Odd Lots2026.06.181 hr

Jeremy Grantham on AI Capital Warfare and the Anatomy of Market Bubbles

Original title · Jeremy Grantham on How to Tell If a Bubble Is About to Burst

用阿法問答深入這集 ✨

或自己問：

The rapid rise of AI has blinded retail investors to a key structural shift: the Magnificent Seven are morphing from secure vertical monopolies into a capital-heavy, highly competitive dog fight. While the technology is revolutionary, market history warns that transformative innovations—much like 19th-century railroads—routinely wipe out initial investors before delivering on their structural promise. Navigating this momentum-driven market requires recognizing the classic warning signs of a monumental bubble.

The secure vertical monopolies are turning into a dog fight

Grantham was sharpest when pointing out how big tech's business models are shifting. Historically, these companies dominated distinct, near-monopolistic niches—like search or social media—with sleepy regulators letting them compound capital-light cash flows. Today, they are entering a capital-heavy collision course over AI, boasting about hundreds of billions in competing investments. For retail investors, this is easy to overlook; instead of peaceful oligopolies, we are entering a margin-eroding capex war where seven giant players are fighting for the same first-mover crown.

Railroads and the internet prove real tech still busts

A common defense of tech valuations is that the earnings power is real, exemplified by Nvidia’s revenue soaring from $11.7 billion in 2019 to $130 billion in 2025. Yet, Grantham’s historical lens connects this directly to the 19th-century railroad boom and the 1999 internet bubble, both of which featured genuinely world-changing tech. In those episodes, investors still lost nearly everything—Amazon famously fell 92% post-crash—before the physical infrastructure was recycled to change society. Realizing a technology's revolutionary potential does not protect you from a devastating capital loss when a speculative bubble bursts.

Speculative leaders rolling over is the ultimate tell

While critics label him a perma-bear, Grantham’s actual track record shows only two absolute "abandon ship" calls: in July 2008 and late 2021. This latest bear market was uniquely interrupted by ChatGPT, sparking an unprecedented capex boom that revived animal spirits. To time the eventual turn, the historical signal worth tracking is when previous speculative leaders—like meme stocks or pre-revenue plays trading at astronomical price-to-sales ratios—begin to decline quietly even as the broader index powers ahead.

The key thing to remember is that massive capex spending has temporarily delayed a standard recessionary cycle, but the long-term mathematical reality of valuation mean reversion remains undefeated. Watch for a divergence where speculative high-flyers quietly lose ground while the major indices hold their highs—historically, that has been the market's greatest warning sign.

Assets mentioned in this episode

NVDANvidia▼ 看空
Nvidia is framed as having stumbled into a monumental hardware head start, benefiting from the sudden alignment of its gaming chips with the needs of modern AI workloads. While acknowledging its dramatic revenue growth, the stock is viewed as a highly vulnerable bubble candidate because competitors are aggressively developing rival silicon and architectures that could eventually bypass this head start. Under this view, its current premium is highly speculative and unlikely to sustain a multi-decade valuation anchor.
MSFTMicrosoft◆ 分歧
Microsoft serves as an example of a stock that historically qualified as cheap under institutional value models due to its strong return on equity and near-monopoly pricing power. However, looking forward, its pivot into a capital-heavy battle over the AI market alongside other tech titans is treated with caution. Rather than enjoying a peaceful, highly profitable monopoly, it now faces a messy, competitive dogfight that could dramatically depress future returns compared to its historical performance.

Key questions

Why is the current market setup compared to 1929 and 2000?→

The market is showing a rare 'internal divergence' where broad indices rise while speculative high-flyers and junk stocks roll over. This pattern has historically preceded major crashes, signaling that liquidity is withdrawing from the market periphery.

💡 詢問阿法，看詳細逐字稿分析→

How is the business model of the 'Magnificent Seven' changing?→

They are shifting from capital-light, high-margin software monopolies into a capital-heavy arms race for AI dominance. This massive, competitive capital expenditure risks eroding margins, moving these firms toward a lower-return environment than investors currently expect.

💡 詢問阿法，看詳細逐字稿分析→

Why does the notes suggest that revolutionary tech is often a bad investment?→

Historical manias like railroads and the dot-com era show that transformative technology does not guarantee investor success. Early pioneers often face massive capital destruction as profits revert to the mean before the industry stabilizes and matures.

💡 詢問阿法，看詳細逐字稿分析→

Further research

Tickers and signals often linked to this episode's themes in public sources · AI-compiled, not investment advice

AI Capex Cycle Sensitivity

Capital expenditure projections for the Magnificent Seven in 2026 have expanded to over $660 billion, transforming corporate capex into the dominant liquidity driver over traditional interest rate signals.

US stocks

NVDA
NvidiaBenefitsAs the primary supplier of advanced GPUs, Nvidia directly monetizes the massive AI capital expenditures of hyperscalers, serving as the central beneficiary of the infrastructure spend cycle.
MSFT
MicrosoftPressuredThe massive capital expenditure required to secure GPUs and build out AI infrastructure places significant pressure on Microsoft's near-term free cash flows and operating margins as it struggles to monetize these investments at the same scale.

Risks

A sudden slowdown in generative AI adoption or a shift toward less compute-intensive model architectures could trigger a sharp contraction in hardware orders, leading to massive inventory write-downs.

Watch list

Quarterly capital expenditure guidance updates from hyperscalers (Microsoft, Alphabet, Amazon, Meta)
Nvidia GPU delivery lead times and supply-chain backlog data
Debt issuance and private credit financing activities by major cloud providers to fund data centers

Internal Market Divergence

A technical pattern where speculative growth names decline while index-weighted giants remain elevated has historically served as a critical leading indicator of structural market corrections.

US stocks

VTV
Vanguard Value ETFBenefitsVanguard Value ETF benefits from defensive sector rotation as capital flees speculative high-fliers in search of historical valuation support and steady earnings.
ARKK
ARK Innovation ETFPressuredARK Innovation ETF holds highly speculative, non-profitable tech names that historically roll over first when market divergence begins, signaling eroding liquidity beneath index leaders.

Risks

A broad-based economic re-acceleration paired with cooling inflation could lift speculative valuations across the board, resolving the divergence without a major index correction.

Watch list

Market breadth indicators such as the percentage of S&P 500 components trading above their 200-day moving average
Performance ratio of equal-weighted S&P 500 (RSP) compared to cap-weighted S&P 500 (SPY)
Weekly fund flows into defensive value equity funds vs. speculative technology ETFs

Tech Business Model Transformation

The transition of Big Tech from asset-light software monopolies to capital-intensive AI physical infrastructure operators threatens to compress structural operating margins.

US stocks

AVGO
BroadcomBenefitsBroadcom benefits as a key developer of custom AI ASICs, allowing it to capture high-margin revenue from the hyperscaler arms race without taking on the depreciation burden of data center ownership.
AMZN
AmazonPressuredAmazon faces structural margin pressure as its capital expenditure rises to a projected $200 billion in 2026, transitioning its high-margin cloud model into a capital-heavy physical infrastructure race.

Risks

A rapid commercialization of proprietary AI features could drive highly scalable, high-margin subscription revenues that completely offset the increased depreciation and infrastructure expenses.

Watch list

AWS and Google Cloud segment operating margins in quarterly earnings releases
Year-over-year growth of depreciation and amortization expenses relative to revenue growth for hyperscalers
Capital expenditure-to-revenue ratios of major tech conglomerates

This section is AI-compiled from public sources, may be inaccurate or outdated, is for research reference only, and is not investment advice.

Want the full episode?