Professional money managers often complain about what they call the amateur’s advantage in investing. Meaning, the pros have a powerful disincentive to call a bubble—or to be bearish in general—because their livelihoods depend on staying invested while everyone else is making money. A fund manager who misses the top of a market can lose clients faster than one who rides a bubble all the way down. In other words, the amateurs—those of us without quarterly reports to explain—are freer to act when things start to look shaky.
I’ve been thinking about that lately because I recently sold a piece of property and now have to decide where to put the money.
Over the years, I’ve relied on a few simple rules to guide my investing:
When in doubt, buy index funds.
Only invest in individual stocks you truly understand.
Avoid speculative bets.
Keep enough cash to survive a one-year fiscal catastrophe.
These principles have served me well through several cycles. But lately, I’ve started to question whether they’re sufficient for what feels like an unusually unstable moment. A few trends, in particular, have me rethinking my assumptions:
The risk of an AI bubble
Chaotic and short-sighted U.S. fiscal and economic policy
Venture capital and private equity hubris
Broader cyclical forces that may be turning.
Let’s take these one at a time. I am not an investment advisor, and nothing that follows should be taken as investment advice. Follow my path at your own risk.
Are We in a Bubble?
I recently interviewed Cal Newport, who offered what might be called the bearish optimist’s case for AI. He argued that even at its apex, the total AI market may be closer to $50 billion than the trillions investors are pricing in. Newport was responding to AI evangelists like Gavin Baker of Atreides Management, who has claimed that entrepreneurs aren’t merely chasing “tens or hundreds of trillions of dollars in value,” but are “in a race to create a Digital God.”
Consider the math. AI companies are spending staggering sums on capital expenditures—Morgan Stanley projects nearly $3 trillion by 2028—yet Americans currently spend just over $10 billion a year on AI services. A recent MIT study found that 95 percent of organizations are seeing zero return from their investments in generative AI. And that’s before a flood of new entrants—many cheaper, faster, and Chinese—crashes into an already glutted market. AI could very well become a commodity.
You can believe AI is a bubble while still being bullish on the underlying technology. As investor and bubble historian Jeremy Grantham has pointed out, “The more serious a new technology is, the more guaranteed you are to have a bubble.” History is littered with such moments—periods when transformative ideas outran the patience of capital. The railway mania of the 1840s bankrupted half of Britain’s rail companies even as it laid the tracks for a century of growth. The dot-com boom of the late 1990s destroyed trillions in paper wealth but ultimately gave us Amazon, Google, and the modern internet. Remember, even Amazon lost more than 90 percent of its value in that crash—and that was the success story. Countless others—Pets.com, Webvan, eToys—simply vaporized.
What makes this AI bubble potentially more dangerous is that the underlying “infrastructure,” if we can even call it that, is fleeting. A GPU lasts maybe 3-4 years while we’re still running freight trains on rail lines laid in the 1800s. Jim Chanos, the prominent short seller, has accused Meta of overstating profits by stretching depreciation schedules for its rapidly obsolescing AI hardware. He claims Meta’s depreciable base at mid-2025 was being written off over 11–12 years—a timeline wildly at odds with the actual lifespan of a GPU. In other words, the accounting assumes permanence where none exists.
One could argue that GPUs can simply be replaced, but the problem is that they are the expense. Some estimates put GPUs at 60 percent of total data-center costs. If those investments don’t pay off quickly, companies will have to eat the losses. And unlike the great infrastructure booms of the past, this one creates astonishingly few jobs. Railroads and power grids employed armies of engineers, welders, and conductors — paving the way for entire new cities and towns. The data-center economy employs a handful of coders and a warehouse full of machines that don’t vote, pay taxes, or shop at the local diner.
The bubble argument is, in many ways, independent of the question of whether AI is truly revolutionary. In fact, we may be more vulnerable if the technology is a game-changer. Imagine a double shock: first, the bubble bursts, wiping out trillions in market value; then, the long-term employment crisis kicks in as AI eliminates millions of jobs. A plausible and catastrophic scenario is that we get the crash first, and the job losses second—a slow-motion quicksand that drags growth down just as we think we’re recovering. The parallel isn’t Silicon Valley 2000; it’s Japan in the 1990s: decades of stagnation following a speculative high.
The danger is that these AI stocks are not a corner of the market; they are the market. Since the release of ChatGPT in 2022, the total U.S. market cap has risen by $21 trillion. Just ten firms, most of them AI-heavy—Nvidia, Microsoft, Meta, Google, and their peers—account for 55 percent of that gain. When the foundation of the rally rests on so few pillars, even a modest crack can bring the whole structure down.
VC/PE Hubris
Derek Thompson recently pointed out something comically alarming. Thinking Machines, an AI startup helmed by former OpenAI executive Mira Murati, just raised the largest seed round in history: $2 billion at a $10 billion valuation. The company hasn’t released a product and has refused to tell investors what it’s even building. “It was the most absurd pitch meeting,” one investor told The Atlantic. “She was like, ‘So we’re doing an AI company with the best AI people, but we can’t answer any questions.’”
Now, Murati could very well succeed—just as some subprime borrowers in 2007 made their mortgage payments. But when you pool together too many bets with no standards, rigor, or transparency, you’re courting disaster. This is what hubris looks like in slow motion: an industry convinced it’s reinvented capitalism, racing to fund the next frontier while forgetting every lesson of the last collapse.
What makes this moment particularly dangerous isn’t just the size of the checks—it’s the suspension of disbelief behind them. Venture capital is supposed to price risk, not ignore it. But right now, too much money is being deployed on vibes, pedigree, and fear of missing out. When investors start bragging not about diligence but about access, you know you’ve entered the part of the cycle where the animal spirits aren’t just howling—they’re stampeding.
This money doesn’t just sit on Sand Hill Road anymore. The same funds pouring billions into opaque AI bets are financed, in part, by the retirement accounts of teachers, firefighters, and state employees. Pension systems, desperate for yield in a low-return world, have piled into private markets—venture, growth, and private equity—because that’s where the big numbers have been. But in chasing those headline returns, they’ve also absorbed the opacity, illiquidity, and fantasy accounting that come with them.
That’s where the systemic risk lies. In the short term, inflated valuations and creative accounting make everyone look smart. Yet when liquidity dries up—when IPO windows close and exits vanish—the marks come due. Pensions are left holding assets they can’t sell, with paper gains that quietly evaporate. And when that happens, it’s not just the overconfident VCs who pay the price; it’s the public budgets that backstop those funds.
The trouble isn’t confined to venture capital’s absurd valuations. In private equity, the very plumbing of the system—the way money goes in and supposedly comes out—is breaking down. The Economist captured this problem in a recent piece, “What It Means to Be Illiquid.” Since 2023, private-equity funds have returned only 3.3% of the value of their investments each quarter, far below the long-term average of 5.6%. Venture capital has been even worse. To generate liquidity, funds have resorted to increasingly desperate tactics: selling assets to themselves through continuation funds, taking out net-asset-value loans against their portfolios, and bundling unsellable stakes into collateralized fund obligations—a private-market cousin of the mortgage-backed securities that helped tank the global economy in 2008.
These maneuvers don’t create real liquidity; they create the illusion of it. As The Economist put it, “only naïve or delusional institutional investors see these developments as anything other than signs of distress.” That distress will eventually reach the people whose savings fuel the system—pensioners, teachers, and civil servants who have no idea their retirements are riding on the hope that a secondary-market buyer shows up before the music stops.
And the timing couldn’t be worse. As Trump threatens to slash federal funding for universities—many of which are already trying to unload their private-equity stakes—and targets state budgets with spending cuts, those same state pension funds could soon be forced to do the same. If that happens, we could see a fire sale of illiquid, overstated assets: universities, endowments, and pension systems all rushing for the same narrow exit at once. In that scenario, the unwind could look eerily familiar—a slow-motion replay of the subprime crisis, complete with the same toxic blend of leverage, opacity, and misplaced confidence.
Erratic Government
The federal government is shut down as I write this, but markets barely flinch. Traders, like sleepwalkers in a burning house, have learned to ignore the smell of smoke.
The deeper problem isn’t this shutdown or the next. It’s that Washington itself—not Wall Street or Silicon Valley—has become the biggest source of economic volatility. The nation’s fiscal and trade policies now move with the erratic pulse of one self-interested, unstable, and uninformed man.
Take tariffs. Once a surgical tool of industrial policy, they’ve devolved into emotional reflexes. In April 2025, the administration imposed “reciprocal tariffs” on nearly every major trading partner—friends, foes, and neutrals alike—turning trade policy into the emotional reflex of a government offended by arithmetic. Critics noted that the formula behind the policy was so opaque and mechanical that some accused the administration of building it with ChatGPT: when asked how to impose “fair tariffs” to offset trade deficits, large language models often produced the exact same equation. One week it’s a trade deficit that demands punishment; the next it’s a courtroom in Brasília, where a 50 percent tariff was levied on Brazil for proceeding with the Bolsonaro coup trial—a decision rooted more in political theater than economic principle. Tariffs imposed Monday, waived Wednesday, re-imposed Friday: that’s the rhythm now. Allies like Japan and Canada learn their fate on Truth Social, while companies try to build supply chains in a policy wind tunnel.
The irony is that this was sold as a way to “contain China.” Instead, Beijing has emerged stronger, with steadier trade access to several allies and new diplomatic leverage. At the 2025 SCO Summit in Tianjin, Xi Jinping hosted Russia, India, and even North Korea in a tableau of global realignment. Meanwhile, China is using U.S. unpredictability to brand itself as the adult in the room—an anchor of stability in a turbulent system Washington once led.
The results at home are measurable. Americans now face an average tariff rate of 17.9 percent, the highest since 1934, the year of Smoot-Hawley. The Yale Budget Lab estimates the 2025 tariffs will raise consumer prices by 1.7 percent—about $2,400 per household—and nudge unemployment up by nearly a full point. It’s as if we’ve imposed economic sanctions on ourselves
Meanwhile, the deficit—already swollen by unfunded tax cuts and the “Big Beautiful Bill”—is projected to exceed $2 trillion this year, with total debt near $37 trillion. Federal borrowing now costs more than the entire defense budget. As investor Larry McDonald has warned, the ownership structure of that debt is changing in dangerous ways: the Federal Reserve already holds $8.5 trillion, and foreign demand for Treasuries is evaporating (the consequence of telling the rest of the world to go fuck themselves). The likely outcome is a quiet form of financial repression—artificially suppressing rates below inflation to monetize the debt, effectively taxing savers to subsidize Washington.
That process has already begun. The Fed has now cut rates to 4.00–4.25 percent, with two more cuts expected before year’s end, even as inflation remains above target. It’s an act of desperation disguised as calibration: an attempt to feed a bond market that can no longer support itself. As McDonald puts it, “the beast in the market still needs feeding.”
If inflation settles into a higher long-term range, the entire market logic of the past two decades collapses. The deflationary world that made tech and passive investing (more on this later) feel like free money gives way to a landscape where tangible assets rule: copper, energy, uranium. The financial fantasy of the software era ends where arithmetic and scarcity reassert themselves.
The consequences extend beyond markets. Every extra point of interest costs the Treasury roughly $300 billion per year—money that could have gone to infrastructure, education, or healthcare. Soon, the U.S. government may find itself working mostly to pay its creditors.
Some in the administration whisper that the “cleanest” fix is to let the dollar fall, inflating away the real burden. But that’s a mirage: it makes exports slightly cheaper and everything else—from iPhones to insulin—more expensive. A nurse in Sioux City won’t celebrate a better export balance when Christmas gifts cost twice as much.
Then there’s the data. The firing of the head of the Bureau of Labor Statistics might sound bureaucratic, but it’s existential. This is the agency that measures jobs, wages, and inflation—the dashboard for a $28 trillion economy. Corrupt that data, and you corrupt trust itself. Turkey tried this: replace technocrats with loyalists, massage the numbers, weaponize the press. The result was capital flight, a collapsing currency, and the slow destruction of the middle class.
America is not yet Turkey. But we’re testing the limits of credibility. Once investors begin to doubt the honesty of our data or the coherence of our policy, risk premiums rise, capital retreats, and the long bull market—already obese with leverage—meets the floor of reality. Confidence, like oxygen, is invisible until it’s gone.
And while this unfolds, the U.S. has never been more alienated from its allies. Europe, Japan, and South Korea—the partners who once underwrote our credibility—are now hedging their bets, cutting their own trade deals, and strengthening ties with one another, and with China. When the next crisis hits, we may discover the global safety net that rescued us in 2008 no longer exists.
Cyclical Pressures
There’s something strange about the run of the Magnificent Seven. Their stock prices, at the moment, seem to have detached from the basic laws of investment. The fundamental relationship between what a company earns and what investors are willing to pay for those earnings—the price-to-earnings ratio—has broken orbit. Historically, U.S. stocks trade around 18 times earnings. Nvidia now trades near 53 times. Across the market, valuations have become stretched to levels that make the dot-com bubble look almost quaint. As Apollo’s chief economist Torsten Sløk recently noted, the top ten companies in the S&P 500 are now more overvalued than the top ten companies at the peak of 2000.
In classical economics, markets are supposed to reflect fundamentals: profits, productivity, and the economy’s real capacity to grow. A company’s share price represents a claim on its future earnings, not a mood or a meme. Over time, prices and fundamentals are meant to converge—this is the doctrine of reversion to the mean. When they don’t, bubbles form. And bubbles, by definition, burst.
Yet many investors now argue that fundamentals no longer matter. They insist that we’ve entered a new paradigm—an era where technology’s exponential potential justifies any valuation. But to believe that, you’d have to believe that the most reliable law in financial history—reversion to fundamentals—has somehow been repealed.
History suggests otherwise. Each great market cycle tells the same story with a different cast and setting. In 1929, U.S. equities traded around 21 times earnings, a record at the time; in real terms, the market didn’t recover until 1958. After the 1972 oil shock, it took until the mid-1990s to reclaim prior highs. The dot-com crash of 2000 erased half a decade’s progress, with indices not fully recovering until 2011. In other words, we’ve spent roughly half of the last century just clawing our way back to old highs. Even Japan—whose 1989 bubble made ours look quaint—only recently, three decades later, surpassed its pre-crash peak in inflation-adjusted terms.
These are not mere statistical curiosities. They’re reminders that bubbles don’t just correct prices—they devour time. Every cycle of excess steals a generation’s worth of compounding.
The question, then, is whether AI has truly rescued us from that cycle—or merely delayed the reckoning. The 2022 pullback now looks like a false start, interrupted by the euphoria around generative AI. As one investor put it, “the market has been dragged kicking and screaming” by a handful of tech giants. These firms are now so dominant that they’ve levitated the entire index—masking weakness everywhere else.
Even the skeptics have adapted their rationalizations. As Steve Eisman (of The Big Short fame) recently told CNBC: “What broke the internet bubble was not valuation. What broke the internet bubble was a recession that caused these companies to go bankrupt and do badly. Until there’s something really bad happening… valuation itself is not something I really pay much attention to.” Eisman is right in one sense—valuation alone never kills a bubble. It’s the catalyst that does: a shock, a slowdown, a trade war, a liquidity crunch. One story is all it takes.
Those catalysts always come. Humans are hopelessly pro-cyclical. We extrapolate today’s good news indefinitely. We assume the present is permanent, that what has been working will keep working, that the Magnificent Seven will keep printing profits forever. But markets are coincident indicators, not clairvoyant ones. They reflect the world as it looks in the rearview mirror, not the one around the corner.
And around that corner, history waits—with the same quiet arithmetic it always brings.
Why I Could Be Wrong
“This time is different” are the four most dangerous words in investing, but there are indeed a few dynamics unique to this cycle. Remember, the Japanese market rallied well past historic P/E ratios — wiping out nearly every bear investor — before settling into its decades-long slump. So, even if I am right about the larger trends, timing it is next to impossible.
So, what could keep the party going longer than pessimists suspect?
1. Corporate Impunity
Antitrust law is on life support. The so-called Magnificent Seven—Apple, Microsoft, Google, Amazon, Meta, Nvidia, and Tesla—face little meaningful oversight as long as they kiss the right rings. Their dominance allows them to crush competitors, manipulate markets, and accumulate power in ways no robber baron of the past could have imagined.
I take this seriously. It’s plausible these firms could remain extremely profitable even through a broader downturn. Their monopolies in critical verticals—cloud, search, operating systems, chips—give them a kind of structural immunity.
But for the first time, these seven giants are all competing in the same arena: artificial intelligence. Each one is racing to dominate the same frontier, using the same scarce resources—data, compute, and talent. History suggests that when monopolists go to war, margins shrink and vulnerability rises. Their impunity may keep them in the game, but AI may slow them all down and eat away at margins.
2. Investor Resilience
Today’s investors are, frankly, hard to scare. “Buy the dip” has become a secular creed. As The Atlantic’s Roge Karma noted, this generation of traders has endured the worst pandemic in a century, the highest inflation in forty years, and the sharpest rate hikes in two decades—and still come out ahead. The scar tissue that once kept markets humble—the trauma of the 1930s or 2008—just isn’t there.
The Wall Street Journal captured this well: younger investors, raised in an era of zero rates and endless rallies, have only ever seen markets go up. Success has bred boldness, and boldness has bred complacency. According to JPMorgan, individual traders now account for about 20 percent of all options activity, even higher than during the 2021 meme-stock frenzy, and nearly a fifth of total market volume—roughly double their share in 2010. They don’t read 10-Ks; they read vibes.
It’ll take more than a few bad earnings reports to break that psychology. The real test will come when liquidity itself runs out—when inflation, unemployment, and demographics finally collide. As older investors retire, liquidate 401ks, and draw down savings while younger ones face rising costs and fewer jobs, people will have to sell to survive. This is why I sort of agree with Eisman. It will take a catalyst. Until then, this new retail stoicism will keep feeding the bull.
3. The Passive Revolution
Thirty years ago, nearly all U.S. mutual-fund money was actively managed. Brokers picked stocks, tried to beat benchmarks, and often failed. Today, roughly half of all fund assets sit in passive vehicles—index funds that buy every stock in the S&P 500 automatically, indifferent to price or fundamentals. Most retirees now hand their savings to Vanguard, Fidelity, or BlackRock, which funnel billions into the market every two weeks as paychecks hit 401(k)s.
This, to me, is the most persuasive reason the bubble could persist. Passive investing has created a perpetual-motion machine for asset prices. When markets fall, these investors don’t sell—many don’t even notice. During the 2020 crash, fewer than 1 percent of Vanguard’s clients sold any equity at all. The automatic inflows act like a ventilator, keeping valuations alive regardless of the patient’s condition.
But even machines need power. As America ages, the current inflow dynamic will reverse. Fewer workers, more retirees. When withdrawals begin to exceed deposits, the “steady bid” that’s propped up markets for a decade will fade. The result won’t be a crash so much as a slow suffocation—a liquidity vacuum where the permanent buyers become forced sellers. That’s when the illusion of stability vanishes, and the great passive era meets gravity.
What Will I Do?
My usual investment strategy is not particularly sexy. It’s heavy on index funds, with about 10–15 percent in individual stocks. I use Wealthfront for the indexes and Public for the individual positions, plus a mix of high-yield savings and treasuries. Over the past few years, the individual stocks have crushed the index funds—mostly because I was overweight in the Magnificent Seven.
At the start of this year, sensing that U.S. valuations were stretched, I began shifting toward international equities. My balance now looks like this: 29 percent foreign developed stocks, 27 percent U.S. stocks, 22 percent emerging markets, 13 percent bonds and treasuries, and 10 percent dividend-growth stocks. The foreign developed shares have outperformed the U.S. side this year, roughly +29.8 percent versus +14.8 percent. I’ve also been gradually rotating my individual holdings from tech-heavy names toward commodities—oil, metals, and agriculture. (This doesn’t include a separate pot I keep in cash and high-yield savings.)
For the first time in my investing life, I’m considering stepping away from index funds. Why? Because I think something truly destabilizing is coming. Maybe in a year, maybe two. Maybe it takes a form we haven’t even mentioned yet—like a wave of student loan defaults that ricochet through consumption and credit markets.
What worries me most isn’t my own portfolio; it’s my parents’ generation. I have (hopefully) decades before retirement. If my holdings drop by half, it’ll hurt, but I’m not relying on that money right now. Boomers, on the other hand, have been planning their final decades around assets that could be far more fragile than they realize. If we enter a downturn even half as long as Japan’s lost decades, many retirees won’t have the time to ride it out.
If I were them, I’d start repositioning now—trading exposure for endurance. That means shifting some equities into a mix of high-yield savings, bonds, commodities, and international stocks. Not abandoning risk, but rebalancing toward resilience.
And even though not retiring, that’s essentially what I’ve already begun to do—quietly, methodically, without sacrificing too much in tax liability. Each month, I’ve been inching further out of U.S. equities and deeper into assets that more match the thesis I outline above.
Why that particular mix of assets? Well, that’s another post.