Rising long term bond yields are often interpreted as a warning about persistent inflation, large government borrowing requirements and uncertainty over future interest rates. But a different explanation is gaining attention: investors may also be demanding higher returns because they expect the massive investment in artificial intelligence to increase economic productivity and growth.

Jacob Manoukian, United States head of investment strategy at JPMorgan Private Bank, has argued that this more positive interpretation deserves greater attention. In his view, the bond market may be anticipating a stronger productivity cycle as companies spend heavily on artificial intelligence infrastructure, data centres, computing equipment and related technologies. The argument is not that higher yields prove an artificial intelligence productivity boom is coming, but that markets may be incorporating at least some expectation of stronger future economic growth.

That interpretation matters because the same artificial intelligence investment is simultaneously creating a large increase in corporate borrowing. Companies building data centres and other infrastructure are raising substantial amounts of debt, adding to the supply of bonds competing for investor capital. This creates two forces that can push yields higher: investors may expect stronger future growth, while companies are issuing more debt to finance the investment required to produce that growth.

Recent Federal Reserve analysis provides some support for the economic side of the argument. Business fixed investment increased at a strong pace in the first quarter of 2026, with much of the increase linked to infrastructure supporting artificial intelligence services. The central bank has also noted strong productivity growth and continued investment in high technology.

Higher yields may contain a growth signal

Bond yields reflect several factors at the same time. Inflation expectations can push yields higher because investors demand compensation for the declining purchasing power of future interest payments. Government borrowing can increase the supply of bonds that need to be absorbed by investors. Expectations about central bank interest rates also influence yields, particularly at shorter maturities.

Long term yields, however, can also respond to expectations about economic growth. If investors believe an economy will become more productive, they may expect stronger corporate earnings, higher investment and potentially greater demand for capital. A stronger economy can justify higher interest rates over time and therefore higher long term bond yields.

This is the more optimistic interpretation of the current bond market that Manoukian has highlighted. Instead of assuming that rising yields are entirely a consequence of fiscal concerns and inflation, investors could be partly pricing in the economic gains expected from artificial intelligence investment.

There is already evidence that the investment cycle is substantial. The Federal Reserve has reported that business investment has been particularly strong in areas connected with artificial intelligence infrastructure, while construction spending on data centres has increased sharply. The technology investment is also supporting demand for computers, electronic equipment and machinery.

The important qualification is that investment is not the same as productivity. Companies can spend enormous amounts on new technology without immediately generating equivalent increases in output. The ultimate economic return will depend on how effectively businesses use artificial intelligence to reduce costs, increase production or create new products and services.

Corporate borrowing is adding pressure to the bond market

The artificial intelligence investment cycle is also changing the supply side of the bond market. Technology companies that historically relied heavily on large cash balances are increasingly turning to debt to finance data centres and other infrastructure.

AI related debt issuance has exceeded $220 billion this year, roughly twice last year’s amount, while total United States corporate bond issuance has reached about $1.68 trillion, up nearly 27 percent from the same period in 2025. )

This matters because investors have a finite amount of capital to allocate across different bonds. If large technology companies issue substantial quantities of long dated debt, they compete with the United States government and other borrowers for investors’ money.

That does not mean corporate borrowing automatically pushes Treasury yields higher. Investors can expand their overall allocation to bonds, and foreign buyers, banks, insurers and other institutions can absorb additional supply. But a substantial increase in corporate issuance can change relative pricing, particularly if investors demand higher compensation for holding corporate debt instead of government securities.

The effect is particularly relevant at the long end of the market because much of the borrowing required for data centres and other infrastructure involves large amounts of capital that may need to be financed over extended periods.

Manoukian has suggested that artificial intelligence related debt issuance could approach a substantial share of United States Treasury coupon issuance by the end of the year. If that occurs, the competition for long term bond investors would become more significant, although the two types of debt would continue to have very different risk characteristics.

Productivity gains could change the economic calculation

The strongest version of the optimistic argument is that artificial intelligence investment could eventually increase the economy’s productive capacity. If companies can produce more goods and services with the same amount of labour and capital, productivity growth could accelerate.

Higher productivity can support faster economic growth without necessarily creating the same inflation pressure that would result from an economy growing simply because demand is increasing. Greater productive capacity can allow businesses to meet additional demand while increasing output more efficiently.

This is one reason the Federal Reserve has been paying close attention to artificial intelligence investment. Its recent analysis said strong productivity growth was helping offset relatively subdued labour force growth, while high technology investment was contributing to economic activity.

However, the timing remains uncertain. Large investments in technology infrastructure can take years before their full economic benefits become visible. Data centres must be built, computing systems installed, software developed and businesses reorganised around new tools. Workers may also require training before artificial intelligence produces meaningful improvements in output.

The productivity effect could therefore be substantial but uneven. Some companies may achieve major efficiency gains while others spend heavily without generating comparable returns.

The bond market is also confronting inflation and debt

The productivity explanation should not be used to dismiss the more traditional reasons for higher yields. United States inflation remains well above the Federal Reserve’s 2 percent target. The personal consumption expenditures price index increased 3.7 percent over the year through July, while core inflation remained at 3.3 percent. ([Reuters][4])

Federal Reserve Chair Kevin Warsh has also recently indicated that policymakers could raise interest rates if inflation does not move sufficiently toward the target. Market expectations for a September increase rose sharply following his remarks, demonstrating that monetary policy remains an important driver of bond yields. ([Reuters][5])

Government borrowing is another factor. Large fiscal deficits mean that the Treasury needs to issue substantial quantities of debt, creating a continuing supply of government bonds that must be absorbed by the market. Concerns about long term fiscal sustainability can also cause investors to demand higher yields.

The current rise in long term yields should therefore be viewed as the result of several forces operating simultaneously. Artificial intelligence productivity expectations may be one of them, but inflation, monetary policy and government debt remain significant.

Semiconductor valuations show another side of the AI cycle

The bond market is not the only place where investors are reassessing the artificial intelligence investment cycle. Semiconductor companies, which supply much of the hardware required for artificial intelligence systems, have experienced substantial price corrections even while the longer term investment outlook remains strong.

Manoukian argues that the correction has created potential opportunities because markets may already be pricing in a peak in semiconductor earnings. He believes earnings could continue to increase if companies achieve the sales already expected by analysts and investors remain willing to value those businesses at similar multiples in future years.

That argument highlights the difference between an industry’s long term growth prospects and the price investors are willing to pay for them. Artificial intelligence can generate strong demand for chips and computing infrastructure while individual stocks can still become overvalued.

The same distinction applies to the bond market. Strong future productivity could eventually support economic growth, but that does not mean every company investing in artificial intelligence will generate sufficient returns to justify its borrowing.

Investors are favouring income over long duration

The uncertainty surrounding long term yields is also influencing fixed income strategies. JPMorgan Private Bank currently prefers shorter duration credit, where investors can receive income without taking as much exposure to movements in long term interest rates.

That positioning reflects the difficulty of determining where long term yields will eventually settle. If artificial intelligence investment produces stronger productivity and growth, yields could remain higher than investors had previously expected. If inflation remains persistent or government borrowing continues to increase, yields could face additional upward pressure.

At the same time, a slowdown in economic activity or a failure of artificial intelligence investment to generate expected productivity gains could produce the opposite outcome. Lower growth could eventually reduce long term yields and increase demand for government bonds.

The bond market is therefore caught between competing interpretations of the same investment boom. Artificial intelligence spending is increasing corporate borrowing and bond supply, but it could also increase productivity and economic growth. The first effect can push yields higher through financing demand, while the second can justify higher yields through stronger future economic activity.

That makes the current rise in yields more complicated than a simple story of rising government debt or inflation. Manoukian’s interpretation remains an investment view rather than a settled explanation of market movements, but it highlights an important possibility: **the bond market may be demanding higher returns partly because investors expect today’s artificial intelligence spending to produce a stronger economy tomorrow.**

(Adapted from TradingView.com)

Categories: Economy & Finance, Regulations & Legal, Strategy