Citi sees AI price effects as key issue for Warsh By Investing.com

Citi says roughly half of recent 2% to 3% U.S. real GDP growth is being driven by AI-related investment, with AI capex up more than 1.5 percentage points of GDP over the last two years. The note warns growth could fall below 1% or turn negative if that tech buildout slows, while April CPI also showed tech-component price spikes feeding inflation measures. The Fed is expected to remove its implicit easing bias at the June FOMC meeting, with Chair Kevin Warsh favoring tighter inflation filtering and less forward guidance.

Analysis

The market is starting to price AI capex as a macro input rather than a sector theme, which is the key second-order shift. That creates a fragile setup: the same spend that supports headline growth also concentrates growth dependence in a narrow set of technology supply chains, so any pause in hyperscaler capex would hit both GDP and the earnings revisions of semis, networking, and data-center infrastructure at once. The more important implication is that this is a duration-sensitive growth engine: if policy stays hawkish and real rates remain elevated, financing intensity rises just as the marginal return on incremental AI spend likely falls.

A hawkish Fed with less forward guidance is more than a rates story; it raises the probability that equity multiples become the transmission mechanism for tighter financial conditions. That matters because the AI buildout has not yet created a labor-market buffer, so the economy can absorb a tech slowdown with less immediate job loss, but the equity impact could still feed back into consumption through wealth effects. In other words, the downside path is not classic recession first, unemployment later; it is capex slowdown, multiple compression, then slower consumption.

The inflation nuance is important because it weakens the case for an easy “transitory tech disinflation” narrative. If measured inflation is being distorted upward by a few hardware categories, the Fed may overestimate persistence in goods inflation even as underlying demand is cooling elsewhere. That raises the odds of a policy mistake in the next 1-2 meetings: hold too tight into a decelerating growth impulse, or misread the signal and delay easing until the market has already repriced growth downward.

Key Decisions for Investors

• Overweight semis and AI infrastructure suppliers only via a basket with explicit hedges; pair long SOXX vs short IWM for the next 1-3 months to isolate capex concentration from broader domestic demand weakness.
• Reduce exposure to long-duration growth equities that need multiple expansion to work; use QQQ puts or put spreads into the June FOMC as a hedge against a hawkish guidance shift over the next 2-6 weeks.
• Long front-end rates relative to intermediate rates via a 2s10s steepener only on a tactical basis if the Fed removes easing bias but growth data stay firm; if claims or capex roll over, reverse quickly because the trade breaks on recession pricing.
• Watch AMZN as a second-order AI beneficiary/loser: if cloud customers slow incremental spend, AWS growth can decelerate before headline capex does; prefer selling short-dated upside calls against any rally rather than outright long exposure into the next earnings cycle.
• Avoid chasing consumer cyclicals here; if equity wealth effects start fading, retail and discretionary names become the most direct channel for a capex-led slowdown over a 3-6 month horizon.