Bitcoin’s AI Bet: Will Interest Rates Rise or Fall?

At first glance, job losses caused by artificial intelligence look bad for Bitcoin. Fewer paychecks mean less money to invest, while lower household spending can weaken economic growth and push investors away from risky assets. However, some Bitcoin analysts argue that the later policy response could benefit Bitcoin: central banks may cut interest rates, governments may increase spending, and additional liquidity may flow into scarce assets. The chain sounds clean, but every link in it remains uncertain.

Two AI Paths for Fed Policy

New macroeconomic modeling on AI, jobs and monetary policy points to two different economic outcomes, each with different implications for interest rates and Bitcoin.

• The AI-slump path: Weak demand and rising unemployment encourage policymakers to cut interest rates and support spending.

• The productivity-boom path: Stronger output and business investment can keep real interest rates higher.

These scenarios also sit inside the deeper question of whether AI could eventually change the role of money. The academic research below explains what could move the economy from one path to another before we assess the consequences for Bitcoin.

The Productivity Paradox

New research asks a strange question: what if machines made everything more efficient, but the economy still shrank? Economists Luca Fornaro and Martin Wolf say that's not a contradiction, it's exactly what could happen.

Picture how much an economy could sustainably produce when its available workers and machines are fully employed. Economists call this potential output. Actual output is what the economy produces and sells in practice.

Potential and actual output do not have to move together. For example:

• A family car purchase makes the difference clear: A factory can produce 100 cars, but families worried about losing their jobs to AI buy only 70 cars. The factory therefore produces 70, even though it could make 100. Its potential output is therefore 100 cars, while its actual output is 70.

The example shows how AI automation can shift income away from workers and toward the people and companies that own the machines and software. This matters because the two groups use income differently, and that difference helps determine whether the economy grows or merely becomes more efficient on paper.

The Income Split That Slows an Economy Down

Workers tend to spend most of what they earn. Wealthy owners tend to save more of theirs. So when automation makes something like car manufacturing more efficient, and AI takes over work people used to get paid for, more of that income flows to owners instead of workers. 

Workers collectively earn less, and spend less, even though the economy can now produce more cars than ever.

• Higher capacity, weaker demand: Automation raises how much the economy could produce, but the shift in income reduces household spending. Businesses respond by producing less, leaving actual output below its potential.

• A self-defeating incentive: Fornaro and Wolf’s model shows why automation can make sense for each company but create problems across the economy. 

One firm can lower costs and become more competitive by automating. When many firms automate, however, the resulting loss of worker income can weaken overall demand and reduce sales, including for the companies that automated.

The second point is sharper: efficiency at the firm level does not guarantee growth at the economy level. Fornaro and Wolf call this a paradox of productivity: what works for one company doesn't automatically work when every company does it.

Why Can’t Companies Just Lower Prices and Sell More?

Fornaro and Wolf do not directly test whether widespread price cuts could prevent an AI slump. Their model instead examines the loss of spending caused when automation shifts income toward owners who tend to save a larger share of what they earn.

• One company can win customers: Cutting prices may persuade buyers to choose one company’s products over a competitor’s.

• Every company faces the same limit: If workers across the economy have less money, businesses are competing for a smaller pot of customer spending. Lower prices cannot replace the income those workers lost.

• The study focuses on restoring demand: Fornaro and Wolf examine monetary easing, employment subsidies and labor-tax cuts as ways to support spending and employment.

How an AI Slump Forms

Left alone, the model above produces what the authors call an AI slump: high unemployment alongside rising productivity.

• Workers take a double hit: The car factory now produces 70 cars instead of 100, so fewer workers are needed. AI automation means a smaller share of the income from each car goes to wages, while more flows to profits and the owners of the machines.

• Capital owners receive some protection: The factory sells fewer cars, but owners receive a larger share of the remaining income. 

Workers lose both jobs and wage income, while profits provide owners with a partial cushion.

This is the paradox of productivity playing out within the authors’ model. Enough companies automate to cut costs, spending drops across the board, and even the companies that automated can end up selling less than they did before. None of it is a forecast. Fornaro and Wolf are explicit: a slump is not inevitable, just one plausible branch, contingent on how policy responds.

The Other Path: An Investment-Led Boom

Run the same model with one change: the central bank actively adjusts interest rates, whichever direction is needed, to keep the economy at full employment, instead of holding rates fixed. The outcome flips from a slump into a boom.

Businesses drive that boom early on, investing in the machines and software needed to use AI. Workers don't feel it right away. Automation still shifts a chunk of income away from wages, and in the paper's own numbers, that early income hit is sharp enough that workers' spending drops at first, even as the economy grows. But as that investment builds up the economy over time, workers' spending eventually recovers and rises above where it started.

Take the car factory again. AI raises how many cars it's capable of building. Left to itself, weak demand means the factory only builds a fraction of that. But when the central bank supports spending, businesses keep investing, expect more customers, and production climbs back toward that higher ceiling.

The technology is identical in both scenarios. What changes is the policy response. In the paper's model, this shift alone is enough to turn a possible AI slump into an investment-led boom. Worth noting: Fornaro and Wolf's model doesn't extend to quantitative easing or balance sheet expansion, it only covers interest-rate policy and, separately, fiscal tools like employment subsidies. Any read-through to "money printing" and Bitcoin is a separate question this research doesn't directly address.

Two Headaches for Central Banks

Holding full employment through an automation wave is not simple. Fornaro and Wolf identify two policy problems. The first is a short-run rise in inflation.

• Automation weakens demand for workers: When machines can perform more tasks, companies need fewer people at the existing wage. Keeping the same number of people employed requires the cost of labor to fall relative to the cost of using machines. In the model, this happens through a decline in real wages, wages after accounting for inflation.

• Paychecks rarely get smaller: A company may avoid cutting a worker's salary from $50,000 to $45,000. But if the salary stays at $50,000 while prices rise, that paycheck buys less. The number stays the same, but the worker becomes cheaper to employ after inflation.

• Prices rise instead of salaries falling: The central bank supports spending so companies can keep more people employed. Salaries stay the same, but businesses charge higher prices. Workers keep the same paycheck, though the money buys less. The economy protects more jobs, but inflation initially rises above target.

This effect fades over time. As automation raises productivity further out, that gain starts pushing prices back down, and the early inflation spike gives way to inflation closer to target. 

The central bank still faces a trade-off along the way: supporting employment can mean accepting temporarily higher inflation, while tightening policy to control inflation can instead weaken employment.

The Second Problem: Company Spending Runs Out of Steam

At first, AI creates a burst of business spending as companies buy software, machines, data centers and other infrastructure. This investment adds demand to the economy and can increase pressure on prices.

But companies do not continue building at the same pace forever. Once the first wave of investment slows, business spending falls back while workers still have less income to spend. The economy then moves from an early inflation problem to a later shortage of demand.

The timing here matters. In the authors’ model, the central bank does not need to cut interest rates during the early investment boom because business spending is already supporting demand. Lower rates only become necessary later, once the investment rush fades and weak household spending starts dragging on the economy.

• Early on, rates do not need to fall: business investment is doing the work of supporting demand, so the central bank has no reason to cut. Rates can even need to rise slightly during this phase, since investment itself is running hot.

• Later, the central bank steps in: once the investment wave slows and household spending is still weak, the central bank lowers borrowing costs. Cheaper loans are meant to encourage families to buy homes or cars, and persuade businesses to invest again, replacing some of the spending that disappeared.

• Eventually, rate cuts may not be enough: Fornaro and Wolf show that the interest rate needed to restore demand can fall below zero. Central banks find it difficult to reduce policy rates that far, leaving them with less room to support the economy.

• The result can be a liquidity trap: borrowing is already about as cheap as it can get, but households and businesses still don't spend enough. Production and employment stay weak, with the central bank left with fewer tools to help.

Can AI Raise Inflation?

Yes, especially during the early transition. AI can also reduce inflation later. The outcome depends on whether worker displacement or productivity gains have the stronger effect.

• The productivity effect pushes inflation down: AI helps workers and machines produce more efficiently, reducing the cost of each product.

• The displacement effect pushes inflation up: AI allows machines to perform tasks previously done by workers. If the central bank supports enough spending to maintain full employment, workers must remain affordable relative to machines. Since paycheck amounts resist falling, rising prices can reduce wages after inflation instead.

What Fiscal Policy Adds

Monetary policy is not the only lever. Fornaro and Wolf point to employment subsidies or labor-tax cuts as a useful complement: they lower firms' labor costs and boost worker income, holding down inflation while supporting demand.

What Would You Give Up to Feel Safe From AI?

The economic debate can feel distant from everyday life. A Swiss survey of nearly 6,000 adults asked people to choose between imaginary career options for a son or daughter at age 40. The jobs offered different salaries and different odds of being replaced by automation. No real jobs or money were involved.

• People placed a value on safer careers: The respondents’ hypothetical choices suggested that reducing the risk of an occupation being replaced by automation within the next ten years by 10 percentage points was worth roughly 17% of median annual earnings. No real jobs or salaries were exchanged.

• People worried by different amounts: How much people would give up in pay to feel safer from automation depended on who they were. Men, younger people, and those with more education were generally willing to give up less for the same peace of mind.

The results don't show what workers will actually do in real life. They show that people already take the threat of automation seriously enough to put a price on avoiding it.

A Warning From the Last Computing Wave

The AI boom is not the first technology wave to make expensive equipment become outdated quickly. Cambridge economist Seda Basihos studied what happened during the computing boom that began in the 1990s.

Companies kept swapping out computers and software long before the old ones actually broke, because newer versions did more. At first, that helped businesses get more done. Over time, however, companies had to replace equipment more frequently while workers needed time to develop the skills required to use it. Productivity growth slowed down, and workers ended up with a smaller slice of the economic pie.

The study is not about generative AI. Its lesson was simpler: faster technology does not guarantee permanently faster growth. An early boom can fade when equipment becomes outdated quickly and workers need time to catch up.

The table below lists the economic scenarios that come from Fornaro and Wolf’s model. The possible Bitcoin effects are a separate interpretation informed by Cipolaro’s analysis, not findings from the academic study.

Bitcoin's Two Conditional Paths

This is where the macroeconomic argument connects directly to Bitcoin.

NYDIG's Greg Cipolaro lays out a similar fork in a research note, applied to price. That part is his own view, not a conclusion of the papers above.

• Widespread job losses could eventually help Bitcoin: If AI replaces many workers, households earn and spend less, weakening the economy. Governments may respond with additional spending, while central banks may cut interest rates to make borrowing cheaper. Cipolaro argues that the extra money and easier financial conditions have sometimes helped Bitcoin’s price rise. The initial job losses may hurt Bitcoin first if investors sell risky assets during the economic shock. Any support from lower rates or additional liquidity would arrive later, and even then, a Bitcoin recovery would not be guaranteed.

• The productivity boom could work against Bitcoin: If AI helps the economy grow without causing widespread job losses, interest rates may remain higher for a while. Fornaro and Wolf’s model shows why. At the beginning of an investment-led AI boom, businesses spend heavily on software, machines and infrastructure, adding demand to the economy and pushing the interest rate needed to keep it balanced higher. Cipolaro adds that if unemployment remains low, central banks have less reason to cut rates. Suppose a government bond pays 5% while inflation is 3%. An investor still earns 2% after accounting for rising prices, while Bitcoin pays no interest. Higher returns from bonds and similar investments can therefore make Bitcoin less attractive, particularly to large investors. In Cipolaro’s view, a strong AI driven economic boom could therefore become a headwind for Bitcoin’s price.

Neither path guarantees a Bitcoin price move. Lower rates do not always push Bitcoin higher, and higher rates do not always push it lower. Cipolaro presents them as patterns seen in the past, not rules the market must follow.

Cipolaro also points to steam power, electricity and computing. Each technology caused fears that jobs would disappear permanently, but the economy eventually found new uses for workers and produced more. He uses that history to challenge the assumption that AI must destroy demand forever. He does not claim that the same outcome is certain this time.

Readers who want access to Coinjuice research and trade setups can review the available membership options on the pricing page.

Where the AI Bet Leaves Bitcoin

AI is not automatically good or bad for Bitcoin. Its effect depends on which economic path develops. A rough transition could see jobs disappear faster than new ones arrive, eventually pushing the Fed to cut interest rates and support spending. A smoother transition could allow business investment and productivity to carry the economy, keeping returns from savings and bonds higher.

Bitcoin enters this debate because of its fixed supply and the fact that it pays no interest. Cipolaro argues that Bitcoin may become more attractive when interest rates fall and competing returns decline. Higher rates can create a headwind by giving investors more ways to earn income elsewhere, although they do not prevent Bitcoin from rising. Whether either relationship remains dependable in practice is a contested question that this series will revisit.

Neither path proves that policymakers made the right or wrong choice. The outcome depends on how quickly AI spreads, which workers it displaces, how income is distributed and when policymakers respond.

The indicators worth watching are therefore not AI headlines alone. Labor-market data, business investment and central-bank decisions will reveal which path is developing. The headlines will continue either way. Bitcoin’s AI bet ultimately rides on the policy response.

For readers who want to prepare for changing interest rates and market liquidity, Coinjuice’s ebook on trading without leverage explains how to build and manage positions without the added pressure of leveraged debt.