- Costs both ways: 0.12% on the perp (our fee tier) and 0.25% on the coin. The exit was charged as if we closed at the 30 day mark.
- Pass, judged on 13 November: more than 3.5% a year after every fee, per coin and combined, with the margin never below 1.5 times what Kalshi requires.
The result
| Funding received | Fees, in and out | Net | A year | |
|---|---|---|---|---|
| Bitcoin | $14.87 | $7.40 | +$7.74 | +6.28% |
| Ether | $5.30 | $7.40 | minus $1.99 | minus 1.61% |
| Both ($3,000) | +$5.75 | +2.33% |
Net is funding minus fees, plus a few cents from the perp and the coin not moving in exact step (plus $0.27 on Bitcoin, plus $0.11 on Ether).
Bitcoin cleared the bar. Ether did not, and together they made 2.33% a year against 3.5%. Failed. The margin was never in danger: the closest it came was 2.4 times the requirement.
Why the two coins split
Of 90 funding payments, Bitcoin's shorts were paid 73 times, paid out once, and got nothing 16 times. Ether's shorts were paid 41 times, paid out 10 times, and got nothing 39 times. Bitcoin had steady demand to be long; Ether mostly did not. Same trade, same costs, and the only difference was who was crowding the other side.
The fee is the whole story
Both coins paid the same $7.40 to get in and out. On Bitcoin that was half of a month's funding; on Ether it was more than all of it. A carry trade earns by the day and pays its fees once, so a short test punishes it. That is why the follow up test holds a 1x version for 90 days, with its verdict on 16 January.
What to take from it
- Check funding history per coin before you run the trade. Positive on one coin says nothing about another.
- Work out how many days of funding your fees cost. If it is a month, a month long test can only just break even.
- Compare with doing nothing. Cash at 3.5% a year was the bar, and the combined trade lost to it.
We do run a version with real money: a Bitcoin only hedge since 18 October, reported weekly in State of the Lab. It is a test, not a recommendation.
Want to run rules like this yourself? Start with a $5 first month, Ping included. Ping counts a paper fill only when a real trade proves it could have filled.
Where this comes from: our paper ledger for this lane, 14 October to 13 November 2026: Kalshi's published funding rates and live perp prices, the coin priced at Kalshi's reference index, with the fees stated above. Results are our own computation. Nothing here is advice. It is the record.