Speaker
Tobias Carlisle
Appearances over time
1 episodes
Episodes
1
Podcasts
Quotes & moments
Every major market valuation metric tracked by Advisor Perspectives is simultaneously at its most overvalued level in the entire historical dataset.
Small caps have outperformed the Magnificent 7 over the last 12 months and year-to-date, a fact Carlisle calls a narrative violation most investors don't know.
The spread between the most expensive and most undervalued stocks is in the 95th percentile, meaning it has only been wider on 5% of historical occasions.
The current large-cap growth market dominance has persisted for over 10 years since 2015, preceding a potential extended value market cycle.
Unlike fiber optic cable or railways which last 25+ years, AI computing hardware depreciates and becomes obsolete in roughly 5 to 7 years.
Even Google, a historically cash-rich company, is now raising $80 billion in capital to fund AI infrastructure, a shift Carlisle called 'bonkers.'
Energy currently represents only about 3% of stock market capitalization, well below its historical average of approximately 12%.
Rebalancing monthly adds little return versus quarterly, while annual rebalancing risks missing major bottoms like March 2009 by several months.
Meta spent approximately $12 billion on the metaverse before reversing course — Carlisle cites this as evidence that current AI CapEx is more discretionary than it appears.
Carlisle traced the same investor over-excitement pattern through the telegraph boom, electronics boom, railway boom, dot-com boom, and now the AI boom — each ending with a crash.
US small, micro, and mid-cap companies experienced an earnings recession from 2022 through late 2024/early 2025, with earnings falling or trading sideways while large-cap earnings rose.
The Acquirer Fund ETF (ZIG) rebalances quarterly and Carlisle has the discretion to rebalance more frequently as an active fund, but chooses not to.
Small caps have outperformed the Magnificent 7 over both the last 12 months and year-to-date — a narrative violation almost nobody is talking about. Equal-weight indexes, small caps, and value are all signaling the same thing: a market leadership shift away from large-cap growth may already be in progress.
Every major market valuation metric is simultaneously at its most overvalued in the historical dataset — but that doesn't mean you should exit the market. Overvaluation signals lower forward returns and volatility, not an immediate crash. The real opportunity is in the parts of the market that haven't been bid up.
The entire AI infrastructure buildout may end up enriching users rather than investors. Competition among AI providers makes commoditization inevitable — most people will use cheaper, older models rather than pay for the cutting edge. The value creation is real, but investors may not capture it.
Fiber optic cables and railways last 25+ years. AI data center hardware becomes obsolete in 5 to 7 years. This changes the economics of the buildout entirely — the returns need to come faster, and the depreciation risk is far higher than in prior technology booms.
The Mag 7's AI capital expenditure is more discretionary than it looks. Just like Zuckerberg pivoted away from the metaverse after spending $12 billion, these companies could pull back if investor sentiment shifts. The real demand signal is coming from stock markets, not from actual consumer need.
Three signals are converging that may mark the top of the AI boom: the SpaceX IPO is forcing the S&P 500 to absorb a massive new entrant, OpenAI and Anthropic seeking $80 billion are pulling liquidity from the market, and Time Magazine just put AI on the cover — the same indicator that called 'death of equities' in 1979.
The pattern is the same every time: telegraph, railways, electronics, dot-com, and now AI. Investors get too excited, bid up picks-and-shovels sellers alongside the technology itself, and eventually everything returns to earth. The human behavioral response has never changed, even as the technology itself has transformed.
Value investing has worked consistently in non-disrupted industries throughout the current tech boom — the underperformance is concentrated in disrupted sectors. The challenge is identifying disrupted industries in advance. For non-disrupted businesses trading at discounts, value metrics have been remarkably reliable.
Energy stocks represent just 3% of market cap versus a 12% historical average. AI data centers will consume energy voraciously, and Carlisle thinks nuclear is the only long-term solution — though natural gas bridges the gap. Oil equities are doing the opposite of what everyone expects right now.
The process is grounded entirely in financial statements: look at what a company earns on assets, its reinvestment rate, and its payout policy over 5-10 years. Combine with the Acquirer's Multiple (which includes balance sheet items) to find the best risk-adjusted opportunities. Equal weight everything because you can't know in advance which names will deliver.
Rebalancing monthly adds almost no extra return over quarterly, but annual rebalancing risks catastrophic timing luck — like having a September rebalance date in March 2009. Quarterly also captures a little momentum before cutting winners. It is the practical sweet spot for systematic value ETFs.
Michael Mauboussin's base rate framework showed that OpenAI's growth projections would be unprecedented in corporate history. Then Anthropic came along and obliterated those already-unprecedented projections. The tech giants keep breaking the models used to value them.
Analysis
What they talk about
- Business 55%
- Technology 45%
Connections
Shows they appear on and people they share episodes with. Drag to explore.
