Will the AI boom lead to lower interest rates?
Kevin Warsh, who currently chairs the first rate-setting meeting at the Federal Reserve Board, argues that the boom in artificial intelligence will allow for significant declines in US interest rates. His new colleagues and many others are not so sure.
Warsh, who was appointed by Donald Trump with the expectation that he would try to implement the interest rate cuts that Trump constantly and persistently demanded, argued that artificial intelligence would be “structurally disinflationary” and that the US central bank should make a proactive decision that artificial intelligence would reduce the inflation rate, rather than reacting to the data of the moment.
He argued that AI would unleash a productivity boom, “the most productivity-enhancing wave of our lifetime,” that would enable non-inflationary growth and allow interest rates to fall.
Other Fed officials beg to differ.
Philip Jefferson, the Fed’s vice chairman, said earlier this year that “all other things being equal, persistent increases in productivity growth will likely lead to at least a temporary increase in the neutral interest rate.” (The neutral rate is the interest rate that neither limits nor stimulates growth, allowing the economy to expand while maximizing growth with a stable inflation rate).
By the time the Fed’s Federal Open Market Committee meeting concludes this week, Trump is unlikely to get the rate cut he so desperately wants.
Another Fed Governor, Michael Barr, also linked high productivity growth to higher interest rates and said it was unlikely AI would be a reason for the Fed to lower its policy rate. Meanwhile, San Francisco Fed President Mary Daly said accelerating AI-driven productivity growth would require a higher neutral rate because investment demand would grow faster than the supply of savings.
There is little debate about whether artificial intelligence will increase productivity rates, but there is serious disagreement among economists about how this will play out and what this could mean for inflation and interest rates. The weight of opinion seems to be in favor of higher inflation and higher interest rates.
The argument in support of a higher neutral rate is relatively simple.
Whether or not AI leads to higher productivity, what really matters in terms of inflation and interest rates is whether it increases the rate of productivity faster than the costs of using it.
We are in the early stages of AI deployment. Companies developing the technology and its applications are investing unprecedented amounts, at a dramatically accelerating pace, to build the infrastructure needed to train their models and deliver the necessary computing power.
Last year, investment in AI in the US was approximately US$375 billion ($533 billion). This year it is expected to be around 750 billion dollars. It could exceed $1 trillion next year. Goldman Sachs said $7.6 trillion could be spent on artificial intelligence in the next five years.
Most of this spending is on computing power and data centres, which also requires investment in energy grids and water supply.
This has flow-on effects, from the rising cost of semiconductors causing significant shortages in the face of demand causing extraordinary price increases, to rising prices of commodities like copper, and rising consumer costs for everything energy and electronics. It also increases the cost of labor with artificial intelligence knowledge and skills.
With AI companies competing aggressively for capital (equity and increasingly debt), the cost of capital is set to rise more broadly. The significant increase in yields on 10- and 30-year US Treasury bonds since OpenAI launched ChatGPT in November 2022 may reflect some of this AI impact.
Whatever the ultimate impact of AI on productivity and inflation, it will be inflationary in the transition from here to there because the costs of deploying and deploying AI will – at least at this stage – inevitably rise faster than any rate of productivity gains. Economists call this “productivity J curve.”
Economists will also say that even if AI significantly increases productivity, returns on capital, and economic growth, it will likely reduce savings rates and push the neutral interest rate higher rather than lower.
The Fed may already be doing this, although its current policy stance and three rate cuts last year do not yet reflect this. A neutral rate above the policy rate (currently the target range for the federal funds rate is 3.5 percent to 3.75 percent) would mean that monetary policy adjustments are stimulative.
With an inflation rate of 3.8 percent and rising (thanks to Trump’s tariffs and the war in the Middle East), the Fed shouldn’t be stimulating the economy.
By the time the Fed’s Federal Open Market Committee (FOMC) meeting concludes this week, Trump is unlikely to get the rate cut he so desperately wants. In fact, it is more likely that the Fed will signal the next rate hike.
The fact that there will likely be a transition phase lasting years before the putative productivity benefits of massive investments in AI reach a meaningful degree, and that this phase will likely be inflationary, means that Warsh’s theory of productivity-driven disinflation is unlikely to be tested any time soon.
Even if he wants to advance the case, his colleagues at the Fed are unlikely to cooperate, and he has only one vote on the FOMC.
The risk for the Fed is that it will fall behind the curve in tightening monetary policy in parallel with the possible increase in the neutral interest rate.
But under these circumstances, monetary policy has an Article 22 element.
The demand for capital to finance AI developments and related infrastructure is huge and growing. So far, investors’ appetite for all things AI has been satisfied.
Initial public offerings proposed by the three major AI start-ups (SpaceX, Anthropic and OpenAI) will amount to over $200 billion, valuing them at over $4 trillion.
Even megatech companies like Google parent Alphabet, Amazon, Meta, and Microsoft are finding that their spending on AI is sucking up their cash flows, forcing them to take on debt and, in some recent cases, raise equity capital.
This is happening at a time when the US government, like most developed economy governments around the world, is struggling with debt and increasing that total every day.
The US’s gross government debt will reach US$4 trillion (about 125 percent of GDP) by September, and budget deficits of up to 6 percent of GDP will ensure total debt levels continue to climb.
The U.S. savings rate is falling rapidly, from 3.7 percent of disposable personal income in the first quarter of this year to 2.6 percent in April; Households are squeezed by rising costs of living resulting from Trump’s tariffs and the Iran war.
While rates are relatively low and stock prices are at record highs (driven by the frenzy around AI), it is possible to finance huge capital demands from AI companies.
If interest rates rise to offset rising inflation (and futures markets have begun predicting a rate hike in December), this will increase capital costs for the energy and water needs of AI companies and data center infrastructure companies. And potentially, if further rate hikes are deemed likely, it could puncture the stock market and the cycle of rapid and significant increases in value that AI companies rely on to gain access to capital.
If the Fed is forced to raise interest rates to counter AI-fueled inflation and end the stock market rally, debates about productivity and interest rates may remain academic, at least in the short term.
The Business Briefing newsletter delivers big stories, exclusive news and expert insights. Sign up to receive it every weekday morning.
Stephen Bartholomeusz is one of Australia’s most respected business journalists. He was most recently co-founder and deputy editor of the Business Spectator website and associate editor and senior columnist at The Australian.Connect via email.
