UpTrajectory Review

GMO, the Boston-based asset manager co-founded by Jeremy Grantham, has used its quarterly letter to warn that the artificial-intelligence trade is a bubble primed to burst. The headline alone tells us the letter draws a direct line between today's AI enthusiasm and prior manias Grantham has spent a career cataloging, from dot-coms to housing. What the available text does not give us is the letter's specific evidence, valuation math, or timing call, so we cannot responsibly quote or paraphrase its internals. What we can say is that GMO's framework has been consistent for decades: identify parabolic price moves detached from cash flows, flag the narrative that justifies them, and wait. Grantham called the 2000 top and the 2008 housing crash early enough to be both right and mocked; he also missed the post-2009 recovery for years. That track record matters when weighing this call.

For a small-business operator, this is not an abstract market debate. The AI buildout is touching your cost structure whether you participate or not: cloud-compute pricing, the cost of software vendors bolting AI features onto their products, the availability and price of capital for tech-adjacent suppliers, and even local commercial real estate demand near data-center hubs. If GMO is right that a break is coming, the downstream effects include tighter credit for anything AI-branded, a repricing of the semiconductor and energy infrastructure that has been bid up on hyperscaler spending, and a cooling of the labor-market frenzy for AI talent that has pulled salaries up across adjacent roles. If GMO is wrong, the cost of waiting is that competitors who adopted early compound productivity gains you deferred. Either way, the letter is a prompt to audit your own exposure, not a signal to panic.

What is genuinely new here is not the bubble claim itself, which has circulated for over a year, but the source: GMO's quarterly letters are read closely by institutional allocators, and Grantham's name still carries weight with pension and endowment committees. The letter lands at a moment when the market's concentration in a handful of AI-linked megacaps is at historic extremes, which makes the timing of the argument notable even if the substance is familiar. Where we are skeptical is of any precise timing implication readers might draw. Grantham has been early on nearly every major call, and 'coming break' could mean months or years; businesses that restructure around a downturn that arrives in 2027 will have traded two years of growth for a hedge they may not need. The letter's value is as a stress-test for your assumptions, not as a market clock.

The second-order effects split the business community unevenly. If AI-linked capital spending contracts, the pain concentrates in suppliers, commercial landlords near data centers, construction trades, and the regional utilities that have been planning generation additions around hyperscaler demand. It also hits the venture-backed startup ecosystem, which has been funding AI-native tools on the assumption of endless cheap capital; a repricing there would tighten acquisition pipelines and slow the flood of point-solution vendors competing for your software budget. On the other side, businesses that have deferred AI adoption on cost grounds may find the correction brings pricing back to earth, making 2026 or 2027 a better entry point than 2024's froth. The letter also has a portfolio implication for owners whose personal wealth is concentrated in index funds: the same concentration that powered returns is the concentration GMO is warning about.

What to watch next is concrete. First, hyperscaler capex guidance in the next two earnings cycles: if Microsoft, Google, Amazon, and Meta soften their spending language, that is the earliest reliable signal that the buildout is pausing. Second, credit conditions for AI infrastructure borrowers, where spreads have been remarkably tight; widening there precedes equity repricing. Third, whether GMO's letter gets echoed by other institutional voices or dismissed as permabear repetition, which will tell you how crowded the skeptic trade has become. For operators, the actionable step this quarter is an exposure inventory: list every vendor, lender, landlord, and customer whose fortunes are tied to AI spending, and ask what your plan looks like if that spending halts. That exercise costs an afternoon and works whether or not Grantham is right this time.

Takeaway: Audit your business's exposure to AI-linked vendors, customers, and capital now, because GMO's warning is less about timing than about how concentrated the downside has become.