Pendle's high yields were the price of points. Yield-token holders got back $730M less than they paid

01 · The tradeA 560.85% Pendle YT claim, and what the record shows

Hold a dollar token like Ethena's USDe, and it pays you a few percent a year. Some apps let you sell that future income today. Others let you buy it from someone else.

On 4 November, a trader who posts as @0xSearchie wrote that he made an average 560.85% APR on dollar tokens this way. He buys the income on Pendle Finance, an app built for this trade, and holds it to the end. He waits for good prices with limit orders and picks markets where the token's issuer pays extra rewards.

Pendle splits a token that earns interest into two parts that run until a set date. The principal token, or PT, pays back one unit of the token on that date, so buying it locks in a fixed rate. The yield token, or YT, collects all the interest until then and is worth nothing after. The prices of the two set a rate, the implied yield. YT buyers pay that rate up front and hope the token earns more.

Many YTs also earned points, a promise of a future token from projects like Ethena. That made YTs a cheap way to farm: one YT earns the points of a whole token for a fraction of its price. Ethena's launch post says its ENA airdrop, free tokens sent to its early users, went to Pendle YT holders and not to PT holders.

Bitquery records every Pendle trade on Ethereum, the chain that holds most of Pendle's deposits. We followed every market that has expired there, and every wallet that held its YTs. For each market, we compared the yield buyers priced in with what the token went on to earn. For each wallet, we added up what it paid and what it got back.

The record is hard on YT buyers. In 87% of markets, the token paid less than buyers had priced in. Most of those markets also offered points, and across all points markets, YT holders got back $730M less than they paid. In markets with no points and interest paid on-chain, one buyer in four made money.

02 · Priced in vs paidIn 87% of markets, the token paid less than buyers priced in

The median market priced in 8.8% per year. The median token paid 3.6%, well under half of that.

Points decide the size of the gap. In dollar-token markets without points, the gap was about one percentage point, and even there the token beat the priced-in rate in only 31% of markets. So PT buyers, who locked in the fixed rate, usually came out ahead.

In markets with points, the token beat the priced-in rate just 6% of the time, and dollar-token markets with points priced in about twice what they paid. Buyers paid the difference for points.

Most of these tokens pay out by rising in value against the coin they track, and we read that rise on-chain. Some paid no interest at all, only points. For the few that paid in separate reward tokens, we used Pendle's own daily figure.

03 · The price of pointsYT holders in points markets got back $730M less than they paid

We followed every YT from the wallet that got it to the day it expired. When a wallet bought or sold through Pendle, we used the price it paid. When it made its own YTs or moved them in from another wallet, we used that day's market price. Then we added the interest and reward tokens it claimed, all in dollars on the day.

About a quarter of a million wallets held YTs in points markets. Together they got back $730M less than they put in, counting interest, sales and rewards. Fewer than one in thirty came out ahead.

Most of that money went early. Markets that ended between April and June 2024 alone account for $226M, the tallest bar below, and the losses shrink to a trickle after 2025.

Two checks that use no wallet data come out within 4% of that total. One multiplies the gap between the priced-in and paid rates by the YTs outstanding each day. The other values every outstanding YT at its market price each day.

That $730M is what holders paid for points. Whether it was worth it depends on what they became, and some did pay out. Ethena sent its ENA airdrop to YT holders, with half released over six months. The Eigen Foundation says its second round of EIGEN covered people who used Pendle before 15 March 2024, through lists drawn up by the apps that held their staked ETH. One of those apps, ether.fi, later pulled most of its ETH out of EigenLayer. We do not put a value on airdrops here, because points live off-chain, where our data cannot follow them.

04 · Without pointsWithout points, 25% of Pendle YT buyers made money

Points markets are hard to judge, because the rewards never show up on-chain. So we also looked at the markets that offered none and paid their interest on-chain. There, you can see the whole trade.

Across them, one YT buyer in four made money. The median buyer lost a quarter of what they put in, and together they came out $5.1M behind.

We also ran a simple test. Say you bought the YT on a market's third day, at the market price, and held it to the end. That made money in 16% of markets. Buying a week later, a month later or halfway through does not help much: the odds stay below one in five. The PT bought on day three returned a median 2.9% over the 110 days left.

Those are typical buyers. The post is about beating them.

05 · What helpsBuying below the token's own yield lifted the share in profit to 43%

The post names two habits: buy with limit orders, and buy below the yield the token already earns. We tested both on every position in those markets where we could see the price paid.

Buyers who paid a rate below the token's own yield over the past month came out ahead 43% of the time, against about a quarter for everyone else. Their median result was a loss of just 2%. But only 542 positions pulled it off.

Limit orders helped less: buyers who used them for most of their position made money 31% of the time, against 27% for those who bought from the pool. Traders use them all the same. Across all markets, limit orders carried more than a third of the money spent on YTs, at a median 1% below that day's market price.

We could not test the post's other habit, markets where the issuer pays extra rewards. Some of those rewards reach holders outside the YT, where our wallet records cannot see them. The habits we could test raised the odds, yet most buyers who used them still lost money.

06 · Pendle nowPendle's deposits are down 90% from the peak

Points also built Pendle. Its deposits peaked above $10 billion at the end of August 2025 and are down 90% since. Ethereum still holds three quarters of what is left. We traced one part of that story, the unwinding of loops built on Ethena's USDe and Pendle PTs, in the $7 billion Ethena loop.

Fewer people use Pendle on Ethereum, too. The number of wallets using it each month is down 88% from its peak in the summer of 2024.

07 · GapsWhat we could not see

We do not value points or airdrops, so our results show only the on-chain side of the trade.

Some markets paid extra rewards outside the YT, such as the USUAL tokens that came with Usual's USD0++ markets. Our wallet results leave those markets out.

A wallet is not always one person. Vaults and trading bots hold YTs for many users.

This piece covers Ethereum only. Pendle also runs on Monad, Arbitrum and other chains. We also left out 160 Ethereum markets that exist on-chain but do not appear in Pendle's app, and every market still open, including the ones the post describes.

08 · The methodHow we measured it

Our data covers every Pendle market on Ethereum. We kept the ones listed in Pendle's app that had expired by early November 2026 and traded for at least a month, 455 in all.

For each market, we compared the rate it priced in with the interest its token paid. We left out the last two weeks, when the priced-in rate jumps around as time runs out.

For each wallet, we counted every YT it bought, sold, made or moved, more than half a million purchases in all, plus the interest and reward tokens that holders claimed. Pendle keeps part of the interest as a fee, which rose from 3% to 5% over the period, and our figures are after that fee. Dollar values use each day's price on Ethereum exchanges.

The full Ethereum history behind this piece is also sold as files in Bitquery's Ethereum dataset.

FAQ

What is Pendle in crypto?

Pendle, also called Pendle Finance, is an app that splits tokens that earn interest into a principal token and a yield token. People use it to lock in a fixed rate, to bet on rates going up, or to farm points.

How does Pendle work?

Pendle splits a token that earns interest into a principal token (PT), which pays back one unit on an end date, and a yield token (YT), which collects all the interest and points until that date. The prices of the two set the implied yield, the fixed rate a PT locks in.

Do Pendle yield tokens make money?

Usually not, going by Pendle's expired Ethereum markets. In 87% of them, the token earned less than buyers priced in, and in markets without points, one YT buyer in four made money.

How much did Pendle YT holders pay for points?

In Ethereum markets with points, YT holders got back $730M less than they paid, before counting the points. Some of them later turned into airdrops, such as Ethena's ENA.

What did Pendle principal tokens return?

Bought on a market's third day and held to the end, the median PT returned 2.9% over the 110 days left, the fixed rate it locked in.