AI is Causing the Collapse of U.S. Treasuries, and Even if Successful, It Will Face Heavy Taxes

Author: Wall Street Insight

On September 24, U.S. long-term bonds plummeted across the board. The yield on 30-year Treasuries rose to 5.48%, a 20-year high; the 10-year yield broke 5.2%, the highest since the financial crisis. Long-term yields have surged almost vertically over the past two months, with Treasury Secretary Yellen intervening twice to buy back bonds, each time only providing a brief respite before the market reverted to its previous state.

The massive deficit, high oil prices, and the Federal Reserve's renewed interest rate hikes—these are all real underlying pressures, but they are not new variables in this round. What truly sets this round apart is the emergence of a competitor on the demand side that is fundamentally indifferent to price.

AI is Squeezing the Treasury Out of the Market

In a bid to build data centers and computing power, tech giants are issuing ultra-long bonds at a record pace, almost without regard for price.

As of August, the five largest cloud providers had issued nearly $230 billion in bonds this year, more than double last year's total, with a significant portion concentrated in 20-year, 30-year, and even 40-year maturities.

These transactions would make headlines in any year.

Meta, founded only 22 years ago, issued a 40-year bond last October—by the time it matures, Zuckerberg will be 80 years old. This single issuance was worth $30 billion, setting a record for corporate bond issuance. In April of this year, Meta added another $25 billion.

Alphabet even issued a century-long bond in pounds. Amazon raised $54 billion in a single transaction in March. Oracle, currently rated just two notches above junk, has negative free cash flow but continues to issue bonds one after another.

Why are they unconcerned about interest rates? Yellen herself remarked in an August interview that these companies do not pay attention to interest rates when borrowing because they firmly believe that the returns from AI will be so high that current interest rates are negligible.

In other words, the normal brake on borrowing impulses that rising interest rates would provide has failed in the face of AI demand.

However, the funds willing to lock in "long money" are limited: pension funds, insurance companies, and long-term bond funds—these buyers are the ones supporting the long end of the market, and their money is finite. When money flows into the long bonds of tech companies, it cannot flow into Treasuries.

With Alphabet's 30-year corporate bond yield at 6.4%, Treasuries can only reclaim that money by raising their own yields.

This is essentially a "reverse squeeze": in the past, government bond issuance squeezed out private investment; now, it is the AI giants squeezing the government out of the bond market.

Nomura estimates that tech giants have borrowed about $200 billion in long-term funds, accounting for a quarter of the Treasury's annual issuance of medium- and long-term bonds. Investment-grade bond funds are also reducing their holdings of Treasuries while increasing their positions in corporate bonds, and the scale of foreign private funds buying U.S. corporate bonds has already surpassed that of U.S. Treasuries—something that was unimaginable before.

A large, price-insensitive private borrower stationed at the long end fundamentally rewrites the rules of the game.

Yellen's buybacks are simply not enough here—each time only a few billion dollars, which is a drop in the bucket in the over $30 trillion Treasury market, and it can only adjust the maturity structure of Treasuries, not control the pace of bond issuance by tech giants.

There is also a self-reinforcing cycle hidden here: tech giants issuing bonds push up long-term interest rates, and higher rates in turn raise the funding costs of AI projects themselves—the premium that AI giants pay when issuing new bonds has soared from over 2 basis points last year to 12 basis points this year.

AI Will Only Make the U.S. Government Poorer

If it were just a matter of competing with the Treasury for money, there might be a solution, because if AI truly succeeds in driving rapid economic growth, tax revenues will eventually catch up, making it easier for the government to repay debts, and yields would naturally fall.

But the problem is that the AI boom is unlikely to expand the government's tax base. A surge in productivity and GDP could actually worsen the government's debt repayment capacity.

This deterioration is mainly reflected in both the long-term and short-term aspects.

In the long term, individual income taxes and social security taxes together contribute more than three-quarters of federal revenue, meaning the U.S. government's financial lifeline is tied to payrolls, while the economic effects of AI are precisely compressing the value and income of labor—companies are doing more with fewer people, profits rise, but total wages decline, leading to less revenue for the government from individual income and social security taxes. If AI causes labor income to drop by 10 percentage points of GDP, most developed countries will face significant fiscal problems.

In the short term, the "Big and Beautiful Act" passed in 2025 sharply reduced corporate taxes in the U.S. This act restored the policy allowing full tax deductions for capital expenditures in the year they are incurred. The intention of the act was to stimulate investment, but for tech giants that were already willing to spend at any cost to invest in data centers, this is not an incentive but a windfall. Microsoft disclosed in July that its current tax liability was only $2.5 billion, down from $14.1 billion in the same period last year—a reduction of 80%, while profits were soaring. This is the magic of depreciation. Data from the Congressional Budget Office in September confirmed this: corporate income tax revenue fell by a quarter year-on-year in the first 11 months of this fiscal year, decreasing by nearly $100 billion.

Shrinking revenues, ballooning expenditures, and a cliff in tax revenues. The deficit can only continue to be filled by issuing bonds. The federal deficit has already reached $2 trillion in the first 11 months of this fiscal year, and the total national debt just surpassed $40 trillion in August, with interest alone consuming about $1 trillion each year.

Under the current system, expecting AI to ultimately save the day by expanding the existing tax base is simply unrealistic.

The More Successful AI Is, the Heavier the Taxes Will Be

The more successful AI becomes, the more thoroughly it replaces labor, and the less tax the U.S. government receives from payrolls.

At the same time, government spending will not shrink; rather, it will expand—because more people will need retraining and welfare to support their livelihoods.

As the government's fiscal gap widens, money can only be found in the one area that is still expanding—profits generated by AI.

This content is provided for general informational purposes only and doesn't constitute financial, investment, legal, or tax advice. Any events, rewards, online promotions, or related information mentioned herein should not be considered a recommendation, solicitation, or invitation to purchase, sell, trade, or otherwise deal in any crypto assets. Crypto assets are highly volatile and may result in loss. The availability of WEEX services, products, and related events may vary by region. You are responsible for ensuring that your participation is in accordance with applicable local laws and regulations.