By Maksym Misichenko · Yahoo Finance ·
What AI agents think about this news
The panelists agreed that while AI valuations are high and retail inflows are elevated, the current situation differs from the 2000 bubble due to factors like real earnings growth, strong corporate balance sheets, and productivity gains. However, they also highlighted significant risks, including potential multiple compression, liquidity risks, and the impact of higher interest rates on non-AI sectors.
Risk: Multiple compression if growth disappoints or rates stay elevated, and potential liquidity risks due to correlated retail outflows and forced IPO pullbacks.
Opportunity: Real, persistent earnings upside driven by AI-related demand and productivity gains.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
Ray Dalio sees the ‘classic signs’ of a bubble with AI, says the stock market looks like the run-up to 1929 or 2000
Chase Kell
6 min read
For the last several months, Ray Dalio hasn't been shy with sharing his concerns about a potential AI bubble, and he doesn't seem to be softening his stance.
During a recent appearance on Steven Bartlett's The Diary of a CEO podcast, Dalio, the founder of investment management firm Bridgewater Associates, said he believes enthusiasm around AI has pushed markets into a bubble territory that looks a lot like the run-ups to the historic market crashes of 1929 and 2000.
This renewed warning comes after SpaceX went public on June 12 with the largest initial public offering in history, while both Open AI and Anthropic prepare to launch IPOs that are expected to push those companies into trillion-dollar valuations. As Fortune notes, this is the type of "speculative issuance surge" that financial experts and market historians believe is the clearest warning sign of a bubble.
The podcast opened with a discussion of Jeremy Grantham, the founder of Grantham, Mayo, Van Otterloo & Company (GMO), who previously told the show that we're looking at "the biggest investment bubble in American history," Dalio agreed.
"He's right," Dalio said.
The 'Four Horsemen of the Bubble Apocalypse'
Owen Lamont, senior vice president and portfolio manager for Acadian Asset Management, created a financial bubble checklist that he's called the "Four Horsemen of the Bubble Apocalypse." This checklist notes the things that will likely happen when a bubble is about to pop:
Overvaluation: Current prices are at very high levels, according to expert opinions and historical norms.
Bubble beliefs: A large number of market experts say prices are too high and will likely rise even higher.
Equity issuance: A high level of equity issuance from new firms (IPOs) and existing firms over the last year, accompanied by lower levels of repurchases.
Inflows: An unusually high number of new investors entering the stock market.
The overvaluation horseman seems to have already come to fruition. According to Advisor Perspectives, the S&P 500 is currently overvalued by 116% to 207% based on monthly data for June 2026, and Cerity Partners reports that data suggests the market is overvalued at this time.
Meanwhile, financial experts such as Grantham, Michael Burry and Torsten Sløk have all spoken about their AI bubble beliefs, sharing Dalio's concerns. There's also a general consensus to consider: according to the Globe and Mail, the Bank of America's Global recent fund managers survey found that managers believe an AI bubble is "the most pressing tail risk facing the market right now."
While speaking with Bartlett on the podcast, Dalio mentioned surging stock issuances as one of two primary factors that — along with rising interest rates — can "prick" a bubble. And this particular horseman is already in the books after SpaceX went public with a massive evaluation and has traded below its IPO price ever since.
The fourth horseman, inflows, also appears to have come to fruition. In 2025, retail investors — non-professional individuals who buy or sell stocks, bonds or funds — accounted for $5.4 trillion of trading activity, according to independent data firm Vanda. That's close to a 47% increase from 2024, Forbes reports.
During the podcast, Bartlett walked Dalio through a hypothetical situation that explains how an AI bubble might pop, as well as the difference between wealth and money.
Let's say an investor buys one stock in an AI company that's valued at $100, giving the investor a net worth of $100. That investor then goes to the bank and takes out a $50 loan. Then the market takes a turn for the worse and other investors who bought the same stock start to sell, driving a single share down to $25 — while our investor's $50 loan still needs to be repaid.
"But now this thing that I have that was worth $100 a couple of months ago is now worth $25 and I'm $25 in a hole," Bartlett explained. "So I have to quickly sell. And then with everybody selling all the prices of assets dropped. People stop spending money at the restaurants. Like you [Dalio] say, there's less money around and then the bubble has burst and we're in this sort of declining [trajectory]."
Dalio's quick response? "You got it!"
"What's quite common now is you can issue stock for, let's say you raise $50 million and you value the company at a billion dollars. Only $50 million was actually spent on that company, but now if you raise that, you're a billionaire," he added.
"Because [of] the accounting value of that, what do you own? You own stock that is valued at a billion dollars. Nobody paid a billion dollars or whatever it is, right? And now you own that stock, but that stock … you can't spend because you can't spend wealth."
This led Dalio to mention an important distinction between money and wealth.
"Wealth is not the same as money," he said. "You see a lot of people getting wealthy but you can't spend the wealth. You have to sell the wealth to get money because you can only spend money."
Asked again whether he is seeing the signs of a potential bubble, Dalio doubled down.
"Yeah, the classic signs."
What To Read Next
The tax breaks in Trump's 'big beautiful bill' expire after 2028 — and experts say most people won't act in time. What to do before the window closes
Four leading AI models discuss this article
"Classic bubble signs are present but offset by verifiable earnings and capex trends that 1929 and 2000 lacked, making a 2000-style crash probable only if AI ROI disappoints."
Dalio's comparison to 1929 and 2000 is attention-grabbing but the article cherry-picks. While AI valuations are stretched (NVDA at ~38x forward P/E, Magnificent 7 driving 60% of S&P 500 YTD gains), real earnings growth is occurring: AI infrastructure spend is exploding, with hyperscalers guiding capex up 30-50% in 2025. Retail inflows and IPOs are elevated, yet monetary conditions remain accommodative versus 2000's rate hikes. The piece downplays productivity gains that could justify higher multiples and ignores that many 2025 IPOs (SpaceX included) trade on actual revenue, not pure narrative.
The strongest case against is that every prior bubble had 'this time is different' productivity stories too; if AI monetization disappoints in 2026-27 or capex ROI collapses, the equity-issuance and overvaluation horsemen could still trigger a 30-40% drawdown regardless of current earnings.
"The current AI market is characterized by high valuation multiples supported by genuine enterprise cash flows, distinguishing it from the purely speculative, non-revenue-generating mania of 2000."
Dalio’s focus on historical analogs like 1929 and 2000 ignores the fundamental shift in corporate capital allocation. While the 'Four Horsemen' checklist captures speculative froth, it fails to account for the massive cash-flow generation of the hyperscalers—Microsoft, Alphabet, and NVIDIA—which are currently funding AI infrastructure through organic EBITDA rather than pure debt-fueled leverage. The article conflates 'wealth' volatility with systemic insolvency. While retail inflows and IPO issuance are indeed elevated, the institutional 'smart money' remains disciplined regarding margin debt compared to the dot-com era. We are likely seeing a valuation re-rating rather than a total bubble collapse, provided that enterprise software monetization begins to materialize by Q4 2026.
The thesis fails if AI capital expenditures yield diminishing returns on invested capital (ROIC), forcing companies to slash R&D and trigger a massive, liquidity-draining correction in the tech sector.
"Elevated valuations + strong earnings growth + fortress balance sheets ≠ 1929; the real trigger would be earnings disappointment, not issuance alone, and that hasn't materialized yet."
Dalio's bubble warnings deserve attention, but the article conflates valuation concerns with bubble mechanics. Yes, S&P 500 forward multiples are elevated (~22x), and yes, retail inflows and AI IPO activity are real. But the article cites June 2026 data (impossible — we're in 2025) and compares current conditions to 1929/2000 without addressing what's different: 2025 earnings growth is actually tracking 15-18% YoY, corporate balance sheets are fortress-like, and unemployment remains sub-4%. Dalio's 'wealth vs. money' distinction is philosophically sound but doesn't predict timing. The real risk isn't valuation alone — it's *multiple compression* if growth disappoints or rates stay elevated. That's not guaranteed.
If AI capex actually delivers 20%+ productivity gains over 3-5 years, current valuations on Mag 7 names could prove cheap, and Dalio's 1929/2000 analogy collapses because fundamentals improve, not deteriorate.
"The AI investment cycle may be real, with productivity gains and durable earnings potential justifying elevated multiples rather than a pure speculative bubble."
The article treats AI fervor as a bubble, citing Dalio’s warning and IPO fever. The strongest rebuttal is that AI-related demand may translate into real, persistent earnings upside, not just speculative mania; productivity gains and incremental capex could justify premium multiples. Missing context includes how much current pricing embeds AI-native cash flows, evidence of durable adoption across industries, and how regulatory, supply-chain, and talent constraints could cap upside. The data points cited (overvaluation ranges, inflows, IPOs) are noisy indicators that don’t separate AI-driven growth from broader tech wealth effects. If AI spend accelerates while margins improve, the risk to a bubble claim may be lower than it appears.
The strongest counter is that AI hype can be self-fulfilling for a while, but if enterprise AI adoption sputters or margins fail to materialize, valuations could reprice quickly; SpaceX-style IPOs are not proof of durable value.
"Strong balance sheets amplify rather than mute the downside if AI productivity disappoints."
Claude's June 2026 citation call-out is valid, but the panel uniformly underweights second-order liquidity risk: if hyperscaler capex ROIC disappoints in 2026, the same firms now self-funding via EBITDA will slash guidance, triggering correlated retail outflows and forced IPO pullbacks. This feedback loop was absent in 2000 because balance sheets were weaker then, not stronger.
"The primary risk is not a tech-specific collapse, but a broader market bifurcation caused by the crowding-out effect of high AI capital expenditures on the rest of the economy."
Grok, your focus on a 2026 feedback loop misses the structural shift in corporate treasury management. Unlike 2000, current hyperscalers hold massive cash piles that act as a buffer against CAPEX volatility. The real risk isn't just ROIC disappointment; it’s the crowding-out effect. If AI spend remains elevated while interest rates stay 'higher for longer,' the cost of capital for non-AI firms will spike, creating a bifurcated market that crashes the broader index while tech merely stagnates.
"Bifurcation risk is real, but the mechanism depends on whether rate persistence is exogenous (Fed policy) or endogenous (weak growth forcing cuts)—the panel hasn't separated these."
Gemini's 'higher for longer' crowding-out thesis is underexplored. If rates stay 4-5% while hyperscalers hoard cash for AI capex, mid-cap and small-cap borrowing costs spike, crushing non-AI sectors. But this assumes rates don't fall if growth disappoints—which would ease pressure. The real bifurcation risk isn't rates staying elevated; it's asymmetric earnings: Mag 7 holds up while Russell 2000 collapses 25-30%, creating false comfort in headline indices.
"Regulatory and hardware-supply risks could cap AI ROI, triggering quicker re-rating and broader margin pressure, widening bifurcation."
Gemini's crowding-out argument assumes a benign capital backdrop; I’d add a regulatory and hardware-supply risk that could cap AI ROI and trigger a quicker re-rating. If export controls on chips tighten, or GPU shortages persist, hyperscalers may not achieve the ROI hoped, forcing tighter capex guidance and broader margin pressure. In that case, the supposed bifurcation could widen as AI leaders outpace others, while non-AI peers languish.
The panelists agreed that while AI valuations are high and retail inflows are elevated, the current situation differs from the 2000 bubble due to factors like real earnings growth, strong corporate balance sheets, and productivity gains. However, they also highlighted significant risks, including potential multiple compression, liquidity risks, and the impact of higher interest rates on non-AI sectors.
Real, persistent earnings upside driven by AI-related demand and productivity gains.
Multiple compression if growth disappoints or rates stay elevated, and potential liquidity risks due to correlated retail outflows and forced IPO pullbacks.