Kevin Warsh is chairing his first rate-setting meeting at the US Federal Reserve this week, and he has a theory he wants the committee to buy: artificial intelligence will let interest rates come down. Most of his new colleagues are not convinced.
Donald Trump appointed Warsh expecting rate cuts, which the president has demanded again and again. Warsh calls AI “structurally disinflationary” and says the central bank should judge the economy by where AI is taking prices, not only by the latest data. He describes the technology as “the most productivity-enhancing wave of our lifetimes,” one that would let the economy grow without pushing up inflation.
Other Fed officials read the same trend the opposite way. Vice chairman Philip Jefferson said earlier this year that “all other things being equal, persistent increases in productivity growth are likely to result in an increase in the neutral rate, at least temporarily.” The neutral rate is the interest rate that neither slows nor speeds up growth. Governor Michael Barr has tied faster productivity to higher rates too, and told listeners AI was unlikely to be a reason to lower the policy rate. Mary Daly, who heads the San Francisco Fed, argues that demand for investment would outrun the supply of savings.
Economists agree AI will lift productivity. They split on what follows, and the weight of opinion leans toward higher inflation and higher rates.
The cost of building first
The reasoning starts with timing. What matters is whether productivity gains arrive faster than the bills for deploying the technology, and the industry is still at the start of that build-out.
US investment in AI came to about $US375 billion ($533 billion) last year. This year it is expected to reach about $US750 billion, and next year it could pass $US1 trillion. Goldman Sachs has said $US7.6 trillion could go into AI over the next five years. Most of the money pays for computing power and data centres, which in turn need power grids and water supply.
Shortages of semiconductors have pushed chip prices sharply higher. Copper is dearer. Households are paying more for electricity and for anything electronic, and workers with AI skills command higher pay. AI companies are also competing hard for capital, both equity and a growing share of debt, so borrowing costs should climb across the economy. Yields on 10 and 30-year US Treasury bonds have risen materially since OpenAI launched ChatGPT in November 2022, and some of that may be AI’s doing.
Economists call the stretch before any payoff a “productivity J-curve”: costs first, gains later. Even if AI eventually lifts returns and growth, it would probably cut savings rates and push the neutral rate up.
Where the Fed stands
The Fed cut its policy rate three times last year. The target range for the federal funds rate is now 3.5 per cent to 3.75 per cent. If the neutral rate has moved above that, policy is stimulating the economy, at a time when inflation is 3.8 per cent and rising because of Trump’s tariffs and the war in the Middle East.
A cut at this week’s Federal Open Market Committee meeting looks most unlikely. The Fed is more likely to hint that its next move could be a rise. Warsh has one vote on the committee, and the years-long wait for AI’s productivity benefits means his theory is unlikely to be tested soon.
A trap for AI funding
Raising rates carries its own risk. The big three AI start-ups, SpaceX, Anthropic and OpenAI, are planning public offerings that would raise more than $US200 billion and value them at more than $US4 trillion. Alphabet, Amazon, Meta and Microsoft are finding AI spending is soaking up their cash, forcing them to borrow and, in some recent cases, sell shares.
Governments are borrowing heavily as well. US gross government debt will reach $US40 trillion by September, about 125 per cent of GDP, with deficits of 6 per cent of GDP. Ordinary families have less to fall back on: the US savings rate dropped from 3.7 per cent of disposable personal income in the first quarter to 2.6 per cent in April, as tariffs and the Iran war push up living costs.
Money for AI stays easy to find while rates are low and share prices sit at record highs. Futures markets have started pricing a rate hike in December. A hike, or the prospect of several, could puncture the sharemarket boom that AI companies rely on to raise funds. If the Fed has to raise rates to fight AI-driven inflation, the argument over productivity may stay academic for now.
Sources
- Will the AI boom lead to lower interest rates?
- Will the AI boom lead to lower interest rates?
- There Are Three Paths Forward for the AI Boom. Two of Them Lead to Lower Interest Rates
