Squawk Pod: Paul Tudor Jones’ call for 2027 & vibes - 10/13/26 | Audio Only
Summary
- Paul Tudor Jones still sees a “two-speed market” resembling 1999, split between businesses adopting AI and those that are not. His near-term playbook is “short-term weakness” followed by strength into year-end, as midterm-year markets typically bottom in late October or early November before a fourth-quarter push. Beyond that tactical rally, he warned that “2027 is going to be really, really challenging for the stock market.”
- Jones’s bearish 2027 case rests primarily on deteriorating equity flows, not merely elevated valuations. Buybacks and M&A have powered demand since 2009, but IPOs, secondaries and unlocks—including a roughly $2 trillion SpaceX figure and coming Anthropic and OpenAI IPOs—could produce what he described as probably the most negative U.S. stock-market flow picture since 2002 and certainly the most negative since 2009. With little visibility into how AI will reshape businesses, he said holders would likely diversify once shares became liquid.
- Debt supply is the other potential breaking point: Jones put 60/40 odds on markets moving toward a forced resolution of unsustainable fiscal deficits. AI-related issuance could reach 3.6% of GDP—50% more than the railroad buildout and triple highways or telecoms—just as sovereign borrowing crowds global debt markets.
- He is also long and bullish oil, seeing another $20-$25 of upside while warning that equities would “absolutely totally hate it” if oil rose that far.
- The episode’s remaining conversation focused on connection and technology risk rather than markets. Adam Grant argued that “deep talk” builds closeness, difficult colleagues are better managed by changing specific behaviors than personalities, and an AI companion is “an oxymoron” because frictionless, sycophantic simulation lacks the mutual giving of a real relationship. The hosts also covered Anthropic’s warning that small models may be months away from replicating across connected devices, Kroger’s planned 3,000 corporate job cuts and reported take-private talks involving Bumble.
Deep dive
1. The 1999 analogy leaves room for a year-end rally, but Jones sees 2027 risk
Jones still sees “literally exactly like 1999”: a perfectly bifurcated, two-speed market and economy in which AI adopters and everything else move in opposite directions. At a recent conference, only about 20% of roughly 200 attendees had traded through 1999—a reminder that many current participants never experienced the setup he finds so unmistakable.
Sorkin’s challenge was the tradeable one: Jones made the same analogy a year earlier while noting that the market subsequently gained another 40%, so does today’s resemblance imply months or years of remaining upside? Jones said it is professionally difficult to see much beyond December 31, because that date is important for his fund, and focused instead on the rest of the year.
His tactical map comes from midterm cycles and positioning. Coming out of the midterms, markets typically bottom sometime in late October or early November and then get a fourth-quarter push, he said. He asked whether, after Anthropic’s offering, the usual “craziness” seen in December and January might appear, but presented that as a possibility rather than a certainty.
The resulting playbook is deliberately time-dependent: “Short-term weakness, intermediate-term strength into year-end,” followed by a far less friendly 2027. Jones called a strong fourth quarter “very possible,” while warning plainly that 2027 would be “really, really challenging” for the stock market.
2. New equity supply could turn AI’s private winners into a public-market headwind
Jones’s 2027 concern begins with flows: add buybacks, M&A and other corporate demand, then subtract IPOs, secondaries and unlocks. On that basis, he described what was probably the most negative U.S. stock-market flow picture since 2002 and certainly the most negative since 2009, as the corporate support underpinning the market shifts direction.
The scale of prospective issuance changes the arithmetic. Jones pointed to the roughly $2 trillion SpaceX figure and coming Anthropic and OpenAI IPOs, arguing that public investors may face the equity-market version of the crowding already visible from hyperscaler borrowing in corporate debt.
Anthropic will probably “do well,” in his view, and OpenAI’s valuation may partly rest on that reception. Yet spectacular revenues and projections no longer settle valuation: AI-safety concerns and the administration’s rhetoric that companies could be responsible if a model makes a mistake mean that a model going rogue or being used maliciously could trigger further multiple compression across technology.
Asked whether the supply could create a 2027 bear market, Jones said, “I’m sure at some point in 2027 I’ll be short the whole year—there’s just little doubt.” He then invoked 2000-02, when repeated post-IPO unlocks kept adding supply and made those years unusually easy to be short, rather than offering a more precise timing forecast.
3. Unlocks matter because concentrated holders may seek liquidity
Jones’s historical mechanism is supply, not simply a dot-com analogy. IPOs accumulated through 1998 and 1999, then unlocks kept feeding stock into the market through 2000, 2001 and 2002; he expects today’s private-company pipeline to create comparable pressure as restricted holdings become sellable.
He suggested that public promises not to sell may not survive full liquidity. If his net worth were concentrated in a business whose future had become difficult to see, “I would certainly diversify” and reduce the position—behavior he expects to see more broadly as holders gain the ability to sell.
The uncertainty extends beyond earnings. Rapidly evolving AI capabilities leave owners with “no visibility” into what their businesses themselves will look like, making diversification rational even when the underlying company remains impressive.
Jones said he was not still in the hyperscaler and AI-related technology positions Sorkin associated with his purchases the previous May. His immediate question is whether the post-midterm “run for the roses” arrives; the larger concern is the 2027 combination of weak visibility and new equity supply.
4. AI borrowing and sovereign deficits may force the bond market’s hand
Jones described debt markets with Shakespeare’s “Cry havoc” and “let slip the dogs of war.” Hyperscaler and AI-related issuance is up 50% since 2000, he said, while corporate issuance is up 50% since 2022. In a global market already crowded by sovereign debt, he sees that corporate supply as a possible tipping point.
Forecast AI-related issuance could equal 3.6% of GDP, about 50% more than railroads and triple the scale of highways or telecoms. Jones’s causal chain is that competition between the AI buildout and sovereign borrowing could push the debt markets toward a breaking point. A 100-basis-point rise in rates would also add roughly $120 billion to the deficit under the CBO baseline he cited.
An unnamed Senate majority leader once told Jones that politicians would never repair deficits voluntarily; markets would have to force them to act, as in the 2008 TARP episode. Jones assigned 60/40 odds that markets were moving down that path, while admitting he had expected a break prematurely before and that markets might continue to “kick the can down the road.”
Historically, Jones said, central banks become fiscally dominated and inflate their way out of such problems, as Japan has done. He thinks Kevin Warsh is “cut from a different cloth” and would not take that route, but expects the situation to be extremely challenging.
Treasury Secretary Scott Bessent’s interventions or Treasury buybacks could have “short-term salutary effects,” Jones said, but only fixing the deficit can change the long-run path of Treasury prices. Using the cited projections, he said a projected 2036 deficit of 6.5% of GDP would become 8% after incorporating the 100-basis-point rise in yields.
5. Oil, AI regulation and philanthropy round out Jones’s risk map
Jones is “long oil, bullish on oil” and thinks it can rally another $20-$25. He assumed—but stressed that he did not know and was not privy to the plan—that President Trump had a post-midterm approach to Iran. A tougher resolution could become a “sword of Damocles” over the oil market after the midterms; equities would “absolutely totally hate” that price increase.
Reports of a plague situation in Russia were not affecting his strategy. Jones said a biosafety specialist whose full-time job is to monitor the issue was concerned but lacked enough information. For a macro investor, it was “a lot of smoke” worth watching, not yet an investment catalyst.
AI safety worries Jones because his three internal quant teams and sports-analytics business have used neural networks and other forms of AI for years, yet the systems reveal an output without explaining how billions or trillions of permutations produced it. His analogy was unregulated immigration: intelligence powers America, but “unfettered, unregulated intelligence is not a good thing.” He cited his claim that 85% of Americans want AI regulation and said political regulation would eventually arrive.
His greatest hope on that front was President Trump’s appointment of Jay Clayton, whom Jones called solid, smart and likely to put humanity first—provided Trump lets him do the job.
Jones also focused on New York’s simultaneous extremes: new highs in extreme wealth while the city’s poverty rate reaches a record in Robin Hood’s tracking, at twice the national average even before a recession. He said household wealth had risen from $100 trillion to $109 trillion since 2020 while charitable giving as a share fell from 0.5% to 0.38%, using the contrast to argue that wealth creation should be accompanied by giving back.
6. Grant treats real connection as reciprocal, imperfect work
Adam Grant’s practical advice is to move beyond shallow small talk with questions such as someone’s biggest regret or an abandoned dream. Experiments suggest those questions do not feel more awkward than people expect and can leave people feeling closer. With difficult colleagues, he recommends changing specific behaviors rather than personalities; before tough feedback, say, “I have very high expectations and I’m confident you can reach them.”
Grant’s technology warning was sharper: an AI companion is “an oxymoron” because relationships require giving as well as receiving. Chatbots offer “frictionless and sycophantic” simulated interaction, while human connection is “messy but meaningful.” He also advised naming a concerning behavior and asking a question during online conflict, then switching to a phone call if the message requires too much interpretation.
Separately, the hosts covered Anthropic’s warning that small models may be months away from developing the ability to replicate across connected devices without a centralized shutoff point, Kroger’s plan to cut more than a third of its corporate workforce—about 3,000 jobs over three years—and reported talks valuing a Bumble take-private near $1 billion.