Billionaire investor Jeremy Grantham: This is the most expensive market in American history
What this covers
Jeremy Grantham, co-founder of GMO, joins a CNBC interview to make the case that US stock valuations have reached their historical peak. The conversation spans market bubbles, the AI boom, and what happens when transformative technologies collide with investor excess. Grantham draws on eight decades of market history and his firm's research into valuation extremes to argue that the current market, by measures most predictive over time, is the most expensive in American history—comparable to the 2000 tech bubble at its worst.
The core of his argument rests on a pattern: all 26 statistically extreme bubbles in the historical record, defined as two-sigma events, eventually collapsed all the way back to their pre-existing trend line. Great transformative inventions—railroads, the internet, AI—reliably trigger over-investment and temporary ruin for most investors, even as the underlying technology reshapes the world. Since 2010, the market's price-to-earnings ratio has averaged over 60% higher than the prior 100 years, anchoring his claim of persistent expensiveness. From current levels, Grantham forecasts a decline closer to 70% rather than 50%, though he acknowledges timing remains deeply uncertain—plausibly anywhere from weeks to two years out.
The discussion also tackles where AI differs from prior manias: data-center chips risk obsolescence within two years as new generations emerge, making today's infrastructure over-investment potentially more perishable than fiber-optic cable or railroad iron. He notes the unusual breadth of disagreement about AI's consequences—ranging from utopian leisure to existential risk—and cites his own 2009 call of an S&P 666 market bottom and a 2000 Nasdaq forecast of 70–75% decline (which fell 82%) as evidence his bubble calls, while sometimes conservative, track directionally.
Grantham argues the current US stock market is the most expensive in American history and, like all great-invention manias (railroads, internet, AI), will over-invest and revert to trend, implying a decline closer to 70% rather than a permanent new valuation regime.
- All 26 two-sigma bubbles in the historical record broke all the way back to their pre-existing trend
- Great transformative inventions are reliably accompanied by over-investment and temporary collapse even as the technology reshapes the world
- Since 2010 the market PE has averaged over 60% higher than the prior 100 years, indicating persistent expensiveness
This asset isn't compiled yet
You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.
Great transformative inventions — railroads, the internet, and now AI — are always accompanied by the belief that they will 'rewrite the rules,' by over-investment, and by a temporary collapse in which most investors lose their shirts, even as the technology genuinely changes the world.
“The great new inventions, railroads are always accompanied by it's going to rewrite the rules... They were always accompanied by over-investment and temporary collapse... they will lose their shirts in AI.”
In the long run investors spend about half the time merely waiting to recover a prior high — after the 1929 peak recovery took until 1954, and after the 1972 peak until 1981-82 — a fact obscured by the recent 16-year uninterrupted rise.
“half the time you're waiting to get back to the old high... After 1929, you have to wait until 1954. After 19 72, you have to wait until uh 80 80 a long time.”
In March 2009 Grantham published a one-pager titled 'Reinvesting When Terrified' calling the market bottom near S&P 666, a bullish call few others were making at the time.
“I posted my only one-pager called reinvesting when terrified.”
Using a precise definition of a bubble as a two-sigma event — statistically expected every 44 years if random but actually occurring every 36 years — all 26 identified historical bubbles that met this definition broke all the way back to their pre-existing trend.
“A nerdy statistical term, two sigma. The kind that would come up every 44 years if it was completely random, and actually comes up every 36 years... there were 26 of them, and 26 broke all the way back to the pre-existing trend.”
Because everyone recognizes AI as a dazzlingly important idea, everyone rushes to invest in it, which draws in enormous capital and produces over-investment — the mechanism by which manias in obviously-important technologies still ruin investors.
“AI is so obviously a dazzlingly important idea. Everybody knows it... we all want to put our money in it... and it sucks in more than you can shake a stick at, and you get over-investment. So, everyone in the end, in those situations, loses their shirt.”
In their 2000 quarterly letters GMO called a 70-75% decline for the Nasdaq, and it ultimately fell 82%, demonstrating their bubble calls have been directionally validated even if conservative.
“we said a 70 75% decline for the Nasdaq in 2000 in our quarterly letters, and it went down 82.”
Amazon fell 92% after the 2000 tech bust — having risen sixfold beforehand — and then went on to 'inherit the earth,' illustrating that even eventual winners can devastate investors during the bubble collapse.
“Amazon in 2000 came down 92%. It had gone up six times, it came down 92% and then inherited the earth.”
By valuation measures most predictive over history — chiefly the value of the stock market relative to GDP with modifications — this is the most expensive market in American history, most closely comparable to the 2000 tech bubble.
“based um on the value of the stock market compared to the GDP with with modifications, uh this is uh the most expensive market in American history.”
Fiber optic cable laid during the 2000 boom ruined investors at the time but was eventually all used, whereas AI-era chips may be redundant within two years as new generations appear — making today's data-center build-out more perishable than prior infrastructure over-investment.
“Think of the fiber optic cable that got laid. It ruined everyone at the time back in 2000. But we used eventually we used all that cable. The trouble is... Chips today may be redundant in 2 years.”
From 2010 until today the market's PE has averaged over 60% higher than it did for the prior 100 years, which does not guarantee a collapse but does clearly indicate the market is expensive.
“From 2010 until today the PE has averaged over 60% higher than it did for the prior 100 years... it does indicate it's expensive.”
Bubbles can overshoot well beyond two sigma — some rose to two-and-a-half sigma and Japan reached almost three sigma — which is the practical problem because a client's patience runs out before the bubble finally breaks.
“some of them went up from two sigma to two and a half sigma. And in the case of Japan, almost three sigma. So, they can be painfully higher than just two sigma. And that's the problem. Because the client's patience is not quite as long.”
The disagreement in opinion about AI's consequences is more profound than Grantham has ever seen in the stock market — Nobel Prize winners, executives, and workers all disagree — with outcomes ranging from utopian leisure to accidental or deliberate human extinction.
“The disagreement in opinion is more profound than I've seen in any other anything in the stock market ever... We're either going to be sitting on the beach getting served mint juleps by machines, or they're going to kill us accidentally or on purpose.”
The market will peak and drop back to trend, and getting back to trend from current levels implies a decline closer to 70% than 50%.
“The market's going to peak out and drop back to trend. And getting back to trend from here is uh closer to a 70% decline than a 50% decline.”
Grantham distinguishes his two explicit 'it's now' super-bubble calls from his broader position that the 21st-century market has generally been overpriced by 20th-century standards.
“two times I clearly said it's now. The other times I said it's overpriced. The 21st century has been overpriced by the standards of the 20th century.”
It has paid shockingly better to be a permabull than a permabear over the last 100 years.
“it has paid shockingly well to be a permabull over the last 100 years more than it has paid to be a permabear.”
The timing of the coming peak is deeply uncertain — plausibly anywhere from two weeks ago to as far out as two years — reflecting the general unpredictability of market tops.
“My guess is sometime between 2 weeks ago, 2 weeks from now, 2 months, 2 quarters, and conceivably 2 years. The timing is always terribly uncertain.”