By Kresmion Research, August 2, 2026
Every Tuesday the CFTC takes a snapshot of who holds what in the major currency futures markets. The snapshot taken on July 28 showed speculators net short a combined 300,673 contracts of euro, yen and sterling futures, which is 19.8 percent of the open interest in those three markets. Shorting the euro, the yen and sterling is, in effect, a single levered bet on a stronger dollar. In the three years of weekly reports Kresmion tracks, that share had never been reached. The prior record was 17.4 percent, set in April 2024.
Three days after the snapshot, the United States bought yen in the open market for the first time in more than a decade.
The record, and what it is made of
The raw numbers behind the 19.8 percent: speculators (non-commercial traders in the CFTC's classification) held a net short of 300,673 contracts across euro, yen and sterling futures on July 28, against 1,519,598 contracts of total open interest. That is the most stretched combined reading in all 158 weekly reports since July 2023, and it beat the prior record by almost two and a half points.
Each leg was near its own floor. On a three-year scale where 0 marks the most short a market has been and 100 the most long, the euro sat at 1.3, the yen at 4.5 and sterling at 9.0. Those are the bottom three readings of the 13 futures markets Kresmion tracks, and the next closest market, silver, sat at 12.2. The euro leg alone, at 72,447 contracts net short, was the second largest euro short of the 158 weeks.
One honest framing note before the story continues. This is one wager expressed three ways rather than three separate discoveries: all three currencies trade against the same dollar, so a record in the combination is a record in the dollar factor. And the yardstick matters. Measured in raw contracts, the euro short looks close to a record. Measured against open interest, which has grown sharply, the euro leg is 8.8 percent of its market, only the eleventh most stretched week in the sample. The record lives in the combined position, and the fair way to say it is that speculators as a group had never been this committed, per unit of market size, to the dollar side of the three big pairs.
Three days, three different hammers
The position was measured on Tuesday. What followed came in three distinct blows, each with its own mechanism.
Wednesday, July 29. The Federal Reserve held its target range at 3.50 to 3.75 percent, with three policymakers dissenting in favor of a hike, the first triple dissent in a single direction since September 2016. A hold had been the base case, but part of the market had priced a surprise increase, and the July 28 Kresmion brief covered exactly that gap. The dollar index slipped 0.6 percent on the day.
Thursday, July 30. The morning delivered a transatlantic growth split. The advance estimate of US second-quarter GDP printed 1.5 percent annualized against a consensus near 2.1, while the euro-area flash estimate beat expectations the same day, 1.0 percent year over year against 0.5 expected, with Germany, Italy and Spain each ahead of forecasts. The euro rose 0.7 percent, sterling 0.6 percent, and the dollar index fell 0.8 percent, its worst day of the week.
Friday, July 31. The Financial Times reported that the US Treasury bought yen to support the Japanese currency, with the New York Fed selling euros for yen through Goldman Sachs and Morgan Stanley, after the Treasury had told banks earlier in the day to stand ready. It was the first joint US-Japan yen-buying operation in more than a decade, and it followed a photographed page of Treasury Secretary Scott Bessent's notes reading "Buy Japanese Yen (JPY) $5-10 bil". By Kresmion's daily mark, dollar-yen fell 1.9 percent to 160.18, one of its dozen sharpest single days in almost five years of daily closes, and it kept falling into the 5 p.m. New York close near 157.6, a drop of about 3.5 percent on the day.
Across the three sessions the dollar index fell 1.56 percent, from 101.38 to 99.80. In Kresmion's daily history that lands in roughly the bottom 3 percent of all three-session moves.
What cuts against the clean story
A record position meeting a record-speed reversal reads like a lesson in crowded trades. Several things complicate that read, and they belong in the story.
The timing is unknowable from this data. The positioning snapshot is from Tuesday and the reversal came Wednesday through Friday, so the data cannot yet say whether the crowd held its short through the fall or covered into it. That answer arrives with the next report.
An official buyer moved the price. Friday's yen move was driven by government intervention, which says nothing about private traders changing their minds. The reported scale, $5 to 10 billion, is small next to the roughly one trillion dollars of yen-pair turnover on an ordinary day; interventions of that size work through signal more than through volume.
July 31 was also month-end, when index and portfolio rebalancing flows pass through currency fixes and can exaggerate a directional day.
And the base rate argues for humility. Sorting all 158 weeks by how crowded the combined position was, the dollar index's forward 20-day return shows no consistent pattern across the quintiles. Extreme crowding, in this sample, has not reliably preceded reversals. The record and the reversal arrived in the same week, and the honest reading is three real catalysts landing on a crowded position, with no evidence from this sample that the crowding itself set the timing or the size of the move.
Two more points cut the other way entirely. The dollar index at 99.80 sits near the middle of its range over the past 90 days, around its 40th percentile, so the week was a fast move into familiar territory, no breakdown. And rate pricing leans against dollar weakness: markets price roughly a two-thirds chance of a September hike and no chance of a cut into 2027, and higher US rates would ordinarily pull capital toward the dollar.
What would change the read
The next CFTC report is taken Tuesday, August 4 and released Friday, August 7, the same afternoon as the July jobs report. It will show what the record crowd did during intervention week. A combined short still near 20 percent of open interest would mean the position absorbed the blow; a sharp drop toward the mid-teens would mean the record marked the top of the trade and the unwind is already underway. Alongside it, watch whether the Treasury or the Bank of Japan confirm follow-on operations, and whether dollar-yen holds below the 163 to 164 area where it traded before Friday.
Key takeaways
| Point | Detail |
|---|---|
| The record | Speculators' combined net short in euro, yen and sterling futures reached 19.8 percent of open interest on July 28, the most in 158 weeks of data |
| The prior mark | 17.4 percent in April 2024; the euro leg alone was the second largest euro short of the sample |
| What hit it | A Fed hold with three hawkish dissents, a US GDP miss against a euro-area beat, and the first US yen-buying intervention in over a decade |
| The move | The dollar index fell 1.56 percent in three sessions, roughly a bottom 3 percent move; dollar-yen fell about 3.5 percent on Friday into a New York close near 157.6 |
| The caveat | Across the sample, extreme crowding has not reliably preceded dollar reversals; the intervention was official flow, and month-end added noise |
| The test | The August 7 CFTC report, released the same day as July payrolls, shows whether the crowd held or ran |
Frequently asked questions
What exactly did the CFTC data show?
The Commitments of Traders report taken July 28 and released July 31 showed non-commercial traders net short 300,673 contracts across euro, yen and sterling futures combined, equal to 19.8 percent of the open interest in those three markets. That is the largest combined share in the 158 weekly reports since July 2023 that Kresmion tracks.
Did the record positioning predict the dollar's fall?
No, and the history says it could not have. Sorting the three years of weekly data by how crowded the position was shows no consistent relationship between crowding and the dollar's forward returns. The fall that followed had three identifiable catalysts: the Fed's hold with hawkish dissents, a US growth miss against a European beat, and a yen intervention.
Why does the US buying yen matter so much?
Because it had not happened in over a decade. Currency intervention by the US Treasury is rare, and a joint operation with Japan, executed through the New York Fed by selling euros for yen, signals that both governments treat the yen's weakness as a problem worth spending on. The reported size, $5 to 10 billion, is small against daily yen turnover, so the operation mattered as a statement of policy more than as flow.
Does this mean the dollar will keep falling?
This note describes what happened and where the numbers sit in their own history; it makes no forecast. The next positioning report on August 7 shows how the record crowd responded, and rate pricing still leans toward a September hike, which pulls in the dollar's favor. This page is information, not investment advice.