Wednesday, June 17, 2026
Economy

Foreign Demand for US Treasuries Rose to 8.5 Trillion, but the Buyers Are Not Who You Think

· · 3 min read
Economy · June 17, 2026

Foreign Demand for US Treasuries Rose to 8.5 Trillion, but the Buyers Are Not Who You Think

The Treasury International Capital report for May landed on Monday, and

the headline number is not the one the market expected

. Foreign holdings of US Treasuries rose to 8.53 trillion dollars, a fresh record, but the composition shifted in a way that

should worry anyone pricing the long end of the curve

. Japan, the United Kingdom, and China still own roughly 31 percent of all foreign-held Treasury debt, yet the share of incremental buying coming from those three has collapsed to its lowest level since 2019.

The Record Hides a Rotation

Total foreign ownership crossed 8.5 trillion dollars for the first time, lifting the foreign share of marketable Treasury debt back toward 30 percent after a decade-long drift lower from the 2008 peak near 50 percent.

The headline flatters the underlying flow

. In the May TIC release, Japan trimmed its position by 12 billion dollars to 1.13 trillion, the United Kingdom added 18 billion to 768 billion, and China held flat at 760 billion. The net increase of roughly 24 billion dollars across the top three is the

smallest monthly add since the pandemic-era flight to safety in March 2020

.

Where the New Money Is Coming From

The buyers filling the gap are

not sovereign wealth funds or foreign central banks

. They are fund-custody jurisdictions. Luxembourg added 31 billion dollars, the Cayman Islands added 27 billion, and Ireland added 19 billion. Together those three custody centers absorbed 77 billion dollars of the 112 billion dollar monthly increase in foreign holdings, or roughly 69 percent of the flow. That is

a structural tell, not a tactical one

. Custody-jurisdiction buying reflects leveraged funds, hedge funds, and private-credit vehicles repricing duration after the May refunding, not official-sector demand stepping in to cap yields.

Why the Auction Bid-to-Cover Is the Better Signal

The TIC data publishes with a

two-month lag

, so the May print is a backward-looking snapshot. The forward-looking signal is the weekly auction bid-to-cover ratio and the indirect-bidder share, both of which act as real-time proxies for foreign demand. The most recent 30-year auction on June 12 printed a 2.31 bid-to-cover, the weakest tail of the year, with indirect bidders taking only 24.1 percent of the issue.

When the indirect share drops below 25 percent and the bid-to-cover falls under 2.4, history says the long end is vulnerable to a term-premium repricing within the next six weeks

.

The 2s10s curve has already responded,

flattening to 22 basis points

, its flattest level since March 2023. The 5s30s briefly inverted to minus 3 basis points intraday. Both moves are consistent with a market that no longer trusts the marginal foreign buyer to anchor the long end. The Federal Reserve owns 4.2 trillion dollars of Treasuries, roughly 15 percent of marketable debt, and its balance-sheet runoff continues at a pace of 40 billion dollars a month.

Every dollar the Fed sheds has to be absorbed by someone, and the TIC data says that someone is increasingly a leveraged fund, not a central bank

.

What the June 17 FOMC Has to Reckon With

Chair Warsh inherits a

market structure that is more fragile than the one Powell handed him

. When foreign official demand was reliable, the Fed could run quantitative tightening without disrupting the long end, because Japan and China would absorb the duration the Fed shed. That backstop is thinner now. The May TIC release shows official-sector holdings flat to down across the three largest sovereign holders, while custody-jurisdiction holdings, the proxy for leveraged and hedge-fund demand, rose by 77 billion dollars. That is a market that

needs a positive carry to hold duration

, and a positive carry requires either higher yields or a steeper curve.

For the June 17 dot plot, the implication is specific. A median 2026 dot at 3.13 percent, the March print, implies two cuts and a flatter curve. A median at 3.25 to 3.38 percent, the market consensus, implies one cut and a steeper curve. The latter is

the configuration the long end needs to attract the custody-jurisdiction buyer back at scale

. Watch the indirect-bidder share at the next 30-year reopening in late June. If it stays below 25 percent, the term premium will keep widening, and the 5.00 percent level on the 30-year becomes

the line to watch before Jackson Hole in August

.