Kevin Warsh’s AI Stock Market Comments, Explained

The comments people are searching for came from the press conference after the Federal Open Market Committee meeting that ended on October 28, 2026. According to reporting by The Motley Fool and other market commentators, Warsh pushed back on a popular argument on Wall Street: that the AI boom is so large and so well funded that interest rates barely affect it.

The line most often quoted is that “interest rates get in the cracks”, meaning higher borrowing costs eventually reach every part of the economy, AI included. Some coverage framed this as a warning to the AI-driven rally; other analysts read it as a general statement about how monetary policy works.

The backdrop explains why markets listened closely. The Fed kept the federal funds target range at 3.75% to 4%, but the vote was 10–2. Governor Lisa Cook and Cleveland Fed President Beth Hammack dissented because they wanted a quarter-point rate increase.

When two voters want to raise rates and the chair says inflation is too high, investors start pricing a chance of a hike rather than a cut. That shift matters most for companies whose valuations depend on profits far in the future, a description that fits much of the AI trade.

For most of 2025 the AI buildout was paid for largely with cash from the biggest tech companies. In 2026 that changed. Meta, Microsoft and Oracle have all raised new equity or debt to fund AI spending, and a growing share of GPU purchases and data center construction is financed with loans, bonds and leases.

That is the link Warsh’s comments point to. When rates rise:

• Financing costs rise for data centers, GPU fleets and the power deals behind them.
• Payback periods stretch, because each year’s cash flow is worth less in today’s money.
• Valuations of long-duration growth stocks tend to fall, since more of their value sits in future years.

Not everyone agrees on the size of the effect. Moody’s chief economist Mark Zandi has said there is some merit to the view that AI demand is not very rate-sensitive, while still warning that higher rates weigh on stocks and raise the cost of capital.

The week after the Fed meeting was rough for AI shares. On Friday, October 30, the Nasdaq fell 2.9% and the Philadelphia Semiconductor Index dropped more than 7%, and Alphabet lost close to $500 billion of market value after its earnings, InvestorPlace reported. Worries about debt-funded AI spending and a less friendly Fed fed into the same move.

On Tuesday, November 3, US election day, the trade bounced. The semiconductor index rose about 3% and Nvidia’s market value went back above $5 trillion. That evening AMD and Super Micro both beat expectations. Searches for Warsh’s comments peaked the same day, as investors tried to work out whether the sell-off or the rebound was the better guide.

We cover the rebound in S&P 500 and the AI trade rebound, and the two earnings reports in AMD Q3 2026 earnings and SMCI stock after earnings.

For a company building AI capacity, interest rates are not an abstract market story. A data center is a large up-front cost paid back over years, so the cost of money changes whether a project works.

Take a simple example: a facility that costs C$50 million and earns steady cash each year. If the financing rate rises by one percentage point, the yearly interest bill rises by about C$500,000 on the full amount borrowed. That can add months to the payback and change which province or state makes sense, because cheaper power can offset dearer money.

That trade-off is what our AI Build Configurator is built for. It shows capital cost by line item, power cost and payback for every Canadian province and US state, so you can test a project before rates move again.