A note on why I wrote this 

Good decisions depend partly on understanding the world in which those decisions must work. Yet many consequential developments occur in fields in which most of us are not specialists.

 

I wrote this essay initially to help my daughters understand why rising bond yields matter — not only for the financial decisions they make, but also for how they think about public policy, what they expect from political leaders, and the questions they ask when deciding whom to support. The exercise raised a broader question that is equally relevant to leaders: how do we make a complex change intelligible enough to judge its significance and decide what deserves our attention?

 

This essay is an attempt to do that by examining the rising cost of capital.

 

 

Something important is changing in the American economy, and it is easy to mistake it for another bout of high interest rates.

 

The interest rate the United States government must pay to borrow money for 30 years through its longest-term Treasury bonds recently reached levels not seen since 2007. That matters because the U.S. Treasury market is one of the foundations of the global financial system. U.S. government bonds provide a benchmark against which many other forms of borrowing around the world are priced, as well as serving as financial reserves and collateral throughout the international banking and investment system.

 

The rate the American government pays therefore influences borrowing costs across the United States and around the world — for businesses financing expansion, families buying homes, consumers borrowing for cars, and governments refinancing their debts.

 

So this is not simply a story about bonds.

 

It is about the price of money.

 

And there are reasons to think that something more fundamental may be happening than the familiar economic cycle in which inflation rises, the Federal Reserve raises interest rates, economic activity slows and borrowing costs eventually come back down.

 

The possibility we need to consider is more consequential: capital itself may be becoming persistently more expensive because demand for money is rising at the same time that some of the investors who traditionally supplied it are becoming less willing or able to do so.

 

The possibility we need to consider is that the market for capital itself is changing: exceptionally large government borrowing and a major new source of private-sector demand are meeting a less dependable pool of traditional buyers.

 

If that is right, the question is no longer simply: When will the Fed cut rates?

 

A more useful question becomes: Who gets access to capital when many powerful borrowers want much more of it at the same time — and who gets crowded out?

 

The simple economics of a bond market

 

Government borrowing can sound remote from everyday life. The underlying mechanism is not.

 

The United States spends more than it collects in taxes. To finance the difference, the Treasury Department sells U.S. Treasury securities — government debt commonly referred to simply as Treasuries. Investors lend the federal government money by buying these securities and receive interest in return.

 

Because they are backed by the U.S. government, traded in enormous quantities and can usually be bought and sold readily, Treasuries have long been treated as among the safest and most liquid financial assets in the world.

 

They also serve another important purpose.

 

Foreign governments and central banks hold large quantities of financial reserves so that they can intervene in currency markets, meet international obligations or protect themselves during periods of financial stress. Because the U.S. dollar is the world’s principal reserve currency, a substantial share of those reserves has historically been held in dollar assets, especially U.S. Treasuries.

 

When investors are eager to buy Treasuries, the American government can borrow relatively cheaply. When investors require greater compensation before lending, Treasury yields rise — meaning the effective interest return investors demand for holding U.S. government debt increases.

 

Those yields then become a benchmark for much of the rest of the economy.

 

Suppose an investor can earn close to 5 per cent lending to the U.S. government, which is generally regarded as one of the safest borrowers in the world. A company asking that same investor for money normally has to offer more. Mortgage rates, corporate borrowing costs and many other forms of credit are similarly influenced by what investors can earn on government debt.

 

Higher Treasury yields therefore ripple outward.

 

That is why a movement in a market most people never directly enter can eventually affect the mortgage on a house, the economics of a new factory, the viability of a property development and the monthly cost of financing a car.

 

It also affects government itself.

 

America’s gross federal debt has now crossed $40 trillion. The number is so large that it is difficult to grasp: it exceeds the value of all the goods and services the U.S. economy currently produces in an entire year, and the debt has more than doubled over the past decade. 

 

The immediate significance is not that the government must repay $40 trillion at once. It is that this enormous stock of debt must continually be financed and refinanced. As older, low-interest debt matures and is replaced at higher rates, more federal revenue goes towards interest rather than public services, defence, infrastructure or tax reduction.

 

The government can therefore become caught in an uncomfortable feedback loop: the more it borrows, and the more expensive that borrowing becomes, the more of its future resources must be devoted to servicing previous borrowing.

 

And because the U.S. government is already borrowing on such a large scale, any additional large borrower entering the same capital markets matters; government debt is only one part of what is changing.

 

A powerful new competitor for capital has entered the market.

At the same time that the U.S. government needs enormous amounts of financing, the build-out of artificial intelligence is creating a large and relatively new source of demand for capital. Goldman Sachs estimates that AI-related debt issuance has approached $500 billion so far in 2026.

 

We cannot yet say with confidence how much this new borrowing is contributing to the rise in Treasury yields relative to government deficits, inflation, energy prices and other forces. But the mechanism matters. Technology companies and the U.S. government compete for overlapping pools of investor money. When another exceptionally large class of borrowers enters the market, investors have more choices, and existing borrowers may have to offer higher returns to attract capital.

 

The precise size of the effect is uncertain. The competitive pressure is not.

 

AI may feel weightless because we encounter it through software. Its physical infrastructure is anything but.

 

Building advanced AI requires data centres, semiconductors, electricity generation, transmission infrastructure, cooling systems, fibre networks and enormous computing capacity. Much of that must be financed before anyone knows precisely how large the eventual economic returns will be.

 

America’s technology companies are therefore entering capital markets on an extraordinary scale.

 

There is good reason to welcome this. Successful investment in AI could raise productivity substantially. If workers and businesses can produce more with the same resources, incomes and living standards can eventually rise.

 

But even productive investment has to be financed.

 

And the amount of capital available at any particular price is not unlimited.

 

Imagine an auction at which several exceptionally wealthy buyers suddenly arrive, all determined to acquire the same limited supply of land. The land does not have to produce more for its price to rise. Competition among buyers alone can bid up its price.

 

Something similar can happen in capital markets.

 

The U.S. government needs immense amounts of financing. Technology companies are demanding immense amounts of financing. Defence, energy infrastructure and the reshoring of industrial production require still more.

 

All of these borrowers are competing, directly or indirectly, for the world’s savings.

 

That competition would matter less if the pool of investors willing to provide long-term capital were expanding just as rapidly.

 

There are reasons to think it is not.

 

The traditional buyers are changing too.

For decades, foreign governments, central banks and major institutional investors have been important buyers of U.S. Treasury securities.

 

Some bought them because America ran large trade deficits: countries that accumulated dollars through trade needed somewhere safe and liquid to invest those dollars. Others accumulated Treasuries as part of their foreign-exchange reserves.

 

Several of those traditional sources of demand are now changing.

 

China is the most obvious example. The economic and geopolitical relationship between the United States and China has deteriorated considerably from the era in which China’s rapidly growing trade surpluses routinely translated into very large purchases of U.S. government debt. Beijing has also sought greater diversification of its reserves and reduced financial dependence on the United States.

 

The Gulf states face a different calculation. Countries that accumulated enormous financial surpluses from energy exports increasingly want to deploy more of that wealth domestically — financing economic diversification, infrastructure and new industries. More recently, the U.S. conflict with Iran has created additional demands for spending on defence and on repairing damaged infrastructure.

 

Japan presents yet another problem.

 

Japan has long been one of the world’s largest holders of foreign assets, including U.S. government and corporate bonds. But the yen has come under severe pressure. If Japanese authorities or financial institutions need to bring money home to support the currency or meet domestic financial needs, some foreign assets may have to be sold.

 

This does not mean China, Japan or the Gulf states will suddenly abandon U.S. Treasuries.

 

The important point is subtler.

 

Buyers who once had powerful structural reasons to keep accumulating American debt may now have stronger competing uses for their money.

 

If they buy less — or in some circumstances become sellers — somebody else must absorb the enormous quantity of bonds being issued.

 

Those replacement investors are likely to be more price-sensitive.

 

They will still lend to the United States, but they may require a higher return before doing so.

 

That changes the balance between borrower and lender.

 

For many years, abundant global savings helped suppress the cost of capital. Governments, companies and investors became accustomed to that world.

 

We should not assume it will automatically return.

 

Why the usual cure may work less well

This is where the present situation becomes particularly interesting.

 

Normally, high interest rates contain the seeds of their own reversal.

 

When borrowing becomes sufficiently expensive, people borrow less. Companies cancel marginal investments. Consumers postpone purchases. Housing slows. Economic growth weakens. Demand for credit falls.

 

At the same time, if inflation is the problem, the Federal Reserve can raise short-term interest rates until demand cools and inflation subsides. Once inflation has been contained, monetary policy can eventually become easier.

 

In other words, the economic system contains brakes.

 

But consider two of the borrowers now adding especially large amounts of new demand for capital: the United States government and the technology companies building AI infrastructure.

 

The government does not behave like a household confronting a higher mortgage rate.

 

Higher interest rates do not automatically cause Congress to reduce the budget deficit. Indeed, they initially make the fiscal problem worse because the government must spend more servicing its existing debt.

 

Technology companies respond differently, but the result may be similar.

 

Their AI investments may be less sensitive to higher interest rates than ordinary business investment because company decision makers may regard the current technological race as critical to their companies’ future. 

 

If a chief executive believes AI could determine which companies dominate that executive’s industry — whether computing, advertising, commerce, software or cloud infrastructure — over the next several decades, postponing investment may appear far more dangerous than paying another percentage point or two for capital.

 

There is also a competitive dynamic.

 

If one company slows its investment while its rivals continue building, it risks falling permanently behind. Each company therefore has an incentive to keep spending partly because it expects the others to do so.

 

The result can resemble an arms race: the cost of participating rises, but the perceived cost of not participating rises faster.

 

So higher interest rates may not quickly discipline either of these exceptionally large borrowers.

 

Someone else must therefore adjust.

 

And that is where this becomes an affordability story

Consider a family hoping to buy its first home.

 

Its income may have risen. The family may have accumulated a deposit. The price of the desired house may not initially have changed.

 

But a much higher mortgage rate can increase the monthly payment sufficiently to put the property beyond reach.

 

Consider another younger household still renting and trying to build the savings required for a deposit. Higher borrowing costs can make houses less affordable at exactly the same time that rent and other living expenses absorb the income from which those savings must be accumulated.

 

Now compare those households with an older, wealthier household that already owns its home outright and holds substantial savings.

 

Higher interest rates may benefit that household. Its bank deposits, bonds and other interest-bearing investments can generate more income.

 

This reveals a distributional consequence of higher rates that can easily disappear inside economic statistics.

 

The same financial environment can increase the income available to people who already hold substantial cash and interest-bearing assets while raising the barriers faced by people who need to borrow in order to acquire assets.

 

Those with financial assets receive higher returns.

 

Those who need to borrow in order to acquire assets — a first home, a business, productive equipment — face a higher entry price.

 

Over time, this can reinforce and exacerbate a divide.

 

People who already own appreciating or income-producing assets have more opportunities to compound wealth. People attempting to enter the system must clear a progressively higher financial hurdle before they can begin accumulating those assets themselves.

 

This is one reason an apparently technical argument about bond yields can become a deeply felt political issue.

 

People may not describe their frustration in terms of Treasury yields or the cost of capital. They experience it as something much more concrete:

 

I earn a reasonable income, so why does buying a home feel further out of reach?

 

Why can my parents earn 5 per cent on their savings while my mortgage would cost considerably more?

 

Why can the largest corporations finance enormous investments while a small company struggles to justify borrowing for expansion?

 

The politics may be poorly defined, but the underlying economic experience is real.

 

Higher interest rates do not affect everyone equally.

 

America does not experience this change alone

There is another complication.

 

U.S. Treasury securities are woven into the global financial system. When American yields rise, investors compare opportunities elsewhere against them.

 

Why lend to another government at 3 per cent if a U.S. Treasury offers close to 5 per cent with what investors regard as relatively low credit risk?

 

Other governments may therefore have to offer higher yields to keep attracting capital.

 

That pushes their borrowing costs higher as well.

 

But the financial relationship runs in both directions.

 

Suppose Japanese authorities decide they need to support the yen. Doing so can require them to mobilise dollar assets or other foreign reserves, potentially reducing their need or capacity to hold U.S. Treasuries. Private Japanese institutions may also reduce overseas holdings as domestic financial conditions change. If those sales include U.S. Treasuries, the additional selling pressure can push Treasury prices down and yields higher.

 

Or consider a European government facing concerns about its finances.

 

If investors become nervous about French or British government debt, they may rapidly reorganise portfolios. Banks, pension funds and investment managers holding those securities may need to sell other assets, raise cash or reduce risk elsewhere.

 

Because global institutions often own assets across many markets simultaneously, stress in one sovereign bond market can trigger buying and selling in another.

 

Large movements in currencies can amplify the process. So can the use of borrowed money by financial institutions and investment funds: when losses appear in one part of a portfolio, assets elsewhere may have to be sold to meet cash or collateral requirements.

 

That is how financial pressure can travel back into the United States.

 

The relationship is therefore not simply: higher American yields → higher global yields.

 

It can become a feedback loop: American financial pressure affects overseas markets → overseas institutions adjust → those adjustments create new pressure in American markets.

 

What appears to be a domestic American problem is better understood as part of an interconnected global repricing of capital.

 

The great uncertainty: what AI produces in return

There is an optimistic version of this story.

 

The enormous investment now being made in artificial intelligence could produce equally enormous productivity gains.

 

Productivity is the ability to generate more output from the same quantity of labour and capital. Sustained productivity growth is one of the few ways an economy can become genuinely richer rather than merely redistribute existing wealth.

 

If AI allows businesses to produce more efficiently, invent faster and create entirely new industries, future national income could rise enough to justify much of today’s extraordinary investment.

 

More income also makes debt easier to carry.

 

In that world, today’s intense competition for capital would represent the difficult financing phase of a major technological transformation.

 

History gives us good reasons both for optimism and for caution.

 

Railways transformed nineteenth-century economies. They connected markets, reduced transport costs, enabled new industries and created enormous gains for society. But many railway companies failed, and many of the investors who financed them lost fortunes.

 

The internet offers a more recent parallel.

 

The technology unquestionably transformed commerce, communications and the global economy. Yet hundreds of companies created during the dot-com boom of the late 1990s, which culminated in the crash beginning in 2000, disappeared, and enormous amounts of investor capital were destroyed.

 

The lesson is not that transformative technologies are bad investments. It is that the economic value created by a technology and the financial return earned by a particular investor are different things.

 

A technological revolution can enrich society enormously while individual companies overbuild, borrow too much, invest at the wrong price or lose to competitors.

 

That distinction will matter greatly with AI.

 

The relevant question is therefore not whether every AI investment will earn an adequate return. We already know that some will not.

 

The more useful questions are:

 

Which parts of the AI investment boom are likely to generate returns sufficient to justify the enormous amounts of capital being committed?

 

Which companies possess durable economic advantages, and which are spending defensively because they are afraid of being left behind?

 

And how much unproductive or duplicated investment will occur before those differences become apparent?

 

What should we pay attention to now?

 

For most people, predicting the precise level of the 30-year Treasury yield is neither possible nor particularly useful.

 

There are more valuable questions to ask.

 

First: Is demand for borrowed money continuing to grow faster than the pool of investors willing to provide long-term capital at current interest rates?

 

Government deficits give us one indication of how much additional financing Washington requires.

 

Treasury issuance tells us how much new government debt the market must actually absorb.

 

Corporate borrowing by technology and infrastructure companies shows how much additional competition for capital is coming from the private sector.

 

These are therefore different observable manifestations of the same underlying question: how much new demand for capital is entering the market?

 

Second: Are traditional buyers of U.S. debt continuing to step back?

 

The behaviour of foreign governments, central banks and major institutional investors matters because enormous issuance is easier to absorb when dependable buyers are willing to keep increasing their holdings.

 

Third: Does the AI investment boom begin producing measurable productivity gains?

 

That is ultimately what could transform today’s exceptional borrowing from an enormous financial burden into productive investment.

 

Fourth: Does American fiscal policy change?

 

A large and growing economy can carry substantial debt. But debt that continually rises faster than the government’s capacity to service it gradually consumes resources that could have been used elsewhere and leaves policymakers with fewer choices when the next crisis arrives.

 

And finally: Where is the adjustment actually appearing?

 

Housing is an obvious place to look. So are automobile sales, commercial property, leveraged businesses, consumer credit and other activities that depend heavily on affordable financing.

 

These are not peripheral indicators. They tell us who is being crowded out when the largest borrowers continue borrowing despite higher rates.

 

A different way to think about the years ahead

For much of the period following the 2008 global financial crisis, interest rates remained exceptionally low.

 

Central banks cut policy rates dramatically and purchased enormous quantities of bonds to stabilise financial markets and support weak economies. In many advanced economies, inexpensive money persisted for more than a decade.

 

Then, when the Covid-19 pandemic struck in 2020, central banks again drove interest rates towards zero and launched another extraordinary round of bond purchases to prevent an economic collapse. In the United States and elsewhere, borrowing costs remained at or near historically low levels for an extended period.

 

Households, companies, investors and governments therefore experienced not one brief episode of cheap money, but more than a decade in which exceptionally low — and at times near-zero — interest rates repeatedly became the economic norm.

 

They gradually became accustomed to it.

 

That assumption entered investment models, government budgets, property valuations and household expectations.

 

We may now be discovering that inexpensive capital was a feature of a particular economic era rather than a permanent condition.

 

If so, one of the most useful changes in thinking is surprisingly simple.

 

Do not ask only: When will interest rates fall again?

 

Ask instead: What if capital remains more expensive than we became accustomed to?

 

That question leads to different decisions.

 

A household might place greater value on financial resilience before taking on a large mortgage.

 

A business might test whether an investment still works if financing remains expensive for several years rather than assuming rates will soon return to previous lows.

 

An investor evaluating the AI boom might ask not simply whether artificial intelligence will transform the economy — it probably will — but which companies, technologies and infrastructure investments will capture enough of the resulting economic value to justify the price being paid for them today.

 

A government might have to confront trade-offs that inexpensive borrowing once allowed it to postpone.

 

And a citizen trying to understand the political consequences might ask a broader question:

 

Who benefits, and who loses, when access to capital becomes more expensive?

That may prove to be one of the defining economic questions of the next several years.

 

The current rise in bond yields may yet prove temporary. Energy prices could fall. Fiscal policy could improve. Foreign demand for Treasuries could strengthen. AI investment could produce spectacular gains in productivity.

 

But there is another possibility worth taking seriously.

 

We may be entering a world in which an extraordinary appetite for capital is colliding with a greater reluctance to supply it cheaply.

 

If that is happening, higher borrowing costs are not simply an inconvenience imposed by the Federal Reserve or another temporary episode of financial-market turbulence.

 

They are the mechanism through which the economy decides whose ambitions can be financed, whose must be postponed, and how the costs of scarce capital are distributed across society.

 

And that is why what happens in the bond market matters far beyond Wall Street.

 

I wonder, what are you assuming about the future cost of capital in the decisions you are making now?