Mary Daly, president of the Federal Reserve Bank of San Francisco, told Axios that artificial intelligence demand could extend energy and supply shocks, keeping inflation elevated longer than the Federal Reserve typically expects. The concern is not merely a temporary spike, but a structural shift where AI, tariffs, and higher energy costs compound to require further monetary tightening.
Daly noted that the Fed usually looks through supply shocks that fade within one to three years. However, she observes that AI-driven pressure on chip and other technology prices appears to be moving beyond that window. "I see it less as a one-off," she said, adding that demand for AI does not seem to be declining. Instead, it appears to be increasing, suggesting relief is further out than standard models predict.
The San Francisco Fed district includes Silicon Valley and many companies driving the current AI boom. Daly reports hearing signs from business contacts that firms are bracing for tighter chip supplies. This trend is altering purchasing and product-design decisions outside the data center sector. Some companies are seeking forward contracts for memory chips to secure suppliers, a notable development in a market where chips typically depreciate quickly and firms rarely lock in supply ahead of time. Others are "reengineer[ing] their products" to rely less on specific chips, creating more flexibility if supplies tighten. Daly described these moves as signals of concern that the shortage could spread more broadly.
The risk lies in AI hardware demand competing with chips used in cars, appliances, and other goods. This dynamic could recreate bottlenecks similar to those seen after the pandemic, when chip shortages prevented manufacturers from finishing vehicles and pushed prices higher. Daly warns that demand for AI-specific equipment could spill into the broader semiconductor market, raising costs for companies with little connection to data centers.
A complicating factor for the Fed is that the companies fueling the AI boom are among the least sensitive to higher interest rates. Daly stated that "these hyperscalers aren't very interest rate-sensitive," though they may become more so as they rely on borrowing to finance their buildout. While the largest AI spending is concentrated among these hyperscalers, other companies investing in the technology are more sensitive to borrowing costs. Therefore, higher rates can still restrain the broader economy and inflation outlook, even if they have less impact on the firms at the center of the boom.
Daly affirmed her support for the interest rate hike taken three weeks ago, calling it necessary given increased inflation risks. She declined to specify whether she advocates for further action but explained that her view will depend on whether energy, trade, and AI shocks abate or compound. If these shocks prove to be conventional and temporary, further hikes may not be needed. However, if they last longer than forecast or if a second round of tariff negotiations adds new pressures, the period of elevated inflation could extend.
Daly emphasized that the FOMC will continue to carefully dissect whether these shocks are rolling off or compounding and whether underlying inflation is gaining momentum. This framing differs from some colleagues who argue monetary policy is out of whack with the broader economy and needs adjustment regardless of geopolitical events. Fed Governor Michael Barr recently stated that the Fed was "out of position" and made an adjustment in the right direction. Daly's position implies openness to pausing rate hikes if geopolitical events cooperate in coming months, whereas other officials see broader inflationary forces compelling continued adjustment.