Will AI Lower Interest Rates? Why DIY Investors Shouldn't Bet Their Plan on It

Kevin Warsh took over as Federal Reserve chair in May 2026 with a clear view: artificial intelligence is “structurally disinflationary,” it lowers costs and raises productivity, and it argues for lower interest rates.¹ Austan Goolsbee, president of the Chicago Fed, looks at the same AI boom and reaches a different conclusion. “The bigger the hype,” he warned in May 2026, “the more rates would need to rise to prevent overheating.”² Two careful people, the same data, opposite calls.

They disagree partly because they are answering different questions. “Will AI lower interest rates?” sounds like one question, but the policy rate the Fed sets, the real neutral rate the economy gravitates toward, the yield on a 10-year Treasury, and the rate on a 30-year mortgage are distinct things that AI can move in different directions at the same time. This guide separates those questions, weighs what the research actually says, and turns the macro debate into something a DIY investor can act on.

“Interest rates” is several different questions

Most takes you will read treat “rates” as a single dial. They are a stack of related but separate prices, and AI pushes on each one through different channels.

Keep this table in mind. A lot of confident commentary is really an argument about one row presented as if it settled all five.

In the near term, the AI boom looks demand-first

Right now the AI buildout is a real-economy investment boom. The New York Fed reports that the major AI firms committed roughly $300 billion to capital investment in 2025 across semiconductors, power grids, and specialized labor.³ The International Energy Agency puts the largest tech firms’ capital spending above $400 billion in 2025 and expects it to rise by roughly another 75% in 2026.⁴

That matters for rates because the productivity payoff arrives only after deployment, training, and workflow redesign, while the data centers, GPUs, transmission lines, and skilled labor are demanded now. The economy gets the spending impulse before it gets the full productivity gain. This is why Goolsbee’s warning lands: when firms and investors race to build ahead of benefits that have not yet shown up in the data, the boom can overheat demand first. Higher productivity is, in his framing, “the positive supply shock of our dreams,” but only once it actually materializes.⁵

If AI raises productivity, inflation pressure can fall

The disinflation case is real. Higher productivity means more output per worker, which can lower unit costs and let the economy grow faster without as much price pressure. The Chicago Fed connects this directly to the 1990s, when faster productivity growth gave Alan Greenspan room to hold rates steady while the economy ran hot.⁵ That is the Warsh thesis in one line: let supply expand, and inflation takes care of itself.

The caution is timing. Mary Daly of the San Francisco Fed points out that the productivity effects of general-purpose technologies show up slowly. Electricity took decades to move the productivity statistics, and most macro studies so far find limited evidence of a large AI effect on measured productivity.⁶ So the “AI lowers rates” path depends on a specific sequence: productivity gains spread broadly, cost growth slows, inflation cools, and the Fed can ease restrictive policy.

Stronger productivity can also raise the neutral rate

Faster productivity growth can lower inflation pressure and raise the real interest rate the economy can sustain at the same time. The New York Fed defines r-star as the real short-term rate expected to prevail when the economy is at full strength and inflation is stable, and its Holston-Laubach-Williams framework ties r-star to trend growth.⁷

The logic is old and direct. The Richmond Fed, revisiting Al Broaddus’s 1990s argument, states it plainly: higher trend productivity implies higher trend output and consumption growth, and higher trend consumption growth is associated with higher real interest rates.⁸ A household expecting a higher future income borrows more today; a firm seeing more profitable projects invests more. Both raise the demand for capital, which raises the rate that balances saving and investment. So “AI lowers inflation” and “AI lowers rates” are not the same claim. AI could deliver lower inflation alongside higher real rates.

Why the experts split: does AI complement labor or replace it?

The cleanest way to see why economists disagree is an IMF working paper, “From Servers to Rates,” by Giovanni Melina and Stefania Villa. Its result is conditional on one thing: whether AI and information-tech capital complement workers or substitute for them.⁹

• If AI complements labor, the same investment raises employment, wages, consumption, and the natural rate, which calls for tighter policy.
• If AI substitutes for labor, the same investment yields smaller wage and employment gains, and the natural rate can fall, warranting looser policy.

Both stories are internally consistent. That is why two credible economists can run similar logic and end up pointing in opposite directions: they are making different bets about how AI reshapes work.

A serious model says AI could raise r-star by about a point

Lukasz Rachel of University College London builds a capital-market equilibrium model of the neutral rate. In it, AI raises businesses’ demand for capital and reduces households’ desired saving, lifting r-star “perhaps by 1 percentage point over the next five years.”¹⁰ AI is one of the two largest upside scenarios in his framework, alongside a reversal of the premium investors pay for safe assets.

That is not a forecast that rates will rise by a point. It means a serious model finds the “AI raises rates” scenario quantitatively plausible, not a hand-wave. If it is even partly right, a return to the ultra-low rates of the 2010s is unlikely, even once supply catches up with demand.

The credible case for lower rates

The lower-rate path is not far-fetched either, and it runs through several channels. AI could compress prices by cutting costs and stiffening competition. If it displaces workers, or makes them fear displacement, households may save more for precautionary reasons. If the gains flow to capital owners and high-income households, aggregate saving can rise, because richer households save more of each extra dollar. And if AI makes many existing business models obsolete, firms may invest cautiously outside a narrow AI sector.

This maps onto the secular-stagnation logic that explained decades of falling rates. Rachel and Smith, in Bank of England work on the global real rate, attribute most of that multi-decade decline to shifts in desired saving and investment: demographics, rising inequality, weaker investment demand, and a wider spread between safe and risky returns.¹¹ In this world, AI creates abundance, but uneven income and weak demand keep rates low. It is a savings glut born of abundance at the top.

The productivity evidence is still early and mixed

All of these scenarios assume some productivity gain. The micro evidence is encouraging on specific tasks. Brynjolfsson, Li, and Raymond found that access to a generative-AI assistant raised customer-support productivity by 14% on average, with the largest gains, about 34%, going to novice and lower-skilled workers.¹²

Moving from task gains to economy-wide productivity is harder. Daron Acemoglu estimates AI’s total-factor-productivity effect over ten years at no more than 0.66%, and less than 0.53% once you account for harder-to-learn tasks.¹³ Spread over a decade, that is on the order of 0.07% per year. The Penn Wharton Budget Model projects AI raising GDP by about 1.5% by 2035, close to 3% by 2055, and 3.7% by 2075, with the boost to annual growth strongest in the early 2030s before fading.¹⁴ The OECD’s survey captures the spread: optimistic scenarios envision large gains, Acemoglu-style estimates are far smaller, and realization depends on adoption, complementary investment, policy, and keeping human skills aligned with the technology.¹⁵

What this means for DIY investors

1. Do not assume the 2010s are coming back

A common mistake is treating 0% policy rates, 3% mortgages, and a decade of valuation expansion as the natural state of the world. The research does not support that certainty. Secular forces pushed rates down for decades, but AI could counteract some of them if it raises growth and capital demand, since trend growth is one of the drivers in standard r-star models.⁷ Build a retirement plan that works across several rate regimes, not just the low-rate world of 2009 to 2021.

2. Bonds still matter, and so does duration

If AI raises real rates, long-duration bonds can take mark-to-market losses. If AI causes a labor shock, a savings glut, or a deflationary bust, high-quality bonds can help. The answer is to match bond duration to the job each dollar is doing: cash and Treasury bills for an emergency fund, short Treasuries and CDs for near-term spending, intermediate duration for medium-term goals, and high-quality nominal bonds or TIPS for retirement ballast. Speculating on the direction of rates is rarely necessary.

3. Separate equity earnings growth from valuation risk

AI could raise corporate earnings, yet higher rates can reduce the price investors will pay for those earnings. “AI is good for productivity” does not automatically mean “AI stocks are cheap.” A globally diversified portfolio already owns the AI winners at market weight, so you do not need to handpick them. The larger risk is overconcentrating in the obvious names after valuations already embed heroic expectations.

4. Stress-test your mortgage and home plans

AI could lower mortgage rates if productivity cools inflation and long yields fall, or keep them elevated if it lifts r-star, fiscal pressure, or housing demand. The actionable rule is simple: buy based on affordability at today’s rate, and treat a future refinance as upside rather than the base case.

5. Treat your career as part of your balance sheet

AI is not only an asset-pricing story. It affects wages, job security, and skill premiums. A software engineer, analyst, designer, attorney, or marketer may already carry large implicit exposure to AI-driven labor change. That argues for more liquidity, less debt, and avoiding concentrated bets that move with your own career risk, such as loading up on your employer’s stock when your paycheck already depends on the same trend.

The calculator below makes the bond and mortgage pieces concrete. Pick an AI scenario, or set your own rate move, and see what it does to your bond sleeve and your mortgage payment.

Who should care most, and who can skip it

This matters most for:

• DIY investors deciding how much bond duration to hold.
• Early retirees relying on fixed-income returns and withdrawal assumptions.
• Homebuyers wondering whether to wait for lower mortgage rates.
• High-income tech workers whose income and portfolio are both AI-exposed.
• Anyone tempted to overweight AI stocks because “AI changes everything.”

It matters less if you already hold a simple, globally diversified, automated portfolio with no big borrowing or duration bets. For you, the honest answer is to keep contributing and rebalancing. The macro debate is interesting, but it does not change a well-diversified plan.

What Summitward recommends

Do not make an active interest-rate bet based on AI. Build a plan that survives four worlds: high-growth with high rates, high-growth with low inflation, labor displacement with a savings glut, and an AI investment bust. A plan that holds up in all four does not need you to call the winner in advance.

1. Keep broad global equity exposure.
2. Avoid concentrating in AI winners already priced for perfection.
3. Set bond duration by time horizon and spending needs, not forecasts.
4. Add TIPS if inflation is a major planning risk for you.
5. Stress-test mortgages and retirement withdrawals at both higher and lower rates.
6. Count your career as part of your balance sheet.
7. Rebalance on a schedule instead of trying to predict.

Frequently Asked Questions

Will AI lower interest rates?

It might, but that is not the most likely single outcome. AI can lower inflation, which would support lower policy rates, but a durable productivity boom tends to raise the real neutral rate because households and firms borrow and invest against higher expected income. AI could also lower rates if it displaces labor and raises precautionary saving. The directions genuinely conflict, so the safer planning assumption is uncertainty, not a one-way move.

Does AI cause inflation or deflation?

Near term, the AI buildout adds demand through heavy capital spending on chips, data centers, and power, which is mildly inflationary. Longer term, if productivity gains spread, AI can be disinflationary by lowering unit costs. Which force dominates depends on how quickly the productivity payoff shows up relative to the spending.

What is r-star, and why does AI affect it?

R-star is the real interest rate that keeps the economy at full strength with stable inflation. It rises when trend growth and investment demand rise, and falls when desired saving outruns investment. AI can move it either way: higher if it lifts growth and capital demand, lower if it displaces labor and pushes up precautionary or high-income saving. One capital-market model estimates AI could raise r-star by about a percentage point over five years.

Should I buy bonds if AI might raise rates?

Bonds still do a job in a plan, but match their duration to the money. Use cash and short Treasuries for near-term needs, intermediate duration for medium-term goals, and longer or inflation-protected bonds for retirement ballast. That way a rate rise hurts less, and you still hold ballast if AI triggers a downturn instead.

Should I wait for lower mortgage rates because of AI?

Buy based on what you can afford at today’s rate, and treat a future refinance as a bonus, not the plan. AI could lower mortgage rates if it cools inflation, or keep them high if it raises the neutral rate, fiscal pressure, or housing demand. Timing a purchase to a rate forecast you cannot verify is a weak basis for one of your largest financial decisions.

Key Takeaways

• “Rates” is not one number. The policy rate, r-star, long yields, and mortgage rates can move in different directions as AI plays out.
• Near term looks demand-first. Hundreds of billions in AI capex add demand before the productivity payoff arrives, which is mildly inflationary.
• Lower inflation is not the same as lower rates. A productivity boom can cool inflation while raising the real neutral rate.
• The experts split on labor. If AI complements workers, rates likely rise; if it replaces them, rates can fall.
• Build for more than one path. Match bond duration to the job, diversify globally, stress-test your mortgage, and count your career as an asset.

Related Guides

• Four Signs of a Bubble: how to read the AI-investment boom without making a timing bet.
• Concentration Risk: why owning AI winners at market weight beats overweighting them.
• The Safe Withdrawal Rate: how rate regimes feed into retirement spending math.
• An Index Fund Is Not a Financial Plan: why diversification is the start, not the whole answer.
• Rent vs. Buy: stress-testing a home purchase against the rate you actually face.

Sources

1. Reporting on Kevin Warsh’s confirmation as Fed chair (May 2026) and his view that AI is “structurally disinflationary.” finance.yahoo.com.
2. Austan Goolsbee, remarks at the Milken Institute (May 2026): “the bigger the hype, the more rates would need to rise to prevent overheating.” finance.yahoo.com.
3. Simone Lenzu, “AI’s Macroeconomic Challenges and Promises,” Federal Reserve Bank of New York, Liberty Street Economics (May 20, 2026); roughly $300 billion in 2025 AI capital investment. libertystreeteconomics.newyorkfed.org.
4. International Energy Agency, “Energy and AI” (2025); largest tech firms’ capex above $400 billion in 2025, expected to rise about 75% in 2026. iea.org.
5. Austan Goolsbee, “Remarks on Productivity Growth and Monetary Policy,” SIEPR Economic Summit, Federal Reserve Bank of Chicago (Feb 28, 2025). chicagofed.org.
6. Mary C. Daly, “The AI Moment? Possibilities, Productivity, and Policy,” FRBSF Economic Letter (Feb 2026). frbsf.org.
7. Federal Reserve Bank of New York, “Measuring the Natural Rate of Interest” (r-star; Holston-Laubach-Williams). newyorkfed.org.
8. Alexander L. Wolman, “Al Broaddus, Productivity Growth and Monetary Policy in the 1990s,” Federal Reserve Bank of Richmond Economic Brief No. 26-15 (May 2026). richmondfed.org.
9. Giovanni Melina and Stefania Villa, “From Servers to Rates: AI, ICT Capital, and the Natural Rate,” IMF Working Paper No. 2025/224 (Oct 31, 2025). imf.org.
10. Lukasz Rachel, “What Next for r*? A Capital Market Equilibrium Perspective on the Natural Rate of Interest,” Brookings Papers on Economic Activity (Fall 2025); AI raises r-star “perhaps by 1 percentage point over the next five years.” brookings.edu.
11. Lukasz Rachel and Thomas Smith, “Secular Drivers of the Global Real Interest Rate,” Bank of England Staff Working Paper No. 571 (Dec 2015). bankofengland.co.uk.
12. Daron Acemoglu, “The Simple Macroeconomics of AI,” NBER Working Paper No. 32487 (2024); TFP no more than 0.66% over 10 years, under 0.53% after adjusting for hard-to-learn tasks. nber.org.
13. Erik Brynjolfsson, Danielle Li, and Lindsey Raymond, “Generative AI at Work,” NBER Working Paper No. 31161 (2023); 14% average productivity gain, 34% for novices. nber.org.
14. Penn Wharton Budget Model, “The Projected Impact of Generative AI on Future Productivity Growth” (Sept 8, 2025); GDP about 1.5% higher by 2035, near 3% by 2055, 3.7% by 2075. budgetmodel.wharton.upenn.edu.
15. OECD, “The Impact of Artificial Intelligence on Productivity, Distribution and Growth,” OECD Artificial Intelligence Papers No. 15 (April 2024). oecd.org.
16. Framing prompted by The Economist, “Will AI lower interest rates?” (June 25, 2026). economist.com.

This article is educational and is not financial advice. It describes scenarios and research, not predictions. Interest-rate outcomes are uncertain; confirm tax and product details with a qualified professional before acting.