The $2 Trillion Question | Tobias Carlisle on SpaceX, the AI Buildout, and the Rotation No One Sees

Episode Summary

Executive Summary: Toby of Acquirers Funds argues that broad U.S. equities are historically expensive, but that this does not justify exiting the market; instead, it favors small/mid-cap value and quality where forward returns look better and valuations remain near long-term lows. The conversation also examines AI capex booms, whether benefits accrue to consumers or firms, and how market concentration, debt-funded spending, and IPOs may echo prior technology manias.

Main Topics: Market valuation extremes and mean reversion (Priority: 5/5): Toby says major market valuation metrics are at or near record overvaluation, but that the right response is rotation, not abandoning equities. He believes high starting valuations imply lower future returns and greater volatility, especially for large-cap growth. Rotation from large growth to value/smaller caps (Priority: 5/5): He sees evidence of a regime shift: equal-weight S&P, small caps, and value have started outperforming mega-cap growth after a long stretch of large-cap dominance. He thinks the market is in an early, volatile transition toward a value-led period. AI capex boom: transformative but potentially low-moat (Priority: 5/5): The discussion centers on whether AI spending will create durable profits for model builders or mainly benefit users. Toby thinks AI is powerful and transformative, but likely to commoditize quickly, with returns accruing more to consumers and operators than to model owners. Funding, concentration, and the Mag 7 (Priority: 4/5): They discuss how AI spending has shifted from cash-flow funded to more debt-heavy structures and SPVs, but Toby emphasizes that the bigger issue is the market’s demand for the spend, not just company financing. He sees this as part of a broader large-growth bubble, not only an AI story. IPO supply and cycle-topping behavior (Priority: 4/5): The hosts discuss major upcoming IPOs such as SpaceX, OpenAI, and Anthropic. Toby suggests that when private winners go public after capturing most of their growth, it can signal late-cycle behavior and a broadening public awareness that often marks tops. How the Acquirers ETFs are built (Priority: 4/5): Toby explains that his strategy uses financial statements to estimate base/bull/bear outcomes, combining value and quality across cyclicals and higher-quality franchises. He equal-weights positions and prefers quarterly rebalancing to balance timing luck and transaction costs.

Key Arguments: Current market valuations are exceptionally high across multiple measures, but that does not automatically mean selling equities; it means seeking better forward-return pockets. Large-cap growth, especially the biggest index constituents, appears stretched relative to history, while small-cap, mid-cap, and value areas look more attractive on long-term valuation bases. AI is unquestionably useful and transformative in practice, but its commercial economics may resemble earlier infrastructure booms where value accrues broadly and commoditizes over time. The biggest AI companies may be in an arms race driven by investor expectations as much as end-user demand; if market pressure eases, capex could slow. Mega-cap growth valuation may be harder to justify than it appears because the price already embeds extraordinary future growth and margin persistence. The current period resembles earlier technology booms in which a narrow set of large-growth stocks dominate, then eventually revert as valuation spreads compress. Value investing still works in disruptive periods, but disrupted industries can underperform if multiples contract faster than earnings stabilize. A disciplined, financial-statement-driven framework can identify mispriced companies even when the market narrative is unfavorable. Quarterly rebalancing is a practical compromise: frequent enough to reduce timing luck, but not so frequent that it increases turnover and fights momentum too much.

Data Points: Market valuation metrics: 6-7 major metrics - Toby references Advisor Perspectives tracking multiple market-level valuation indicators, all near or at their most overvalued levels. Schiller P/E: 44x at the U.S. market high comparison - He says the U.S. could surpass prior valuation highs and compares current levels to historical extremes like Japan and China. Japan valuation peak: 100x - Used as an example that prior all-time valuation peaks are not hard ceilings. China valuation peak: 100x - Same argument: extreme valuations can go beyond previously observed highs. SP 500 market concentration: 35%-40% - Justin notes roughly this share of the S&P 500 is in very large tech names driving index performance and overvaluation. Small caps vs Mag 7: Outperformed this year - Toby says small caps have outperformed the Magnificent Seven recently, contrary to popular narrative. Equal-weight S&P history: Back to ~1990 - The equal-weight versus cap-weight comparison is available roughly from the 1990 launch period. Valuation spread percentile: 95th percentile - He says the spread between the most expensive and cheapest stocks is at an extreme relative level. Valuation spread incl. quality: 10th-15th percentile - With quality adjustments included, the setup still looks highly favorable to cheaper names. Small/mid earnings slump: 2022 to late 2025 - He describes an earnings recession/flat period for small and micro/mid-cap companies during this span. AI hardware useful life: 5-7 years - Toby argues AI compute depreciates faster than rail or fiber-optic infrastructure. Fiber optic useful life: ~25 years - Comparison used to show how AI infrastructure may have shorter economic durability. Rail infrastructure useful life: 25+ years - Another benchmark for comparing AI capex to prior buildouts. OpenAI growth projection window: 3-5 years - Justin mentions a base-rate paper projecting growth over this horizon. Anthropic growth: Exceeded OpenAI projection - They say Anthropic’s actual growth has surpassed a recent modeled projection. Semiconductor valuation example: 55x earnings - Justin cites a stat implying very high terminal expectations for semis. Semiconductor implied growth: 16.5% annually for 10 years - Derived from the 55x earnings example, illustrating aggressive embedded assumptions. Timing of rebalancing: Quarterly - Toby says quarterly rebalancing is his preferred compromise for the ETFs. Large value vs growth swap: 5% turnaround - He mentions a period where value outperformed by roughly 5 percentage points around the SpaceX IPO discussion.

Pivotal Quotes: "It’s entirely conceivable that all of this work goes into creating these incredible AI models and all of the value accrues to the consumer and not to the people who create these models." — Toby: Used to frame the possibility that AI model economics commoditize even if the technology itself is transformative. "Ultimately, those multiples do mean revert growth rates, mean revert. If you believe in mean reversion, then the smart bet is smaller micro value, mid-cap value." — Toby: Core investment thesis on why current market extremes favor value and smaller caps. "I think it’s more the stock market that’s demanding it rather than consumers demanding it." — Toby: Explains his view that AI capex intensity is being driven by investor expectations and competitive signaling.

Implications: Listeners should expect continued volatility and style rotation if valuation spreads keep compressing. The conversation suggests AI may be huge technologically but less monopolistic economically than bulls expect, which could favor value, small caps, and firms that adopt AI well rather than the builders alone.