Fed financing AI boom: who's behind the $3 trillion surge?

Fed financing AI boom is the quiet question lurking behind the loud inflation debate. Federal Reserve Chair Kevin Warsh and others have flagged a real timing problem: AI could lift productivity and productive capacity, but only after the current investment surge strains resources. That tension is genuine. But it may be obscuring something more urgent.

The AI buildout is fueling a sprawling, fast-moving financing system. Its risks, exposures, and weak points are not well mapped. But Morgan Stanley projects nearly $3 trillion in global AI-related infrastructure investment through 2028, with an estimated $1.5 trillion external financing gap that's already absorbing construction, semiconductors, electricity, and skilled labor. In the near term, that pushes up resource utilization and prices. Over time, automation and new capital should raise potential output and cut unit costs, so we can't afford to misread the signals. The error would be treating every pressure sign from this buildout as an inflation problem demanding higher rates. That's a mistake.

Monetary policy shapes future supply

Monetary policy does not merely restrain demand. It can also influence the investment and innovation that determine future supply. Patrick Moran and Albert Queralto showed in a 2018 Journal of Monetary Economics paper that when innovation and technology adoption are endogenous, monetary policy changes firms' incentives to develop and implement new technologies. That means today's rate decisions can directly affect tomorrow's productivity.

The 1990s offer the more compelling historical counterfactual. By the mid-1990s, unemployment had fallen below what policymakers then regarded as its natural rate. Pressure was building inside the Fed to tighten. Chairman Alan Greenspan instead entertained the possibility that the models were wrong, that faster productivity growth had raised the economy's speed limit, and he largely resisted further rate increases. Unemployment continued to fall while inflation stayed subdued.

What the 1990s teach us

How much of that productivity boom would America have missed if the Fed had kept tightening until the economy conformed to its models? We can never know. That is precisely the problem. Inflation caused by excessive accommodation eventually appears in the data. Productivity lost because investment and innovation never occurred does not.

With AI, the damage could be permanent. Data centers, power capacity, human capital, financing expertise, and the businesses that form around them create cumulative advantages. If that investment lands elsewhere, lowering U.S. rates several years later does not necessarily bring it back. Missing a large slice of the AI investment cycle could cause lasting harm to American productivity and competitiveness.

The financial architecture blind spot

Focusing on inflation misses the other half of the AI boom. It's the rapidly changing financial architecture. That's the real story. Traditional monetary-policy models give financial variables very little independent weight, since standard frameworks focus on inflation and employment or the output gap, and financial conditions matter largely only insofar as they forecast those core variables. But that's a blind spot. And it's a costly one.

Recent work with Sergey Sarkisyan shows that credit spreads contain policy-relevant information about financing distortions and firms' cost of capital that inflation and the output gap miss. That's a striking gap. But it reflects a peculiar drift in the Fed's mandate, where the original reason for the institution's existence,protecting financial stability after recurrent banking panics,has quietly become secondary, while inflation and employment now dominate its models and policy debate. So the Fed's founding purpose has faded. It's not gone, but it's no longer central.

Financial stability should rank alongside price stability and employment in the Fed's mission, and sometimes it should outrank both.

High levels of debt, unstable funding, and a severe misallocation of capital can cause far more lasting economic damage than small misses on inflation or employment targets. AI makes this particularly consequential. The Fed needs a much clearer picture of how this investment is financed: the growing role of private markets, the increasingly complicated ties between borrowers and intermediaries, and where debt, maturity risk, and true exposures actually sit.

Better data, better models

It needs better data and better models of how losses could propagate if expected revenues disappoint or today's expensive computing capital becomes obsolete faster than anticipated. That calls for a different allocation of intellectual resources. Since 2008, the Fed has invested heavily in understanding banks, housing, and mortgages. That expertise remains valuable. But the next financial vulnerability is unlikely to resemble the last one.

Private markets and new funding structures deserve comparable analytical depth. A central bank exceptionally well equipped to understand the last crisis isn't necessarily well equipped to anticipate the next one. That's the lesson from 2008. The central failure wasn't simply an incorrectly set federal-funds rate. And policymakers missed the debt, complexity, and interconnections of the rapidly changing mortgage-finance system until the fallout became systemic, which means they didn't see the whole picture until it was already tearing through the economy. So we've got to ask harder questions.

Not subprime, but the lesson holds

AI isn't subprime mortgages. Predicting another financial crisis would be unwarranted. But the institutional lesson is clear, and it's this: when financial innovation moves faster than our models, understanding where risk is accumulating must be a central concern of the Fed, because higher interest rates are no substitute for grasping the actual problem. So don't look for a crash. Just watch where the risk piles up.

In fact, reflexive tightening could bring the worst of both outcomes. If the looming danger is financial rather than inflationary, higher interest rates could expose debt we do not fully grasp while also raising the cost of the productive investment needed for AI to deliver its expected benefits. The result could be a financial vulnerability the Fed failed to understand combined with something much harder to repair: a lasting loss of U.S. technological leadership as investment, expertise, and complementary infrastructure develop elsewhere.

No argument for easy money

None of this argues for easy money. Persistent inflation will demand a policy response, and central bankers shouldn't be picking which AI projects deserve funding, so the real work lies in matching instruments to problems while restoring financial stability to its proper place in the Fed's framework. That means acting without unnecessarily impairing the capital formation on which America's long-run competitiveness may depend. It's a delicate balance. But the task is clear.

The Fed spent the past few years relearning the dangers of underestimating inflation, and the challenge now is not to allow complexity and financial innovation to leave policymakers blindsided. Inflation eventually announces itself. But financial vulnerabilities can remain hidden until they become crises, and missing productivity is harder still to detect, potentially permanent. So the risk we shouldn't underestimate is eroding America's AI advantage before its full productivity gains arrive. It's a quiet threat.