The Bubble Nobody Wants to Name

The Anatomy of a Bubble

When David Rosenberg calls something a bubble, he has earned the right to use the word. He called the housing bubble in 2007, the dot-com bubble in the late 1990s, and the commercial real estate and LBO bubble of the late 1980s. In every case, he was dismissed: the word was too emotional, the concern premature, the case incomplete. And in every case, the name stuck, just not until after the collapse.

Today, Rosenberg is saying it again. And the structure of what he is describing is more dangerous than the tech mania of the late 1990s, not because the technology is less real, but because the bubble has spread far further.

"The bubble was not in the internet technology," Rosenberg told Insight Is Capital. "It was in investor behavior and what they were pricing in for the future. It's the investor behavior that is the bubble, not the technology itself."

The critical difference between then and now is what Rosenberg calls the engulfment of the entire market. During the dot-com era, the mania was concentrated: technology, media, and telecommunications. The rest of the market offered cover. Today, Caterpillar is an AI stock. Utilities are AI stocks. Consumer discretionary, industrials, energy infrastructure — all of them carry the AI trade's correlation profile. The bubble has nowhere left to hide from.

The CAPE multiple has been above 40 for five consecutive months, a level exceeded only at the height of the dot-com mania. And even that comparison understates the current risk: in 2000, many of the companies driving valuations had no earnings at all. Today's market is the second most expensive in recorded history, priced for a future that may not arrive on schedule.

The AI Equation: CapEx, Dilution, and Circular Financing

The capital expenditure numbers being cited across the technology sector have taken on a surreal quality. Nearly $800 billion in planned CapEx for 2026 alone, with comparable figures forecast for 2027 and 2028. Rosenberg does not dispute the ambition. What he questions is the math behind the return on investment, and whether anyone has seriously run it.

"Revenues in the AI space have to go up compounded per year by fifty percent off a high number, per year, for the next five years," Rosenberg said. "If you believe that, then stick with it. If you don't, it's time to take chips off the table."

What makes the current cycle structurally distinct is the emergence of circular financing. The largest hyperscalers, once the most prodigious cash-flow generators in corporate history, have moved to net free cash flow negative as CapEx absorbs what used to be returned to shareholders. Where they were once buying back stock aggressively, several are now issuing equity. The 27% year-over-year earnings growth being cited for the current quarter is partly the product of fiscal tailwinds from Washington's "One Big Beautiful Bill," full expensing allowances and construction incentives that have inflated the GDP and earnings base in ways that cannot be extrapolated into the future. The bill contributed between a quarter and half of this year's US economic growth. That tailwind fades.

The Inflation Myth: Why Tariffs Are Deflationary

When Rosenberg characterizes the current tariff-driven price shock as deflationary rather than inflationary, the room goes quiet. It is a counterintuitive claim. But the economic mechanism he describes is consistent with historical precedent: the Gulf War I oil shock, the energy surge of 2008, the tariff shock of 2025. In each case, the commodity price spike was followed not by sustained inflation but by declining real incomes, contracting consumer demand, and eventual disinflation.

The key variable, in Rosenberg's framework, is wages. Consumer price inflation becomes durable only when it seeps into the labor market and triggers a wage-price spiral. That is precisely what happened in 2022 and 2023, when generous jobless benefits and constrained labor supply pushed wages sharply higher. It is not what is happening now.

The evidence Rosenberg cites is specific: average hourly earnings grew just 0.1% in September, implying negative real income on a monthly basis. New home prices are down 7.5% year-over-year. Mortgage applications have fallen nearly 40%. Core goods inflation, the category directly exposed to tariff pass-through, has decelerated from 1.5% to 0.7%, not accelerated.

The implication for the Federal Reserve under new chair Kevin Warsh is pointed. Where Jerome Powell declined to raise rates into a supply shock and was vindicated by events, Warsh has moved aggressively into an environment where the inflation data does not yet justify that posture. Rosenberg believes the Fed will be cutting rates by next year, following the same pattern observed after every tightening cycle since 1989.

The Bond Opportunity Nobody Wants to Discuss

At a time when the consensus views fixed income with barely concealed contempt, a consequence of the brutal drawdown in duration from 2021 through 2023, Rosenberg has been adding duration to his model portfolio. He acknowledges this makes him deeply unfashionable. He also thinks it makes him right.

The arithmetic is not complicated. With the 10-year Treasury yielding above 4.5%, an investor needs yields to rise a full 100 basis points from current levels just to lose money on a one-year holding period. Meanwhile, the S&P 500 offers a dividend yield of 1%. The equity risk premium, the excess return equities are supposed to provide over risk-free bonds, has turned negative. The last time that happened was in the late 1990s.

The household balance sheet data tell a stark story. American households currently hold 73% of their financial assets in equities, the highest concentration ever recorded, and just 7% in bonds. This is not the result of an intentional allocation decision. It is the passive consequence of a multi-year equity bull market that has never been rebalanced. A classic 60/40 portfolio, left unattended since 2023, is now effectively a 75/25 — and that equity portion is itself concentrated in a handful of mega-cap names that carry the AI trade's full correlation. The average and median stock in the S&P 500 is already down 15% from its highs, even as the index trades near its peak.

The Boomer Reckoning

Of all the concerns Rosenberg raised in this conversation, the most viscerally affecting is the least discussed: the demographic asymmetry sitting inside the equity market's most loyal constituency.

The median Baby Boomer is now 70 years old. At that age, conventional financial planning suggests an equity allocation of roughly 30%. The median Boomer currently holds 55% of their financial portfolio in equities. And more than half of what they own sits in passive index ETFs, which means many may have no clear sense of how concentrated their exposure actually is, or how close to the surface the technology correlation runs.

What makes this structurally different from the dot-com era is the age of the cohort and what that implies about recovery time. When the Nasdaq fell between 60% and 80% from 2000 to 2002, the median Boomer was 45. They had decades of compounding ahead, years of contributions still to make, and a career in which to absorb the shock. Today, at 70, those buffers are gone. A serious drawdown is not a setback to be recovered from. For tens of millions of people, it is a permanent impairment to lifestyle and retirement security.

The social and political implications of a generation of retirees facing a simultaneous equity drawdown are real, and they will not be contained to the financial pages. "Will they find a robot has taken that [cashier] job over?" Rosenberg asked. It is partly rhetorical. But only partly.

What True Diversification Actually Looks Like

Rosenberg's model portfolio, the strategy now formalized in the Corton Rosenberg Global Macro Fund ETF (ROSY), has returned approximately 60% since its inception in early 2023, with a beta of just 0.4 to the S&P 500. It is, in its construction, an argument made in numbers: that capital preservation and rational risk management are not the same thing as being out of the market.

The portfolio holds hard assets: oil, energy infrastructure, rare earth metals, pipelines. Gold remains a holding above $4,100 an ounce, a level Rosenberg characterizes as a "miracle" given headwinds from rising real rates and a strong dollar, one sustained by central bank demand that tells its own story about confidence in reserve currency stability. Fixed income sits at 30% of the portfolio, nearly five times the average household allocation, with an active mandate to extend duration. Cash has been raised to 10%.

Equities remain in the portfolio, expressed through geographies and sectors with real economic substance: Asia (which captures semiconductor and technology supply chain exposure in Taiwan and Korea), global healthcare, aerospace and defense, and biotech. The S&P 500 equal-weight index, he notes, is a meaningfully better approximation of diversified equity exposure than the cap-weighted version currently dominated by ten companies representing over 40% of the index.

Canadian Equities: The One-Eyed King

Rosenberg is constructive on Canadian equities as a relative value play. Canada carries a positive equity risk premium, meaning investors are actually paid to take on equity risk, unlike the US market today. Its index has minimal technology exposure, limiting its vulnerability to an AI trade reversal. Its financial sector has already corrected more than 10% from recent highs, absorbing much of the valuation pressure. And its commodity-heavy composition aligns with Rosenberg's hard asset thesis.

"If we roll over on this AI trade, Canada will be the king with one eye, while the US might not have one at all," Rosenberg said. The 2003-to-2013 precedent, the decade when Canadian equities meaningfully outperformed the S&P 500 following the dot-com collapse, provides a historical analogue that is difficult to dismiss.

"If I was making money in the US market and wanted to deploy any profits, I would put part of it in the Canadian market right now."