Apollo Economist Warns AI Spending Boom Could Keep Interest Rates Higher for Longer

The artificial intelligence investment boom is weakening the traditional influence of higher interest rates on corporate spending, potentially prolonging borrowing-cost pressures as technology companies continue pouring billions of dollars into data centers, chips and computing infrastructure, according to Torsten Sløk, chief economist at Apollo Global Management.

Sløk argued on Thursday that interest rates are unlikely to decline meaningfully while investors continue to expect returns from artificial intelligence investments to exceed prevailing borrowing costs. The implication is that monetary tightening may have less influence on one of the biggest drivers of capital expenditure in the US economy, complicating efforts by the Federal Reserve to control inflation without undermining economic growth.

“The implication for investors is that upward pressure on rates persists until either the Fed raises rates above expected AI returns, or expected AI returns fall below the Fed funds rate,” Sløk wrote in a blog post.

His assessment exposes a growing tension in the financial markets. While rising Treasury yields typically increase financing costs and discourage businesses from committing to large capital projects, technology companies appear increasingly willing to sustain their AI spending because they expect the eventual returns to justify the expense.

That confidence is supporting investment even as long-term borrowing costs rise. Sløk noted that the yield on the 10-year US Treasury note has increased by approximately 125 basis points over the past year, yet Wall Street analysts have continued to raise their capital expenditure expectations for 2027.

That is seen as an indication that higher interest rates alone may not be sufficient to slow AI-related investment while companies believe the technology will deliver returns well above the cost of capital.

Sløk’s argument centers on the relationship between expected investment returns and the cost of borrowing. In conventional business investment, higher financing costs make projects less attractive, particularly when their expected returns are modest. Companies may postpone expansion, reduce capital expenditure, or redirect funds towards debt repayment when the cost of financing rises.

AI infrastructure investment appears to be operating under different assumptions. Sløk said many investors expect returns exceeding 10%, substantially above the Federal Reserve’s benchmark interest rate of approximately 4%.

“The reason is that many AI investors expect AI investments to return more than 10%, well above the 4% Fed funds rate,” he wrote. “With borrowing costs far below expected returns, the AI spending boom is not sensitive to interest rates, and tighter policy is losing its bite on the economy.”

Under this framework, technology companies can justify continued spending even when financing conditions tighten, provided their expected returns remain sufficiently attractive. The result is that monetary policy may have a weaker immediate effect on investment than it would in sectors where borrowing costs are a more decisive factor in capital allocation.

The implications extend beyond technology companies. If AI investment remains strong despite elevated interest rates, it could sustain demand for construction, electricity, advanced semiconductors, networking equipment, and other infrastructure required to operate large-scale computing systems. That resilience could support economic activity while making it more difficult for the Federal Reserve to cool demand through conventional interest-rate policy.

However, expected returns are not the same as realized profits. AI companies and their investors must still demonstrate that the revenue generated by new services, enterprise adoption, and productivity improvements can support the enormous cost of building and operating AI infrastructure.

A projected return above 10% can justify aggressive investment decisions today, but disappointing commercial adoption or declining prices for AI services could weaken those assumptions. If expected returns fall, companies may reassess projects, potentially reducing capital expenditure even without further monetary tightening.

For now, Sløk’s analysis suggests that investors remain confident enough in AI’s future earnings potential to continue expanding their spending plans despite higher yields.

Sløk has previously suggested that interest rates could approach a peak within the following month, citing an improving outlook for oil prices. A potential peace agreement in the Middle East could temporarily ease inflation concerns, lower Treasury yields and reduce pressure on central banks to maintain restrictive monetary policy.

Nevertheless, he does not believe an improvement in the geopolitical outlook would resolve the deeper pressures facing the bond market.

“The fundamental direction of travel for rates remains higher, especially with ongoing fiscal challenges and heavy Treasury issuance adding to the term premium,” he said. “The bottom line for both public and private markets is that higher for longer continues.”

The gap between temporary market relief and structural borrowing-cost pressures is considered central to his outlook. Lower oil prices could reduce headline inflation and ease some of the pressure on government bond yields. But persistent fiscal deficits and heavy Treasury issuance could continue to push up the compensation investors demand for holding longer-dated government debt.

That compensation, known as the term premium, can rise when investors demand greater protection against uncertainty over inflation, the future supply of government bonds and the long-term path of interest rates. Consequently, long-term borrowing costs can remain elevated even when markets anticipate that the Federal Reserve may eventually reduce its benchmark rate.

This situation has created a difficult environment for businesses and investors. Companies financing projects through debt may face higher costs for longer, while government borrowing requirements add another source of pressure to the Treasury market. AI companies with sufficiently high expected returns may continue spending, but other sectors with thinner margins or weaker growth prospects could be more exposed to tighter financial conditions.

There is also a potential feedback loop. If AI investment remains resilient while other areas of the economy respond to higher rates, the distribution of economic activity could become increasingly uneven. Technology infrastructure spending may continue to support demand even as interest-sensitive businesses and households face mounting financing costs.

The major risk is that markets may be relying on optimistic assumptions about AI profitability to justify an investment cycle of extraordinary scale. If those returns materialize, technology companies could sustain spending despite higher rates, reinforcing Sløk’s argument that conventional monetary tightening is losing some of its effectiveness. If returns disappoint, however, the same investment boom could become vulnerable to a sharp reassessment of valuations and capital expenditure plans.

For investors, the message is that a decline in oil prices or a temporary easing of geopolitical tensions may not be enough to produce a sustained fall in borrowing costs. The interaction between AI investment, government debt issuance and the term premium could keep long-term yields elevated even if the near-term inflation outlook improves.

Sløk’s warning therefore extends beyond the next Federal Reserve decision. As long as AI investors expect returns comfortably above borrowing costs and the US government continues to issue large volumes of debt, the forces supporting higher interest rates may remain in place.