AI Has Completely Broken Western Monetary Policy. What Happens Next?

The other day, on October 26, entrepreneur Elon Musk wrote on social media that access to debt financing had been all but cut off worldwide for companies that generate most of their revenue from operating data centers built specifically for artificial intelligence (AI).

I've been analyzing how this situation has been unfolding for a long time, so I'm not surprised. I've watched the global flows of capital into AI gradually dry up. What's more, I'm convinced that US financial authorities have been making extraordinary efforts to keep the American financial system afloat.

Here's the problem. The US Federal Reserve cut interest rates six times over two years. Then, all of a sudden, it raised them in September. Why? Under the conventional logic of monetary policy, when the Fed raises rates, investors should be more willing to buy US debt because yields go up while the risk remains minimal.

But in practice, this monetary policy mechanism has broken down. Capital kept flowing into the AI sector until, just the other day, someone finally pulled the plug on debt financing. On average, each of the giants of the US AI industry is taking a bite out of 4% of the money that would otherwise flow into US Treasury bonds. That's a huge amount.

The Fed's September rate hike, meanwhile, turned out to be a wasted move. The White House then shut off access to debt financing by reaching agreements with the US lending sector. The reason is simple: the US government needs to refinance its debt, and that is becoming increasingly difficult. Capital has to be stopped from flowing into AI.

Tighter Monetary Policy in the US and Japan Won't Tame Inflation Fueled by the US Conflict with Iran

But that's only part of the problem. US financial authorities are also grappling with an unusual feature of the country's rising consumer prices. In a conventional model, a central bank sees prices rising persistently, tightens monetary policy and waits for demand to cool. The catch is that higher interest rates do little to bring down the price of a barrel of oil, the cost of insuring ships or shipping costs, and they do nothing to ease fuel shortages. Rate hikes hit mortgages, consumption and investment, but they don't address the root cause of the price shock.

Nor will raising rates increase oil production in the Persian Gulf or reopen the Strait of Hormuz, which has been closed because of the US conflict with Iran. What it does do, consistently, is make borrowing more expensive for everyone who depends on credit. The rate on a 30-year US mortgage has already climbed above 7% following the Fed's September decision, and the same move is feeding through into the cost of servicing the national debt.

US inflation stood at 3.4% in August by the headline Consumer Price Index and 2.4% by the core index. By September, those figures had risen to 3.7% and 2.6%, respectively. The Fed has made money more expensive than the rate at which consumer prices are rising, yet inflation keeps accelerating exponentially!

Europe and Japan are facing much the same situation, with the same result: mortgage and government debt costs go up, while oil-driven price increases keep showing up on price tags.

Who pays for tightening that doesn't fix the problem? Consumers, developers and government budgets. Who benefits? Owners of computing capacity and energy assets, because what they have to offer remains scarce whatever the interest rate. High rates can cool the mortgage market, but they can't stop corporations that, until very recently, regarded AI as a strategic investment and were willing to borrow at high interest rates to fund it.

Now, though, the money has dried up for them too, and the situation is extremely difficult. US regulators have understood this correctly. A Fed rate hike, the inability to attract enough money into US government debt, and inflation fueled by the US conflict with Iran: that's a combination US elites are hardly happy about.

That's why Washington has urgently asked Moscow to supply diesel fuel. Tankers carrying Russian fuel are critically important: nearly one in every eight barrels of diesel previously exported to the global market came from Russia. The US urgently needs to stabilize the pace of price increases, even as Washington scrambles to figure out how to correct the financial distortion caused by the AI hype.

In the US, Only Data Centers and Power Plants Are Keeping Construction in the Black. Everything Else Is Deep in the Red

The shift of demand away from the US government debt market and toward borrowing by AI companies is ultimately distorting the structure of the American economy as well.

US construction data for 2026 show how these incentives are playing out across the American economy. Residential construction is down nearly 5%. Offices, commercial real estate as a whole and many other segments are also in decline. The only sectors firmly in positive territory are data centers and the energy facilities directly associated with them. Corporate spending on data centers has reached a record annualized rate of about $75 billion, up more than 50% year over year.

Take server farms out of the picture, and nonresidential construction is either flat or declining. Energy construction was up 11% through August 2025 because every data center is effectively joining a queue for gigawatts of power. Companies are responding by building their own power plants or signing long-term contracts. Everything else, from housing and conventional offices to manufacturing, is deep in the red.

AI is pulling construction capacity, labor, materials and capital toward itself, creating imbalances across industries. What's more, by my estimates, US GDP isn't growing at all once you strip out the gains from data center construction and energy facilities built to serve AI.

I always tell my colleagues that an economy in which only one industry is growing isn't a growth economy. It's a bubble economy resting on a single pillar. The Fed itself is beginning to sense this. In September, Jeffrey R. Schmid, president of the Federal Reserve Bank of Kansas City, publicly asked whether the AI ecosystem was becoming "too big to fail." When a regulator raises rates to cool this very sector while simultaneously calling it a potential systemic risk, that's an admission that the conventional monetary policy tool of interest rates is no longer hitting its target.

AI-Powered Investor Decisions Outpace Any Monetary Policy's Ability to Adapt. The Way Forward: Central Bank Digital Currencies

Monetary policy needs a radical overhaul for the AI era. Conventional monetary policy, as we know, operates with a lag. It takes time to have an effect: at least three months for the first impact, and in reality, it can take more than six months before the results of a policy change become apparent.

The AI market works differently. Algorithms detect changes in yields, electricity prices, logistics costs, exchange rates and news almost instantly. AI-powered decisions adapt faster than any monetary policy in the world. That doesn't mean algorithms are always right, but they unquestionably change the speed at which markets move.

The defining feature of today's global situation is that conventional monetary policy is no longer keeping pace with the challenges of the AI era. The biggest of those challenges is the high likelihood of a sharp sell-off in the stocks of major US companies that have bet virtually everything on this technology.

In other words, neural networks aren't going anywhere, but the hype around them will fade. Having spent many years in senior management at Otkritie Bank and Otkritie Broker, I've seen colleagues in the banking sector get carried away with one technology after another, only to become disillusioned later.

A balanced approach is clearly the way to go here. Don't rush to pour huge sums into an innovation without first working through every possible scenario for the returns it might generate. The absence of that kind of discipline is dragging down the businesses of hundreds of US companies that are burning billions of dollars on paid neural network subscriptions but can't earn back what they've invested. At the same time, innovation shouldn't simply be dismissed!

Take blockchain, for example. It's fallen off the radar for many investors, and that's a mistake. I still believe corporate neural networks should be trained on a distributed ledger. It's the only way to keep them under control. And AI itself must not be pushed to the sidelines when markets start shaking as the AI bubble deflates. Countries should start rolling out central bank digital currencies and AI-powered smart contracts now to make monetary policy more responsive.

Smart contracts determine how money is issued and the terms under which it circulates, allowing liquidity to respond to market signals in seconds. Traditional central bank open-market operations can't do that; they take weeks. Against this backdrop, Russia's decision to launch the digital ruble for widespread use on September 1 looks timely. I'm convinced that we can and should continue expanding the capabilities of this third form of the national currency.

Let me remind you that on October 20, the State Duma passed the first reading of a bill designed to establish the legal framework for foreign exchange transactions involving digital rubles and, accordingly, for currency controls. The digital ruble and the Bank of Russia's smart contracts are already a signal that monetary policy itself is beginning to modernize. Given global experience, that's the right move.

Western monetary policy hasn't broken down because Warsh or Lagarde made a mistake. It has broken down because the pace of economic change has become orders of magnitude faster than regulators can keep up with. Tightening doesn't cure inflation when it comes from the oil market, but it does keep hitting mortgages and government budgets.

The next stage in the evolution of global finance will be competition between the monetary policies of different economic systems, built around digital currencies and AI-powered smart contracts. The first to understand this will gain a monetary system that responds to the world as fast as the world itself changes. The rest will be left playing catch-up, and there's no guarantee they'll ever close the gap.