GMO Founder Grantham: AI Highflyers Could Drop 70%, Recommends 60% Allocation to Non-U.S. Stocks
Miles Bennett
GMO co-founder Jeremy Grantham warns US equities are in the biggest bubble in American history, with AI high-flyers facing a potential 70% decline; he recommends allocating 60% of capital to non-US stocks and adding precious metals as a hedge.
Why does he call this the biggest bubble in US history?
Grantham cites the "Buffett Indicator" — total stock-market capitalisation divided by GDP — now at roughly 235%, the highest on record.
This means → by this single metric, today's market is pricier than the 2000 dot-com peak.
He draws a direct parallel to 2000 and warns SpaceX could replay Amazon's 92% post-bubble crash.
How is the AI boom different from past bubbles?
In the railroad and internet booms, Grantham argues, the technology's practical use was already visible; AI's path to commercial payoff remains uncertain.
He points to SpaceX defining its addressable market as one-quarter of global GDP — a claim he calls "a classic sign of a market top."
In plain terms = the closer a bubble gets to its peak, the grander the story — "asteroid mining, massive AI breakthroughs" sound thrilling but remain far from actual profit.
If the bubble bursts, how does it hit the real economy?
Grantham's chain: stock crash → companies lay off workers → household wealth shrinks → consumer spending drops → the broader economy slows.
This means → his concern is not just portfolio losses but a cascading economic downturn triggered by collapsing asset prices.
What does he tell ordinary investors to do?
Put 60% of capital into broad non-US equity indices — his reasoning: "they are much cheaper and have significantly outperformed US stocks since early last year."
Allocate 5%–10% to gold and silver as a hedge.
A small real-estate position is acceptable, but he notes property is "quite expensive by historical standards"; the remainder goes into bonds.
How reliable is his track record?
Grantham correctly called the 1989 Japan crash, the 2000 and 2007 US market tops — a genuine record.
Yet since 2021, he has published near-annual "epic crash" warnings, none of which has materialised.
In plain terms = he is a veteran who has nailed major turning points, but has been crying wolf in recent years. Whether this is the real top signal or another premature alarm depends on whether AI can turn its story into real revenue before valuation pressure becomes unbearable.
Content is for reference only, not financial advice.