New York Federal Reserve President John Williams has indicated that constraints on the supply of equipment critical to artificial intelligence infrastructure may endure beyond two years, adding to potential upward pressure on prices.
Speaking in a Bloomberg interview, he observed that the intensive construction of AI-related facilities is elevating costs for semiconductors and the broader range of goods that incorporate them.
Questioned on whether supply and demand in the AI sector might reach better alignment within the coming two years, Williams replied that the adjustment would probably require more time, provided the current strength of demand continues as anticipated—an outcome he described as a substantial assumption.
These remarks accompanied Williams’ restated assessment that the Federal Reserve may still find it appropriate to raise interest rates once more this year in pursuit of lower inflation.
He noted that his view on the possible need for an additional increase has not shifted.
On the outlook for 2027, he adopted a more cautious stance, indicating that further steps would depend on evolving conditions and that he currently sees no compelling reason to expect additional tightening.
Separately, Williams anticipated that investment in AI would continue at a robust pace even amid higher borrowing costs, as those committing capital tend to be relatively insensitive to rate levels; they regard the technology as likely to deliver major gains in productivity and transformative shifts across the economy.
Alongside these near-term economic effects, questions have arisen about longer-term consequences of rapid AI progress.
Leadership at Palantir has pointed to the possibility that sophisticated AI systems could adversely affect humanity, including scenarios in which generative capabilities grow powerful enough to introduce meaningful risks to human well-being or rights—particularly if rival actors develop superior tools first.
Such observations highlight the tension between technological opportunity and the potential for societal disruption.
President Donald Trump has rejected suggestions that AI development should be deliberately slowed.
He has maintained that effective oversight comes primarily through decisive leadership rather than expanded formal constraints, and that the United States must preserve its advantage because supremacy in artificial intelligence confers decisive strategic benefits overall.
Trump has framed certain safety-oriented appeals as potentially overstated or aligned with other priorities that could disadvantage American interests.
By comparison, Elon Musk and OpenAI chief Sam Altman have articulated substantial concerns about the speed of advancement.
Both have supported arguments for moderating the rate at which increasingly capable systems are developed, citing the risk that progress may outpace measures needed to maintain safety and human control.
Any coordinated effort to reduce the pace of AI development, however, would confront steep practical barriers.
It would require sustained international agreement among leading nations, including China and other major participants, together with a high level of reciprocal confidence. In light of prevailing geopolitical rivalries and the strong incentives for each party to press ahead, such cooperation is widely viewed as neither realistic nor feasible to achieve (at least in the foreseeable future).
