Everything About AI
NextFin News - The Federal Reserve is "sensitive" to artificial intelligence, and its rate policy is starting to show it. That was the message from Jeffrey Rosenberg of BlackRock after the central bank left interest rates unchanged on Wednesday, a move that connects the Fed's tightening campaign directly to the AI investment boom reshaping the U.S. economy.
The Federal Open Market Committee voted 10-2 to hold the federal funds target in a range of 3.75% to 4%, the level set at the September meeting when policymakers raised rates by a quarter point. Only Governor Lisa Cook and Cleveland Fed President Beth Hammack dissented, each favoring a quarter-point hike. The decision arrived less than a week before the midterm elections, and Chair Kevin Warsh used his press conference to insist that politics played no role in the timing.
Stocks barely reacted, the S&P 500 closing at 7,728, up 0.07% on the day, while the 10-year Treasury yield edged up to 5.02%. The muted response is itself a signal: investors were less focused on the hold they already expected than on what the Fed's new language says about the path ahead. The real story is not the pause. It is the transmission mechanism of monetary policy changing underneath the Fed's feet.
The Fed's New Problem: AI Makes Monetary Policy Uneven
Rosenberg, senior portfolio manager for BlackRock's systematic alternatives ETF, said in a television interview that the Fed recognizes financial conditions have tightened but stops short of calling them tight — and he tied that hesitation directly to artificial intelligence. The distinction matters because it exposes the central bank's deepest problem: the same policy stance transmits through the economy unevenly when one sector is building data centers with near-rate-insensitive capital while households and homebuilders face the full force of a 4% policy rate.
Warsh himself gave the clearest signal in his opening statement. "While overall financial conditions are tighter, I'm unwilling and unable to call them tight," he said. The phrasing is deliberate. A central banker who declares conditions tight invites the market to price an easier path ahead; one who refuses to say so keeps the door open for more restraint. But the reason conditions feel only "tighter" rather than "tight" is that the AI buildout keeps absorbing capital regardless of the cost of borrowing.
"While overall financial conditions are tighter, I'm unwilling and unable to call them tight."
The evidence is in the national accounts. Warsh said GDP growth in the first half of the year was revised up from 1.8% to 2.4%, and that third-quarter consumer spending is "likely running above 3% growth." Business investment, he added, is "likely rising at about a double-digit pace," with capital expenditures at a historically high share of GDP and high-tech investment leading the acceleration. That is the footprint of an AI-driven capex cycle running hot enough to offset weakness elsewhere.
There is a historical analog, and it explains why the Fed is so alert. In the late 1990s, technology investment surged while the rest of the economy cooled, and the Fed faced the same asymmetric transmission problem: rate-sensitive sectors slowed, but the tech capex boom kept aggregate demand firm. The difference now is scale and financing. The 1990s boom was largely equity-funded; today's AI buildout is being financed heavily through corporate debt, which means higher rates feed back into credit spreads and issuance volumes rather than just equity valuations. That changes the feedback loop the Fed has to manage.
A Structural Shift, Not a Cyclical Blip
The question investors should be asking is whether the AI investment surge is a cyclical fluctuation that will mean-revert or a structural shift that will not. The evidence points to structural. This is not a one-off inventory restocking or a temporary fiscal sugar rush. It is a change in the economy's capital intensity driven by a technology that requires power, chips, and data centers on a scale that compounds year after year.
BlackRock's own research, published in late August, framed the AI buildout as a test of "financing and cash-flow durability" rather than a broad beta trade — a sign the market is moving from speculation to infrastructure. The firm's chief investment office noted that hyperscaler capital spending has been climbing at a pace that offsets weakness in housing and other rate-sensitive sectors, and that AI-related productivity is already weighing on hiring in technical and professional fields. Twelve-month average payroll growth is just 26,000, with healthcare accounting for virtually all net hiring and employment outside healthcare falling by 45,000 in a single month.
That divergence is the mechanism. AI raises productivity and profits per employee while reducing the need for incremental labor in exposed industries. It lifts aggregate growth while hollowing out the middle of the labor market. And because the capital spending is funded partly through debt markets that are pricing the supercycle as durable, the demand for funding stays high even as the Fed tries to slow the economy. Monetary policy is fighting a demand shock that its own rate tool cannot fully reach.
Torsten Slok of Apollo Global Management put the point bluntly in October: AI's insensitivity to interest rates is a problem for the Fed. He argued the boom is triggering a form of "Dutch disease" in the U.S. economy, draining capital, power, and labor from vulnerable sectors such as housing and autos, and that the only durable fix is expanding the supply of power, chips, and infrastructure. Monetary policy cannot resolve that sectoral imbalance; it can only keep rates higher for longer.
For a cyclical call to hold, you would need three things: a short-term driver, a demonstrated mean-reversion pattern, and at least three comparable historical cycles. The AI capex cycle has none of them. There is no inventory drawdown that refills, no policy impulse that expires, no prior episode in which an economy built out a general-purpose technology and then reverted to its old capital intensity. The productivity task force Warsh appointed is itself an acknowledgment that the Fed no longer trusts its old models to measure what is happening.
The Second-Order Effect: Higher for Longer, With More Volatility
The first-order read of Wednesday's decision is simple: the Fed is pausing, but the bias is toward more hikes. Sixteen of 18 policymakers projected in September that at least one more increase would come this year, and the median projection for the policy rate at the end of 2026 sits at 4.1%. The second-order effect is more important and less widely priced: if AI keeps capital demand high while productivity lifts potential growth, the neutral rate itself rises. That means the policy rate can sit at 4% without feeling restrictive, and the Fed loses the clean signal it once got from the gap between policy and neutral.
This is where Rosenberg's "sensitive to AI" framing becomes a warning rather than a description. A central bank that is sensitive to AI is a central bank whose reaction function is changing. It will tolerate tighter conditions in rate-sensitive sectors because it cannot cool the AI sector without crushing the rest of the economy. The result is a policy path that is higher for longer, more data-dependent, and less guided by the old single-metric framework. BlackRock's fixed-income team has argued the Fed is moving away from treating core personal consumption expenditures as the sole measure of price stability, and Warsh confirmed the shift toward flexibility when he announced that from next fall the FOMC will hold eight meetings a year with at least one dedicated exclusively to strategy.
The bond market has already started to price this. The 10-year Treasury yield was 4.99% as of October 27, with the 2-year at 4.70%, leaving the curve only modestly inverted. Traders were pricing a more than 90% likelihood that the Fed would stand pat in October, but the focus has shifted to the December meeting and beyond. If the AI capex cycle proves durable, the term premium on long-duration debt stays elevated, and the era of cheap long-term funding for everyone except the hyperscalers continues.
There is a third-order implication most investors have not priced: a higher neutral rate revalues every long-duration asset, not just bonds. Growth stocks, commercial real estate, and leveraged buyouts all depend on the assumption that the cost of capital will eventually fall back to the 2010s norm. If the neutral rate has structurally risen, that assumption is the fragile one. The AI winners can fund themselves; everyone else faces a permanently steeper hill.
The Counter-Thesis: Maybe AI Is Deflationary After All
The strongest argument against this read is the one Warsh himself has made. At the IMF-World Bank meetings in Bangkok on October 16, he said he is optimistic that AI will make economies that use it well more productive and wealthier, while warning that "the path from where we are to there is likely to be a very bumpy one." The bull case for lower rates rests on the supply side: if AI delivers a positive supply shock large enough, it can lower unit labor costs and inflation even while demand stays strong. Productivity has already been moving higher, Warsh noted, though he cautioned that "most of that change in productivity precedes everything we're talking about superintelligence."
"I want you to hear from me optimism, but optimism leavened by the seriousness of purpose. A decade from now, people will look back and say this was hugely consequential."
That is a serious argument, and it is the reason the Fed is pausing rather than hiking aggressively. But it requires the supply shock to arrive before the demand shock becomes embedded in wages and inflation expectations. Inflation has run above the Fed's 2% target for close to six years, and September's consumer price index was 3.7% year over year. Core CPI at 2.6% is closer to target, but the remaining pressure sits in categories less responsive to rate policy, such as medical care and shelter. The window for a clean disinflation is narrowing, not widening.
The counter-thesis also depends on the labor market absorbing displaced workers into new roles fast enough to keep aggregate wages from accelerating. So far, the data point the other way: employment outside healthcare is contracting even as aggregate growth firms. That is the signature of a transition cost, not a smooth supply shock. A central bank watching both sides of its mandate cannot ignore it.
The falsifying signal for the "higher for longer" view is specific: if core CPI prints at or below 0.2% month over month for three consecutive months while the unemployment rate rises above 4.5%, the structural-inflation thesis breaks and the Fed's next move is a cut, not a hike. Until then, the burden of proof sits with the doves.
What Comes Next
The short-term path is set by the December 8-9 meeting, where a quarter-point hike remains the base case. The medium-term path depends on whether the AI capex cycle shows signs of rolling over — watch hyperscaler capital expenditure guidance and the share of capital expenditures in GDP, both of which Warsh flagged as trends "bearing watching." The long-term path is structural: if the neutral rate has genuinely risen, the 2% inflation target may be reachable only with a policy rate that feels restrictive to a much larger share of the economy than it did in the 2010s.
For investors, the asymmetry is clear. Rate-sensitive sectors — housing, autos, small caps — remain exposed to a Fed that cannot ease until the AI sector cools. Technology and the AI supply chain benefit from durable capital demand but face rising financing costs as debt markets reprice. Fixed income offers more selectivity and a broader use of starting yield and carry, as BlackRock's fixed-income team has argued.
Warsh ended his press conference by saying the Fed's job is to "get monetary policy right," and that the information coming in "will help us determine where we want policy to be." The information that matters most is no longer just the monthly inflation print. It is whether the AI revolution proves to be a temporary demand spike or a permanent raise in the economy's capital intensity. The Fed is sensitive to AI because it has to be — the technology is rewriting the transmission mechanism of monetary policy, and the central bank is still learning the new language.
Explore more exclusive insights at nextfin.ai.
