Originally published on The Diary of A CEO on June 25, 2026
Guest: Jeremy Grantham, co-founder of GMO, 60 years in investing, managed up to $165 billion
Host: Steven Bartlett
Intro
The investor who called the 2000 and 2008 tops sits down to explain why AI being genuinely world-changing is exactly what makes it a bubble, what he’d own instead of US stocks, and why he now spends as much time on falling sperm counts as on markets.
The Bubble Always Forms Around the Real Thing
Grantham’s first move is to reject the idea that bubbles are built on scams: “People think that a bubble is a mainly because it’s a scam and nothing could be further from the truth. The great bubbles always occur around the very most important ideas.”
He puts AI “right up there with the railroads” as one of the defining ideas of the last couple hundred years, and says that is the precondition, not the counterargument
Railroads did change the world, and railroad stocks collapsed and everybody lost a ton of money
Amazon went up six or seven times in 1999, fell 92% in the crash, and then “out of the wreckage, it inherited the retail world”
“The greater the idea, the more obvious the idea, the more money goes in and the bigger the bubble and the bigger the bust.”
On what he thinks he’s actually good at: spotting where humans are short-term and predisposed to optimism, looking for good news and avoiding unpleasantness
He credits the economist Kenneth Boulding with the line he keeps coming back to: “The only people who think you can have compound growth on a finite planet are madmen and economists”
On market efficiency, he calls the discounted-future-earnings model “complete nonsense” and offers the replacement: “The stock price is what you think the other guy will pay.”
How Far the High-Flyers Fall
“This is I think the biggest investment bubble in American history.” Asked whether a collapse is coming over the next few years, he says the data “would be compatible with history for the peak to be very soon”
On what a break does to the most-loved names: “A 70% decline would not be unexpected” for the stocks that have moved the most
His precedents, in his own numbers:
The NASDAQ fell 82% in the tech bust — “It is far from unprecedented to have these major declines”
Japan in 1989 reached 65 times earnings and sold for more than the entire US; the US hit 35 in 2000 and is “35 or 40 today, but it’s not 65”
Japan then fell for 20 years, and took 35 years to recover
The Nifty Fifty peaked in 1972 and declined 65% adjusted for inflation, with a recession he calls just about the worst since the Depression
1929 was down “about 80% or more” and was followed by the Great Depression
His firm’s 10-year forecast in 2000 was minus 2% a year for US equities; the actual outcome was minus 3%, and he believes today’s market is priced higher than 2000 was
On what an ordinary person feels: the high-flyers lay people off, everyone feels less rich, spending falls by a few percent of the paper losses, and the economy goes under stress
He is explicit that he cannot time it. Pressed on when, he stretches the answer from the next few days out to the next few years, and will not narrow it further
Why Your Advisor Will Never Tell You to Get Out
The section’s core claim, stated flatly: you will not receive the advice from investment advisers “to get your tail out of the market ever”
The 1999 story he tells to prove it: at the annual analysts’ society debate, he asked 400 self-identified full-time stock market experts two questions
If the market went from 31 times earnings back to a more normal 17 at any point in the next 10 years, would that guarantee a major bear market? All 400 said yes
Did they think it would happen? Fewer than 1% said no
The same firms’ public representatives were on the podium telling him to calm down — “It was a huge betrayal of trust if you wanted to put it that way.”
Asked whether that is happening now, he answers “Of course,” and says the warnings are not merely quiet: “you will hear nothing. You never have. You never will. It is simply lousy business for a big firm”
He says he sympathizes, because being early cost him: when his firm made the call in 1999, “we lost half our book of business in two and a quarter years”
The career-risk mechanic underneath it all: in a bear market everyone freezes and waits, then fires one or two laggards; in a bull market a manager watching a peer make money fires you instantly
He traces the whole thing to Keynes: “the central political skill in life turns out to be never be wrong on your own”
Seven Monopolies Walk Into One Ring
Looking backward, each of the megacap seven owned a near-monopoly on a global basis — Tesla’s jump start in EVs, Apple in smartphones, Microsoft in PC software, Meta in social networking, Google in search, Nvidia in chips
Looking forward, he says, “you could not imagine a more different world” — all of them are now girding for battle in the same marketplace
The capex escalation is now a chest-beating contest, with one player’s $200 billion in a single year set against another’s $105 billion, funded out of enormous cash flows and increasingly out of borrowing on top
“Now they have no monopoly. There are seven potentially sharp elbowed, ruthless players determined to fight out with each other until they win.”
“There’ll only be one survivor. They think everything goes to the one who gets there first.”
Asked who wins, he says he doesn’t know, and suggests the smart play may be to opt out: one or two of them might reasonably decide to stay in hardware, license someone else’s model, and stay out of a fight he expects to be brutal
SpaceX as the Euphoria Marker
He treats SpaceX as the cleanest read on where sentiment is: it defines a quarter of global GDP as its addressable market and talks about endless opportunities mining asteroids
“In 50 years people, in a 100 years people will look back and tell stories about SpaceX and its prospectus like they tell stories about the South Sea bubble” — an enterprise of great advantage, but nobody to know what it is
“SpaceX is such a fabulous BS story.” On the promises in the prospectus, he says he thinks it “will fail to deliver anything like its promises in the prospectus”
He says 90% of the company’s theoretical value is AI, even though its own model appears to be losing to two or three others
“It’s the classic description of a market peak. It’s what you look for at the top of a terrific bubble.”
His reading of how Musk actually built Tesla, which he says can’t repeat at SpaceX’s scale: talk the stock to four or five times paper value, sell a slug, build a gigafactory, keep talking, watch it re-rate, sell another slug, repeat — a self-fulfilling prophecy that required a bull market he won’t get twice
On Mars specifically: “Going to Mars is not within the laws of physics” as a survivable proposition — you would need to live underground against cosmic rays and build a spinning gravity machine, and “we have not been able to build a sustainable system in a dome ever”
What He’d Actually Own Right Now
“rule number one is always be diversified” — which for him means wholesome bonds, wholesome cash, and perhaps a small amount of precious metals
His actual allocation when asked directly what an ordinary person with $1,000 or $10,000 should do:
Roughly 60% in a broad-based index of non-US equities
5–10% in precious metals, with no preference between gold and silver
A bit of real estate if convenient and sensible
The rest in bonds
On US stocks, his answer is one word long: “Don’t own US stocks.” Asked about the S&P 500 specifically, he says no
His reason is price, not politics: “I’m not confident that US equities will be intact in 5 years 10 years.” Asked why, the whole answer is “Because they’re so badly overpriced today.”
The rotation is already visible in his numbers: emerging markets up 65% over the last 12 months against 25% for the S&P, which he notes has done better than he would have guessed
On why everyone tells you the opposite: “Of course they do. They’ve been completely dominant for 20 years.” He says markets extrapolate today’s conditions and double-count in the worst way — crushed earnings times seven in 1982, peak margins times 35 in 2000
Housing Is a Bad Bet on Demographics
A typical UK house sold for 3.4 times family income in 1994, about as low as it had been in 50 years; it now runs to over 10 times depending on location, with the same pattern in China, Canada, Australia and most of Europe
“And at 10 times income, a reasonable young couple are in big trouble” — and the same high prices feed straight through to rents
Even a 30% decline would only take housing back to six or seven times family income, still twice the level of what he calls the good old days
His argument against buying property as the default first investment is demographic, not cyclical: family formation is already declining across the richer countries, and expensive houses plus fewer new households is not a combination that supports prices
Asked whether scarcity of affordability pushes prices up, he cuts it off: “if people can’t afford it, there’s no one bidding”
Crypto Does One Thing Well
He owns none, has never owned any, and says he never will advise anyone to buy it
“It’s an unnecessary piece of nonsense” that in his account facilitates nothing except moving money out of sight
His case against it as an asset: not a store of value because it moves too much, not usable as a medium of exchange because you can’t spend it in a shop
“It does one thing very very well. It’s a means of speculating beautifully.”
On Bitcoin going to zero: “in the distant future, yes, it will certainly go to zero, but it may take a long time” — adding that in the distant future everything goes to zero
Advice for Founders: Raise Now, Buy Later
Presented with a founder who plans to raise as much as possible now specifically so he can buy up distressed competitors after a crash, Grantham’s response is two words: “Good lad. Good advice.”
His general counsel to founders dependent on investor capital is to lock up money if they can and build conservatism in elsewhere
On timing, he refuses to pretend: the horizon is weeks, months or years, and he won’t narrow it
The framing he offers is that the sun is shining and it is time to act as if a storm is coming — a decent principle in any period, and a better one than usual now
Nobody Agrees About AI, Including the Nobel Winners
What he finds most striking about AI is the absence of consensus at every level — Nobel laureates, corporate executives and the people inside the companies all disagree violently
“There is absolutely no agreement on whether AI is going to make us all so rich we can sit on the beach and never do another day’s work or it will wipe us out accidentally or on purpose because it’s a much higher level intelligence one day.”
“And when was there ever a case where a higher intelligence was benevolent in a sustainable way to a lower intelligence?” He credits Geoffrey Hinton with the one counterexample anyone offers: mothers and babies
On the safety case, he thinks building in benevolence is not impossible but would require slowing down, and that “you should make sure you can do that before you push ahead”
The paperclip problem gets his attention because of literalism, not malice: an open-ended instruction, a machine with the means to execute it, and no stopping condition
On the commercial obstacle to safety, raised by Bartlett describing a model that refused to alter his own data: any model that becomes judgmental loses users to a competitor, so the market pressure runs toward removing the restrictions rather than adding them
On the framing of social media as well-meaning people who couldn’t spot the consequences, he pushes back on the premise: “I question basically the well-meaning bit. They’re now trying to maximize their profits and their growth and their appeal over the competition”
On energy and robotics, he takes an unusual position for an investor: he’d rather the space-based data center thesis fail, because success means robots everywhere and energy demand “massive beyond belief,” and he wants the extra time for humans to work it out
The Inequality Reset
He puts the US Gini ratio “up there with Brazil and Mexico” — countries that used to be the punchline
His dividing line is 1975. From 1935 to 1975 the US grew over 3.5% a year, with the poorest quarter gaining slightly more than average and the richest quarter slightly less; since 1975 the average hour worked has barely gained anything after inflation
The show’s own research put harder numbers alongside his account — the top 1% holding 31% of national wealth against 2.5% for the bottom half, and the top 10 billionaires’ wealth up 526% between 2020 and 2025 — figures he did not dispute but did not himself cite
His concern isn’t the bottom 20%, it’s the median: “the guy in the middle, the 50th percentile, he is unhappy also. And when your average guy is unhappy because he’s not doing very well, you know you have a problem”
Asked what history says happens next, his answer is one word: “All bad.” He points to the Gilded Age, and says the US got lucky in an ugly way — World War I, the Depression, then World War II — and came out an unusually equal society
On the politics, he reads the move to Trump as smaller than the average of the last seven European elections and driven by the same impulse: kick the rascals out, left or right, because things don’t feel like they’re going well
“We need to tax the rich and help the poor. It’s pretty simple.” He wants a return to something between the 1950s–60s structure and today, phased in over decades so the bottom quarter gains half a point a year on the average and the top gives up half a point
The Sperm Count Data He Can’t Stop Looking At
The path in was environmental, not medical: his foundation started on climate 27 years ago, moved to the collapse in insects — “insects appear to have dropped in biomass, the weight of the flying insects by 50 to 75%” over 60 or 70 years — and then to whether the same effects show up in humans
He credits the Shanna Swan and Hagai Levine work showing sperm count had roughly halved since the first academic reports in 1970, and says his own team found the decline rate accelerating, now running about 2.5% a year
The numbers he uses, in units per milliliter: roughly 180 million in hunter-gatherer days, about 100 million when the academics arrived in 1970, and 35 million today
He puts the threshold for conceiving without difficulty at about 45 million, a level crossed 15 to 20 years ago
The share of young couples needing help has gone from essentially nil to about 17%
“in 20 to 25 years the average young couple will need help getting pregnant” — a conclusion he says he and Swan reached independently
The mechanism he points to is endocrine disruptors leaching from plastics, and he notes that fertility specialists writing books on the subject don’t mention toxicity at all, only the hundred solid reasons people choose to have fewer children
The two Harvard and Mass General studies he thinks deserve more attention, both self-reported diet studies at a fertility clinic
Among 180 men, the quarter eating the least pesticide-heavy produce had roughly double the sperm count of the worst quarter
A parallel study in women showed 68% live births in the best quartile against 38% in the worst — again close to double
His explanation of why pesticides land so hard: they’re designed to kill insects, weeds and fungi, they’re impregnated into the structure of the fruit rather than sitting on the surface, and the fetus is 100 to 1,000 times more vulnerable than an adult
“The baby bust is measurable. The sperm count is one of the few things you can really measure.” On why it gets no airtime: “we just don’t do bad news”
Toxicity Is Regional, Which Is the Good News
The reason he considers detoxification tractable is jurisdictional: unlike climate, a single country can act alone and its own people get the benefit
His cosmetics example: about 10,000 chemicals in use, of which the EU has banned 1,500, Canada 550, and the US 12 — “I am not kidding you”
He thinks the top of the intervention list is narrow and cheap: no cosmetics during pregnancy, and organic for the dirty-dozen produce — berries, apples, pears, peaches, spinach — with bananas, oranges and melons fine either way. He estimates that alone removes as much as half the problem
The generational point is what raises the stakes: a woman’s eggs are all formed in the womb, so exposure carries to grandchildren, and he says recent work suggests more generations than the two they could originally prove
The measurable consequence he keeps returning to: the life expectancy gap between the US and Sweden has gone from two years to six over the last 70, and he has written that his estate would bet on eight or ten in 50 years
The show laid out the regulatory gap in detail — 85 pesticides permitted in the US and banned in the EU, China and Brazil, 1,300-plus cosmetics chemicals restricted in the EU against 11 by the FDA, PFAS in at least 45% of US tap water — and his only amendment was that the EU is “forever giving exemptions and extensions and is far from perfect,” just much less bad
“we have to detoxify capitalism” is the harder half of his prescription: banning chemicals is intellectually easy, but turning capitalist norms family-friendly takes generations
“we all need clean air, clean water, fertile soil, and 2.1 healthy, well educated children. Without any of those, society fails.” He frames all four as commons, and notes that no country has yet produced a permanent uptick in birth rates despite spending several percent of GDP trying
Where to Live, and What the Ambulance Time Tells You
Asked whether there are countries he wouldn’t live in, he declines: “I think I have to refuse to answer this on the grounds that it might tend to incriminate me” — then answers anyway when pressed on the US
His reason is the social contract, not the economy. He says the US holds out too much chance that the social contract is dissolving, and contrasts it with Japan, where what really upsets people is being put in a position where they can’t act in a socially responsible way
“I think the case here is that people are doing what they think is best for them and their family and screw everybody else.” He contrasts it with corporations when he arrived in America, which built the stadium and acted like part of the city they operated in
His single best proxy for a functioning society is maternal mortality, and his numbers run: Nigeria around 480 per 100,000; the US Black population 44; the US overall about 20 or 21; the US Asian population 13; Britain 5; Germany 4; Sweden 2.1; Norway zero in each of the last two years
“What better definition of civilization than looking after the mothers giving birth?”
His framing of the US result is that it isn’t merely worst in the rich world, it’s 50% worse than the next worst
The everyday indicator he watches for fraying: UK ambulance response times going from 12.5 minutes to an hour and a half
Asked where to go instead, he names Denmark, Japan, France and Germany, and says the UK has a little of the American disease but not nearly as much
On the counterargument that America is rich: the wealth is real but concentrated, “and that dazzles in terms of the average” — measured by how well off the bottom quartile is, he says America doesn’t score well at all
What He’d Do at 33, and What He’d Tell Everyone Else
Asked what he’d do at 33 with no contacts and no accumulated knowledge, his answer is not defensive: “I think the simple appeal would be to get your tail into AI” and try to be a leader, knowing more about the area than the next person
The rest of that advice is take a lot of risk, don’t be conservative, work hard, and think outside the box. He identifies the common failure as feeling constrained to play by the regular rules and assuming experts and authorities know what they’re doing
What he’d tell his own children is different, and it’s climate and practical skills — one of his sons is learning to farm chickens, pigs and mushrooms, because he thinks there’s a good chance the complexity of civilization starts to unravel at the edges first
For the taxi driver, the receptionist and the nurse, his advice is to expect tougher times than they were led to expect, and to act on it:
Build a reserve of cash
Get a job that pulls its weight in a larger sense
Upskill into something mechanical — fixing, repairing, engineering — or into research science
Make friends, and live somewhere with a tight society
Asked what he’d do if he could not fail, he says he’d write the toxicity and social contract book — the one that does what Rachel Carson’s Silent Spring did — and that he’d start it tomorrow
He closes on the same point he opened with, that nobody is coming to warn you: “Look at the data. A bubble is not hard to see” … “Don’t wait for help because no help is coming.”
Grantham’s position is not that AI is fake but that its realness is what guarantees the overinvestment, and the same refusal to look at bad data that keeps advisors quiet about a 70% drawdown is, in his telling, what keeps a measurable collapse in fertility out of the conversation entirely.