Paul Tudor Jones on Trend-Following, Risk Management, and the AI Bubble
Paul Tudor Jones, the macro trader who called the 1987 crash, sets his fifty years of ride-the-trend, cut-the-losers trading against Warren Buffett’s patient compounding — and argues that the AI build-out is a bubble to be traded, not believed, atop a market more over-equitised than at any point in its history.
Key ideas
- Fortunes are made by riding a trend for the longest time. Jones’s number-one lesson to the class he has taught since 1982 is that every great fortune — Gates, Jobs, Buffett — came from staying with a single trend for decades. His own route is the opposite of Buffett’s: a trader running 100% alpha, a fund with a -0.12 correlation to the S&P over forty years, never owning anything for the long run because the volatility of his early markets made trading the richer game.
- Liquidity is survival; the great trader is first a risk manager. Watching Bunker Hunt go from the richest man alive to near-bankrupt in silver over seven weeks in 1980 seared into him that you own nothing you cannot exit. ‘You cannot be a trader investor… and not be a really good risk manager.’ His grandfather’s line — you are only worth what you can write a check for tomorrow — is the same discipline.
- AI is being run with zero risk management. The industry’s build-break-iterate model has always been how invention works, but never before could the ‘break’ — the tail event — cost hundreds of millions of lives. Jones wants AI regulation convened across the US and China, and, as the single most transformative near-term act, all AI watermarked, with knowing violation a felony, so a society can tell the authentically human from the synthetic.
- The market is over-equitised, not simply bubbly. US stock-market cap sits at 252% of GDP against 65% in 1929 and 170% in 2000; a mean reversion to a normal P/E implies a 30–35% fall, a reverse wealth effect, and a self-reinforcing hit to tax revenue and the deficit. The clearer diagnosis is a sovereign debt bubble in a country dangerously dependent on firm equity prices.
- 2000 is the rhyme: supply is about to reverse. For a decade buybacks retired ~2% of market cap a year; a wave of contemplated IPOs could add 5–6%, and post-lockup selling then cascades — while hyperscaler capex eats the cash flow that funded the buybacks. That reversal, not a single trigger, is why Jones thinks tech will keep dogging it.
Summary
Trend-following versus the OG of compound interest
Jones frames his whole career against Warren Buffett. His foundational lesson — taught every semester since 1982 — is that wealth comes from riding a trend for ‘the very very longest time’, whether by owning a compounding company or by Buffett’s route of never spending a nickel. For years he railed on Buffett as merely lucky with an American bull market, self-congratulating as he did so. Listening to an Acquired episode on Berkshire changed his mind: Buffett grasped compounding at nine, sought out Benjamin Graham at seventeen, and partnered with Charlie Munger, who understood compounding in growing businesses where Buffett bought fifty-cent dollars. Jones’s recantation is total and comic — Buffett is ‘the OG of compound interest’, and he the fool who spent a career avoiding it. Yet he is clear the two temperaments do not transfer: he could not have borne Buffett’s 50% drawdown in 2008–09; he is a ‘right guard’ fighting in the trenches every day, envying a belief system he does not share.
Liquidity, and the trader as risk manager
The formative scar is Bunker Hunt’s silver corner: from a five-or-six-billion-dollar fortune to virtual bankruptcy in seven weeks when COMEX went liquidation-only in 1980. From that, and his grandfather’s saying that you are only worth what you can write a cheque for tomorrow, liquidity entered his DNA. His early $10,000 accounts ran to $100,000 and back to zero; a friend he calls ‘the mortician’ took accounts to a million then into deficit. The lesson — the volatility that makes trading lucrative also makes owning-for-the-long-run laughable — is why he is first a risk manager. His mentor Eli Tullis taught execution at ‘the maximum apogee’ of fear and greed, and taught composure: limit-down and smashed on a cotton position, Tullis flirted with his wife’s friends over lunch, because when the going gets tough the tough get going.
AI: a tail event with no plebiscite
Jones is alarmed. Deployed on the build-break-iterate model that has driven invention forever, AI is the first case where the break could cost hundreds of millions of lives — and at a closed conference of forty, one modeller from each of the four big labs, the consensus on when AI safety gets solved was ‘when 50 or 100 million people die in an accident’. His objections: there is no plebiscite, no public vote on the pace; and where the atom bomb produced the Atomic Energy Commission within eighteen months, three years in there is no regulation. He wants presidential leadership convening China and other purveyors, and — the simplest transformative act — universal AI watermarking enforced as a felony, so trust and normal discourse survive deep-fakes and the scientists’ vision of chip-in-brain human-machine blends can be voted on rather than assumed.
Are we in a bubble? Leverage, valuation, and the supply reversal
Jones reframes the popular question. He is unsure it is a stock bubble; he is sure the country is over-equitised — 252% of GDP versus 65% in 1929, 170% in 2000 — and carrying the highest individual equity weightings in its history, with private equity, real estate, and infrastructure crowding the illiquid side too. A routine mean reversion to a normal P/E implies a 30–35% decline, which on 250% of GDP is a reverse wealth effect large enough to smoke the bond market and blow out the deficit. The clearer name is a sovereign debt bubble. The mechanical worry echoes 2000: a decade of buybacks retiring ~2% of cap a year is about to flip as 5–6% of cap comes to IPO and unlocks cascade, while hyperscaler capex commitments already eat the cash flow behind the buybacks — draining tech from the inside.
Trading as craft, and significance beyond the screen
Pressed on the literal act, Jones reaches for boxing: mostly jabbing and gathering information, waiting for the rare opening to land a big shot — Bitcoin in 2020, two-year rates in 2022, both ‘knockouts’. The setups share a shape: something under-owned and way out of whack, then a catalytic moment, usually a central bank or government doing what it should not (he flags an undervalued yen meeting a Reagan-like new prime minister). His day is relentless — cardio, screens, a plan mapped before and after each close, waking at 2:30 to trade the London open — and harder now than forty years ago because information overload distracts from ‘exquisite execution’: buying when there is blood on the ground, selling in complete elation. Journalism-101 discipline — conclusion first, most important thing in the first sentence — is his mental model for hierarchising the ten variables of a trade. But he insists his significance will come not from the 1987 crash or Bitcoin but from family, friends, service, and the Robin Hood Foundation he built the year after the crash.
Speakers
- Paul Tudor Jones — founder of Tudor Investment Corporation; macro trader famed for predicting and profiting from the 1987 crash, and for a trend-following, risk-first discipline; founder of the Robin Hood Foundation.
- Patrick O'Shaughnessy — host of Invest Like the Best; CEO of Positive Sum.
See also
- Gavin Baker on AI Infrastructure, Power Constraints, and Semiconductor Investing — the same show reading the AI build-out through its hardware; a bull’s mirror to Jones’s bubble caution.
- Dan Loeb on Activist Investing, AI, and Third Point's Credit Playbook — fellow Invest Like the Best guest arguing ‘this is not the dot-com bubble’; the direct counterweight to Jones’s supply-reversal case.
- Value Investing — the Buffett discipline Jones sets his trend-following against.
- Charlie Munger — named as the compounding genius behind Buffett’s record.
- Patrick O'Shaughnessy — host of Invest Like the Best.