Notes — Jim Grant on the AI Bubble, Decadent Finance, and the Lessons of History

Four questions [Adler frame]

Q1 — What is it about? Jim Grant, editor of Grant’s Interest Rate Observer since 1983, argues in October 2025 that US markets are at or near a major top: the S&P 500 CAPE stands at 40x (second highest in US history after the 1999 dotcom peak at 44.2x), the AI capital expenditure race mirrors the 1990s fibre-optics check-writing contest, private equity is in structural trouble as institutions can’t get their capital back, and the government is running a 6–7% fiscal deficit in conditions that historically define full employment. Interwoven is a discussion of Grant’s book Friends Until the End, a double biography of 18th-century British parliamentarians Edmund Burke and Charles James Fox — offered as a case study in moral courage, rhetorical genius, and what it means to hold an unpopular position.

Q2 — How is it argued? Primarily through historical pattern recognition. Grant does not model his claims; he assembles analogies. The AI capital expenditure race is mapped onto the 1990s fibre-optics contest and the 1870s railroad bubble. The debt problem is mapped onto Wilhelm Röpke’s 1950s diagnosis of inflation as the national economy’s sensitive organ. The Bernard Baruch 1929 myth is debunked — Baruch did not sell at the top; he salvaged 60–70% of capital by acting in 1930. The LTCM story (Nobel laureates who failed) is used to puncture the hubris of institutionalised intelligence. For gold, Grant builds a narrative: Fed political subjugation + fiscal excess + reserve-currency erosion = persistent upside, while acknowledging the story could be wrong.

Q3 — Is it true? Grant’s track record is asymmetric: correct on the dotcom bubble (warned in 1999), correct on mortgage securities (2008), correct on post-GFC inflation (warned from 2014, materialised 2021–22); but frequently early and wrong on timing, and persistently negative in a period (2009–2021) when markets delivered exceptional returns. His core insight — that excessive debt, uncorrected bad actors, and inflated valuations eventually produce corrections — is historically grounded. The timing mechanism is unspecified, which is both honest and unhelpful. His characterisation of private equity’s structural troubles is well-sourced (Nate Kobak at the conference). The CAPE comparison (40x vs. 44.2x in 1999) is the most concrete data point; it is real but contested — CAPE critics argue it ignores interest rate differences and accounting changes.

Q4 — What of it? For the investor: the question is not whether to exit markets entirely but whether to reduce leverage, avoid speculative excesses in private equity and overvalued tech, and maintain what Grant calls “thoughtful risk-taking.” His value is not market timing but the quality of the historical lens. The Röpke passage on inflation as monetary disease is genuinely illuminating. The Burke and Fox discussion is not directly investment-relevant but models the rarest quality in Grant’s actual practice: the willingness to hold an unpopular position publicly, for decades, and not to gloat when right.

Glossary

Decadent finance: Grant’s phrase for the phase of the market cycle in which Fed intervention prolongs booms and forestalls corrections, allowing bad conduct to go unpunished. “Instead of credit being man’s confidence in man, it now demands man’s confidence in the sagacity of his lawyer because the ingenuity of the strong and the cunning.” See Decadent Finance.

CAPE (Cyclically Adjusted Price-to-Earnings ratio): market price divided by the 10-year average of inflation-adjusted earnings, smoothed to remove single-year noise. Like the price-to-earnings ratio — “how many years of profit does this cost?” — except it averages across a full business cycle rather than one year. At 40x in October 2025, second only to the 1999 dotcom peak (44.2x); was 38.6x in autumn 2021.

Check-writing contest: Grant’s phrase for an AI capital expenditure race in which participants feel compelled to spend regardless of demand because stopping means falling behind. Amazon, Microsoft, Alphabet, Meta, Oracle, and CoreWeave will spend $382B in capex in 2025 — up 50% from 2024 and triple 2023 — mirroring the 1990s fibre-optics race that ended in overcapacity and collapse.

Democratisation (of private equity): marketing private equity and private credit to retail investors when institutional investors cannot get their capital back. Nate Kobak: “in finance whenever you hear the word democratising, hide your wallet.” Private equity was capitalised for a near-zero interest rate regime; at 8–12% rates, the assets are overcarried and the distributions have not come in as promised.

Röpke’s inflation thesis: Wilhelm Röpke (1899–1966), German economist and post-war architect of the German economic miracle. Writing in the 1950s: inflation is “a dilation of money, so to speak, a managerial disease of the national economy” — the result of excess claims on production: everyone wanting to invest more than savings permit, wages higher than productivity justifies, governments raising claims on an overstretched economy. The money supply expands because no other part of the system will yield. Like an organ consistently abused until it ceases to resist. Grant’s corollary: under a paper money system, the dollar never regains purchasing power lost to inflation.

Reserve currency privilege: the US dollar’s global acceptance means no hard external limit on US borrowing, unlike the gold standard era when gold outflows constrained expansion. This is the strongest counter-argument to Grant’s fiscal concerns. But it has limits: the 2019 repo market crisis and the March 2020 Treasury market dislocation were early symptoms of stress even in the “deepest of all world security markets.”

Mark-to-market (private equity context): in private equity, assets are “marked” — i.e., assigned a book value — periodically by the fund manager rather than continuously by a public market. This produces a smooth, apparently non-volatile return line. Critics argue this is not non-volatility but non-observation — the asset may have fallen in value but the fund has not yet recognised it.

AI bubble: the check-writing contest

Grant’s central market argument focuses on the AI capital expenditure race. The data: $382B capex in 2025 from the Magnificent 7 (Amazon, Microsoft, Alphabet, Meta, Oracle, CoreWeave) — up 50% from 2024, triple 2023 levels. The Magnificent 7 now account for 31% of S&P 500 total capital spending, up from 19% in 2019.

The analogy: 1990s fibre-optics. Companies laid far more cable than demand could absorb, competing for share in a market whose ultimate size no one knew. When demand did not materialise at the pace investment assumed, the companies collapsed. The same logic: “are you going to get paid for it?” Grant notes that college students leave in spring and AI demand falls measurably — “who else has such a persistent need for plagiarism?” The implication: if the largest identified use case is episodic and price-sensitive, the demand foundation is weaker than the capex trajectory implies.

David Rosenthal (Nvidia employee no. 4) presented the same concern technically: the gap between investment scale and demonstrated willingness to pay is the tell, not the technology itself.

Grant’s position: not that AI is worthless, but that the human response to transformative technology is invariably to overshoot. The railroad boom of the 1870s ended in the Panic of 1873. Fibre optics ended in the dotcom crash. “The promise of a marvellous technology with human characteristics” — those characteristics being herding, momentum-chasing, and check-writing contests. “There is going to be a crash first. You’ll be sorry you ever heard the phrase AI.”

Private equity in structural trouble

Private equity was capitalised for a near-zero interest rate regime: debt at 3% rather than 8–12%. As rates normalised post-2021, the interest burden on portfolio companies increased dramatically, assets became harder to exit, and the distributions institutions expected stopped arriving. The 20,000+ private equity companies globally are now trying to find footing at rates they can’t handle.

The result: secondary sales (unplanned exits at a discount), overcarrying of assets at unreasonable book values, and the search for a new investor base. The new investor base is retail — described as “democratisation.” Nate Kobak’s framing: private equity firms have “maximum aggression at periods of maximum risk” — they are not as smart as their reputation; they are momentum-chasers who happened to succeed during a multi-decade bull run for leveraged assets.

The Ponzi framing: Kobak declined to use the word as “unnecessarily brutal for a sophisticated audience.” But the structure — investors expecting returns that depend on new capital coming in — is at minimum fragile. The Yale endowment model (David Swensen) was sold to colleges and institutions on the basis that private equity would provide returns uncorrelated with public markets. The current position: those institutions can’t get their capital back, and their annual draw obligations are not being met.

Decadent finance and the necessity of corrections

Grant’s deepest argument is that markets require periodic corrections to remain healthy. Corrections serve a disciplinary function: they remove bad actors, correct mispriced assets, and teach hard lessons about leverage and hubris. When the Fed intervenes to forestall corrections — through QE, near-zero rates, emergency facilities — bad conduct goes uncorrected. “Mr. Market is the best disciplinarian.”

The 2009–2021 period prolonged the cycle, prevented the market from “skimming the bad actors off the stage,” and produced the conditions for a more severe eventual correction. “These markets run over the sceptical mind.” Grant’s phrase decadent finance names the regime: credit no longer functions as “man’s confidence in man” but requires lawyers and complex documentation because trust cannot be assumed. See Decadent Finance.

The LTCM case: Nobel laureate-run fund. Emanuel Derman (Goldman Sachs quant, wrote My Life as a Quant) was on a call with the LTCM principals after their blow-up. He was startled by the depth of their sophistication — they asked better questions about their own failure than the Goldman traders trying to value the assets. The lesson: sheer mental power is not invariably the road to riches. This extends Grant’s broader argument about the Fed’s hundred-plus PhD economists who failed to predict that stimulus + QE would produce inflation and a spending spree.

Röpke on inflation: the sensitive organ

Grant invokes Wilhelm Röpke extensively on inflation. The key passage (Röpke, 1950s): “people want to invest more than savings permit. They demand wages higher than the growth of productivity justifies. They want more imports than exports can earn. And above all, the government, which should know better, raises its claims on this overstretched economy higher and higher. Thus, there is a riot of claims and an insufficiency of goods produced to meet them.”

And on money: “Just as there are organs in the human body in which, if consistently abused, ailments slowly but surely accumulate, eventually taking their revenge, so the national economy has its own equally sensitive organ. That organ is money. It becomes feeble and ceases to resist. And it is this infeeblement which we call inflation, a dilation of money, so to speak, a managerial disease of the national economy.”

Grant’s corollary: under a paper money regime, the dollar never regains purchasing power lost to inflation. This distinguishes the fiat era from the gold standard era, when prices fell as well as rose. One reason Trump’s 2024 election victory made sense to Grant: people saw prices not rates of price change. “I don’t care what they’re saying. Look at this. Look at the price of eggs.”

US fiscal context: taking 222 years to borrow the first $16T; then achieving the same in 8 years under Biden and Trump. Running a 6–7% GDP deficit at 4% unemployment — what would historically be defined as full employment and roaring markets. “That doesn’t sound like a well-managed public finance operation.”

Gold

Grant bought his first Krugerrand in January 1980 at approximately $850. Gold is now approaching $4,000. He wears this as “a badge of constructive humiliation” — not a prediction he got right but a reminder of the hazards of narrative-building. The “permanent high plateau” argument he heard in 2011 (S&P downgrade of Treasuries from AAA to AA+) proved wrong for the next decade; gold collapsed from $1,900 to $1,200.

His current narrative: (1) the administration’s attempt to subjugate the Fed — interest rate suppression + dollar weakening; (2) fiscal recklessness without reform. A prominent speculator told him at the conference: “Of course it’s a bubble” — but a bubble can be justified by fundamentals. Grant: “There’s no certitude. There can’t be. People think they have it — you know two things about them: they’re not very old, and they have not had the invaluable experience of having their face ripped off during a bear market.”

Edmund Burke and Charles James Fox

The second half of the episode covers Grant’s book Friends Until the End — a double biography of Burke (1729–1797) and Fox (1749–1806). Brief investor-relevant connections:

The East India Company analogy: Grant draws the parallel explicitly: the Company was as well-hated and well-envied as any modern technology giant. It had its own army and navy, its own monopoly charter, and its agents were paid poorly but allowed side-hustles that became their main business. The incentive result: they enriched themselves rather than their employer, pillaging India. Buffett and Munger’s portraits hang on the wall as Grant tells this story — he describes Charlie Munger’s dictum on incentives without naming it. The Company agents were the original insiders who extracted value from the system rather than creating it.

Moral courage: Burke spent political capital defending the rights of people who could do nothing for him — two men convicted of homosexuality who were stoned in the stocks (one killed), an unknown destitute poet (George Crabbe) who appeared at his door — while prosecuting the powerful (Warren Hastings) over eight years at immense personal cost.

The friendship’s end: Burke and Fox parted irrevocably over the French Revolution. Fox saw it as the greatest moment in human history — the overthrow of tyranny. Burke saw chaos, destruction of civil order, and predicted the Terror. Fox was excited to be bullish; the Terror proved him wrong; but he clung to his position. The friendship ended in the House of Commons — Burke went “incandescent” when Fox quoted his own words back at him during a debate. The deepest lesson Grant draws: a great investor needs the rarest quality Burke embodied — holding an unpopular position for decades without capitulating, and not gloating when eventually right.