Fed Holds Rates at 3.75%–4%, and Warsh Tells AI Investors Rates Still Bite

The Federal Reserve held its benchmark interest rate steady at 3.75%–4% on Wednesday, and Fed Chair Kevin Warsh used the moment to deliver a blunt warning to AI investors: monetary policy applies to them, too.

The Fed said it needed more economic data to guide its next move, with inflation still elevated while growth and business investment stayed strong. For investors, that leaves a harder question hanging over the market: can higher rates restrain the AI investment boom without derailing the wider expansion?

What did Warsh actually say about AI?

Warsh told a news conference on Wednesday that AI investment is not insulated from monetary policy. He pushed back against what he called a narrative that higher rates would have little effect on AI investment because "the returns that investors are expecting are so different."

He rejected the idea that expectations of unusually high returns make financing costs irrelevant. The stakes are concrete: roughly a quarter-point range in the policy rate shapes borrowing costs across the economy, and AI projects — with their multi-year construction timelines and heavy upfront capital — sit squarely in that path.

"With respect to the new technology and this investment surge, do interest rates matter?" Warsh said. "I believe they do, and I don't believe that they're somehow divorced from the rest of the economy."

He added: "Interest rates get in the cracks. Interest rates affect directly, or indirectly, almost everything."

The takeaway is plain: the Fed's own chair does not accept the argument that AI is a special case.

How would higher rates hit AI projects?

Higher rates may not stop companies from investing in AI. But they can influence which projects attract financing, how much capital they receive, and what returns investors demand.

Warsh pointed to widening credit spreads for some assets as evidence that investors are becoming more selective. In credit markets, that could mean a growing divide between borrowers with credible paths to cash flow and those whose projects depend on more optimistic assumptions about future demand and profitability.

That distinction matters as companies commit substantial capital to AI-related projects, including data centers and computing infrastructure. Even when the long-term opportunity remains attractive, higher financing costs can weaken project economics and make investors more sensitive to execution risks.

In short: marginal AI projects, not the whole sector, face the first squeeze.

Is the boom really all about technology?

Warsh also cautioned against attributing the economy's strength to technology alone. "High-tech investment" is leading the acceleration, he said, but spending is strengthening elsewhere too — a signal that capital spending is broadening beyond technology.

That resilience complicates the Fed's task. Strong growth gives policymakers room to keep rates elevated while they assess inflation. But sustained investment and demand could also make price pressures harder to contain.

The implication cuts both ways: the same economy supporting AI spending is the one keeping the Fed cautious.

What does the hold mean for markets?

For markets, the implications are mixed. Expectations of fewer future rate hikes could support Treasury prices. Yet persistent inflation, strong growth, and large fiscal deficits could limit declines in longer-term yields.

Julius Bendikas, U.S. head of economics and dynamic asset allocation at Marsh, said the Fed's decision to hold reinforced the case for patience.

"In our view, there is no need to rush with labor markets balanced, wage growth subdued and longer-term inflation expectations anchored," Bendikas said. "We expect Treasury yields to have room to fall as expectations for future hikes moderate."

Lower Treasury yields could ease borrowing costs even without a Fed rate cut — a partial offset for capital-heavy AI projects. But Bendikas flagged the limits of that optimism: "strong growth, elevated energy prices and ongoing large fiscal deficits create upside risks."

The Fed has bought time, not clarity; the next round of data will decide whether AI financing costs fall or climb further.