PodcastThe Monetary Matters NetworkBen Inker

The AI Bubble Will Pop from Equity Supply Says This $80 Billion Asset Manager | Ben Inker | GMO

Ben Inker sees the AI buildout as vulnerable to overinvestment: intense competition and uncertain returns could weigh on profitability, data-center economics, and AI-exposed stocks. He also highlights how a surge in IPOs could strain the market's ability to absorb equity supply, while warning that concentrated AI credit exposure could create trouble even without widespread defaults. GMO favors cheaper non-US stocks and value over growth, while Inker describes a possible 35% to 50% market decline as a downside scenario, not a forecast.

Artificial IntelligenceMarket ValuationsEquitiesMarket LiquidityInvestment StrategyRisk Management

Preview · 28 more points in the app

Read the Full Summary

Key Takeaways

Inker argues that AI can be transformative without generating attractive returns for investors: low switching costs and comparable offerings may require low prices, while heavy investment risks overbuilding data centers before demand is clear. He points to the internet-era Nasdaq decline of at least 80% as a reminder that a severe market loss would not mean the technology failed to change the world.

Inker describes a conditional downside scenario in which disappointing AI returns lead companies to cut inference spending, easing the compute shortage and potentially making older data centers uneconomic. He says that could hurt hyperscaler returns and trigger defaults or company failures, but a buildout bust need not become a broader financial disaster if losses fall on entities that are not highly leveraged or bank-funded.

Inker warns that planned AI investment could concentrate debt in the sector even without a massive wave of defaults, while data centers increasingly rely on bank, private-credit, and loan markets, potentially including layered lending through non-bank institutions. He doubts the government could backstop an AI downturn without risking the bond market, given high public debt and a large deficit.

Inker argues that markets may struggle to absorb a surge in new share supply because passive investors cannot reposition by mandate, retail investors may not focus on valuation, and active managers can be reluctant to take tracking-error risk. He estimates planned IPOs by SpaceX, Anthropic, and OpenAI could represent 6% to 7.5% of US market capitalization; historically, each additional 1% of market cap from IPOs has been associated with the market doing about 4% worse that year, a challenge rather than a certain decline.

Inker says an AI-focused market decline could reach 35% to 50%, though he is not sure it will happen, and identifies hyperscalers, semiconductor companies, tech hardware, and some software as exposed. He also cautions that AI-related demand can lift companies outside the usual technology group, such as Toto, and that investors would be at risk if they valued its porcelain-related semiconductor business as though it were the whole company.

TECHNOLOGY

AI Bubble Risk — Inker argues that transformational technologies often attract capital before investors know who will capture the returns. AI may change the world yet still prove to be a bubble, he says, because returns can deteriorate as capital pours in.

“Why do we think AI is a bubble? I'd say kind of at the highest level, it is because it is a transformational technological change…”—

AI Competition — Inker argues that AI compute and models look more competitive than winner-take-all businesses because switching costs are low and offerings are comparable. Broad AI adoption requires low prices, he says, making exceptionally high profitability harder to sustain.

“Second, when we look at the basic characteristics of this business, what we see is a lot of this is about either the provision of compute or the provision of…”—

Data Center Overbuild — Inker expects AI investment to lead to excessive data-center construction and large write-offs. Projects take years to build and require forecasting demand for a relatively new technology; the market can turn before the scale of overinvestment becomes clear.

“Well, time will tell, but you're absolutely right. in the internet bubble. The bubble rolled over before it was clear the scale of the malinvestment…”—

ECONOMY

Trading Firm Limits — Inker says multistrategy firms identify opportunities within a particular space but generally do not trade unlike assets because doing so is too risky. This limits their ability to absorb a wider range of market imbalances.

“They do a very good job of figuring out where in a particular space there is the opportunity, but they're really not trading in stocks or assets that are unalike…”—

Risky Credit — Inker says triple-C debt was around the 74th percentile of historical spreads at the end of September. Some risky capital structures created when interest rates were zero no longer work at 5% and may require restructuring; GMO's credit team is cautiously exploring them.

“The triple C's, the lower rated debt, that is no longer the case…”—

AI Credit Exposure — Inker warns that if planned AI investment proceeds, AI debt could become unusually concentrated and create trouble even without a massive wave of AI defaults. Higher borrowing costs and competition for capital are squeezing non-AI borrowers, including housing, autos, and other long-dated investments.

“And if we get all of the investment that is slated to come, this will become a profound problem, right?…”—

MARKETS

Railroad Overbuilding — Inker says excess investment can damage returns across the broader investment base, not just on new projects. He points to the UK railway boom, when six railroads between London and Manchester exceeded what was needed.

“And as more and more capital comes in, even if it was the case that demand was growing faster than supply for some period of time, eventually as everybody who…”—

Equity Supply — Inker argues that markets have less capacity to absorb new share supply when fewer investors adjust portfolios in response to price changes. Passive investors cannot reposition by mandate, retail investors may not focus on valuation, and active managers can be reluctant to risk tracking error as markets become more concentrated.

“Honestly, a big portion of retail is not focused on the price that they are paying on the valuation of the assets…”—

IPO Supply — Inker says planned IPOs by SpaceX, Anthropic, and OpenAI could create an unusually large increase in US equity supply. He estimates that 6% to 7.5% of US market capitalization could come through IPOs. Historically, each additional 1% of market cap from IPOs has been associated with the market doing about 4% worse that year; he stresses this is a challenge, not a certain decline.

“And in aggregate we are talking about assuming open AI and anthropic do come public maybe six seven and a half% of the aggregate market cap of the US which…”—

FORECASTS

SpaceX Valuation — Inker says that if SpaceX trades at roughly 100 times sales, its shares could fall substantially even if the business remains viable. Businesses that fail could fall 100%.

“Right? If you are trading like SpaceX at whatever 100 times sales, there's a long way down you could go…”—

Intro

Ben Inker explains that transformative technologies often attract more capital than can earn attractive returns, and sees AI compute and models as competitive businesses with limited switching costs rather than durable winners.

AI Bubble Setup

Inker argues that passive ownership, index tracking, and growth-oriented investors make markets less responsive to rising prices, so unprecedented IPO supply from SpaceX, Anthropic, and OpenAI could pressure existing shares.

Equity Supply Matters More

He estimates upcoming IPOs could add roughly 6% to 7.5% of U.S. market capitalization, exceeding past records; historical IPO-heavy years were followed by weak markets, though the estimate is not a guaranteed crash forecast.

IPO Wave Math

Inker says IPO supply could make markets perform materially worse than normal, with AI beneficiaries especially exposed; the scale of losses may depend on valuations and how central AI is to each company's business.

Dotcom Fiber Parallel

Using railroads and internet-era fiber as parallels, Inker warns that markets can turn before investors recognize overbuilding, while long data-center construction timelines make forecasting AI demand especially difficult.

Finding Non-AI Bets

GMO seeks reasonably priced businesses with sound returns and possible AI upside, rather than portfolios concentrated in AI; Inker notes that overlooked firms such as Toto can benefit indirectly from semiconductor investment.

Passive Flows and Liquidity

Inker says passive flows and benchmark-conscious active managers limit price-sensitive trading, while value investors are more willing to sell into rising prices; that reduced flexibility makes the market harder to absorb new supply.

Credit Risks and Spillovers

He is cautious on credit overall, noting tight spreads and concentrated AI lending; even without widespread defaults, heavy exposure to AI debt could strain diversification, while high rates also squeeze housing and other borrowers.

Can Government Backstop AI?

Inker doubts the government can fund a major AI rescue without risking the bond market, but says losses may remain manageable if they sit with investors able to bear them and the Fed can cut rates in a recession.

Credit Stress Check

GMO sees little clear distress in public credit markets and acknowledges uncertainty in private-credit defaults; a resilient global economy has helped, though high energy costs are already raising costs and interest-rate pressure.

Energy Portfolio Benefits

GMO favors resource stocks for their diversification and exposure to some inflation shocks, while also valuing high-quality companies; Inker emphasizes that actual portfolio choices should reflect prices paid.

Value Versus Growth Spreads

Inker says GMO prefers non-U.S. equities and value stocks because they are cheaper than U.S. growth shares; wide valuation spreads can reward value as companies change categories over time.

Ex-US AI Winners

GMO remains exposed to Korea and Taiwan through emerging-market holdings, but Inker is cautious about whether recent profit surges are sustainable and has trimmed positions as earnings outpaced share prices.

Big Tech Commoditizes Compute

Inker says U.S. profits are unusually concentrated in large firms with market power, but AI compute and models are more competitive and interchangeable, potentially making hyperscalers' capital-heavy investment less profitable.

What If Mega Caps Go Sideways?

Inker says the largest stocks have historically tended to lag, and current giants could face lower returns as investment needs rise; prolonged sideways performance might gradually shift investors away from U.S. equities.

Memory: Cyclical Peak Earnings Trap

He warns memory makers can look cheap near a cyclical earnings peak because they are commodity producers, and that margin normalization could pressure revenue; their current cash flows nevertheless provide some investment value.

The Major Crash Scenario

Inker describes a downside scenario in which weaker AI returns reduce compute demand, trigger data-center write-offs and defaults, and cause a 35% to 50% market decline; he stresses this is possible, not his forecast.

Conclusion

The host closes by pointing to upcoming IPO unlock timelines, thanks Inker for joining, and directs viewers to GMO's research before delivering sponsor disclosures and thanks.

Related Episodes

Important Disclaimers

Not Investment Advice: For education only — not financial, tax or legal advice.

AI-Generated: Made by AI from source transcripts and may contain errors.

No Guarantees: Past performance does not predict future results; investing carries risk.

Our Commentary · Request Removal