The promise of AI-driven deflation remains a distant dream as mounting infrastructure costs push prices higher across the US economy right now.
Silicon Valley executives including OpenAI CEO Sam Altman and Tesla chief Elon Musk have repeatedly argued that artificial intelligence will make goods and services dramatically cheaper.
Altman recently wrote that “intelligence too cheap to meter is well within grasp,” capturing the optimism that has fuelled trillions in AI investment globally.
SoftBank’s Masayoshi Son went even further, predicting a 40% drop in prices and arguing that “unnecessarily hard work, sweating work, would no longer be needed.”
Those predictions are nowhere near being realised, as corporate adoption of AI remains uneven and slower than boosters promised.
Goldman Sachs Research estimates capital expenditure on the AI buildout will reach $581 billion this year in the United States alone, equivalent to 1.8% of gross domestic product.
That spending is pushing up electricity prices, chip costs, software fees, and data-centre capacity charges well before any meaningful productivity payoff has materialised.
A Census Bureau survey found that only between 17% and 20% of US businesses reported using AI, with adoption far more common at large firms than small ones.
Ronnie Chatterji, chief economist for OpenAI, acknowledged that “it’ll still be a little while before we see it sort of clearly for productivity statistics,” even as adoption gradually increases.
“For it to impact the economy, it has to be adopted by organizations,” Chatterji said, adding that “those organizations have to realize value” before broader gains show up in economic data.
Julie Averill, Lululemon’s former chief information officer, cautioned that the industry’s promises need tempering, saying “the hype is around the ease of the technology in a large organization.”
“The things that have always made implementations in large companies difficult still exist, which is people,” Averill said, noting that building trust in AI models remains a significant internal challenge.
Chatterji said OpenAI’s own data shows power users deploy AI at eight times the rate of average companies, a gap that has grown sharply from two times just three months ago.
Peter Boockvar of One Point BFG Wealth Partners noted that even during the internet boom, the US saw only a 1.5% productivity gain over 30 years, questioning whether generative AI can do substantially better.
“To think that generative AI is going to bring that level of enhancement to the economy, relative to the internet, is tough,” Boockvar said.
The inflationary pressure is now a live debate inside the Federal Reserve, with Fed Chair Kevin Warsh facing pressure from multiple directions as officials weigh interest rate decisions.
Household electricity prices rose 10.1% in the two years leading up to June, outpacing the overall 6.3% increase in consumer prices recorded over the same period according to Bureau of Labor Statistics data.
Minneapolis Fed President Neel Kashkari dissented in favour of a higher rate, stating that “the massive investment in data centers has also added a new demand element to the high inflation Americans are experiencing.”
The cost of computer software and accessories has risen 22.9% since June 2024, while JPMorgan Chase estimates DRAM memory chip prices will have risen 400% by the end of the year compared to 2024.
Warsh appointed Stanford professor Charles Jones, who is on leave at Anthropic, to a Fed task force on AI, alongside venture capitalist Marc Andreessen, who predicts an era of “hyper-deflation.”
Warsh wrote before taking the job that “AI will be a significant disinflationary force, increasing productivity and bolstering American competitiveness,” a view that aligned with President Donald Trump’s push for lower interest rates.
The Fed voted in July to leave interest rates unchanged at between 3.5% and 3.75%, though the decision was not unanimous as some officials pushed to restrain AI-driven price increases.
“The cost and inflationary aspect is really complicating Kevin Warsh’s job,” Boockvar said, arguing that future productivity gains are simply not something the Fed can factor into today’s rate decisions.