The IMF's chief economist warned this week that AI investment isn't just a productivity story , it's a wealth effect that could keep inflation stickier and interest rates higher than the startup world wants to believe.
Pierre-Olivier Gourinchas didn't bury the lead. Speaking ahead of a Bloomberg report published June 26, the IMF's chief economist said the AI investment boom is generating "tremendous valuations" in US equity markets , and that those valuations are making investors and consumers wealthier, more confident, and more willing to spend. That demand is adding to inflation pressures before the productivity gains from AI have shown up in the data. The added investment and consumption, Gourinchas said, are helping to elevate demand and inflation without associated productivity gains to offset it.
The IMF's April 2026 World Economic Outlook puts global growth at around 3.1 to 3.2 percent for 2026 and 2027. Global headline inflation is projected to tick up to 4.4 percent in 2026 before declining again in 2027. That's not a catastrophe, but it's not the clean disinflationary glide path that rate-cut optimists were counting on either. And AI, Gourinchas is now saying clearly, is part of why.
The comparison Gourinchas reaches for is the late 1990s internet bubble. Both eras pushed stock valuations and capital gains to new highs, which fueled consumption that added to price pressures. The difference, he notes, is that today's AI investment isn't primarily debt-financed, which limits systemic contagion risk. A market correction would hit shareholders, not necessarily the banking system. But that caveat shouldn't distract from the more immediate problem: a wealth-driven spending surge is real inflation pressure, regardless of whether it carries a systemic tail risk.
The Fed is watching the same dynamic. If rising equity prices from AI optimism are making households wealthier and driving consumption, the central bank has less reason to cut rates aggressively , and more reason to hold them where they are. Gourinchas himself noted recently that moving away from strong forward guidance on rate cuts is "entirely appropriate," a phrase that politely translates to: don't count on relief anytime soon.
That matters because the rate environment is the single most consequential external factor for early-stage companies right now. Q1 2026 saw nearly $300 billion in venture funding globally, with AI companies capturing roughly 80 percent of it, according to Crunchbase data. But that concentration tells only part of the story. The record round sizes are flowing to OpenAI, Anthropic, xAI, Waymo , firms that can absorb capital at scale. For the broader startup ecosystem, the number of deals actually declined relative to 2022 levels. Founders outside the AI infrastructure layer are raising into a market where capital is concentrated, expensive, and increasingly skeptical of anything that doesn't have a direct AI claim attached to it.
Here's the thing: the AI boom is simultaneously inflating the portfolios of investors while keeping the cost of capital elevated for the companies those same investors are supposed to be backing. It's a strange loop. A founder raising a Series A today looks at a partner whose paper returns from AI bets look excellent , and still gets told the bar for deployment is high and the timeline for exits is long. The wealth effect benefits asset holders. The rate effect punishes asset builders.
The Magnificent Seven are projected to spend roughly $668 billion on AI capital expenditure in 2026 alone, per Wall Street estimates cited by CNBC, with total AI capex potentially crossing $1 trillion by 2027. That level of investment is itself inflationary through energy costs, chip demand, and the construction of data centers at industrial scale. None of it is demand-destroying in the short run. All of it adds pressure to supply chains that are still working through post-pandemic normalization.
The IMF's two-sided warning is worth holding onto. If AI productivity gains arrive as promised, disinflation follows and the current rate environment starts to look like a temporary tax on growth. If those gains disappoint , if the revenue curves for AI companies bend down before the capital expenditure curves do , the market correction that follows could crater demand fast and leave a lot of companies on the wrong side of a very expensive build-out. Gourinchas flagged both paths without committing to either, which is about as honest as macroeconomic forecasting gets.
For founders and early investors, the practical read is uncomfortable but clear: the same AI momentum making venture portfolios look good on paper is the reason the Fed has cover to sit on rates. If you're building something that requires cheap capital to reach profitability, the timeline you're working from almost certainly needs revisiting. The boom is real. So is the bill.
Also read: Washington handed Anthropic back its most powerful model, and the precedent it set is bigger than the lockdown • The AI trade's week of reckoning arrived and the market's message to founders is clear • Washington's Escalating Hardware War With China Is Forcing a Reckoning for America's AI Buildout