2027 will be 'really challenging' for stocks

Robin Hood Foundation has supported Good Shepherd Services, which serves 30,000 children in New York City, for over three decades.

Robin Hood's founding principle anchors Jones's thinking about both opportunity and risk. The foundation operates on the premise that environment and background alone do not seal a child's destiny, a conviction that extends to how he reads markets—systems can shift, outcomes depend on action, not just initial conditions.

This philosophy shapes his skepticism of market consensus. If structural circumstances—like the current glut of AI capital—appear fixed, Jones looks for the inflection point where flows reverse and the narrative breaks.

The stock market today splits sharply into two groups: companies adopting AI and those that are not, and they are moving in opposite directions. This pattern mirrors 1999, when the economy similarly divided into dot-com winners and traditional laggards, with very different trajectories.

What makes the current moment more dangerous, in Jones's view, is that most market participants have no lived experience of the cycle that followed. At a conference he attended recently, he asked 200 traders how many were actively trading in 1999. Only about 20 raised their hands.

That generational gap matters. Traders who never felt the pain of the 2000–2002 crash have no embedded fear of binary outcomes. The current path may look great for the next year or two, but the long-term risk is invisible to most of the market.

Historical patterns from midterm election cycles suggest the fourth quarter should be strong. The market typically finds a bottom in late October or early November, then experiences a seasonal rally into year-end. Jones is thinking tactically about positioning through December 31st, assuming this playbook holds.

This near-term outlook contrasts sharply with his 2027 forecast. The year ahead offers one more window to ride the existing flows, but that window closes when the calendar flips.

The reason Jones calls 2027 "really, really challenging" is the flow math. Since 2009, stock buybacks and mergers have continuously added money to equities. But when he nets those inflows against IPOs, secondary offerings, and stock unlocks, the picture is the worst since 2009—and possibly since 2002. The positive cycle that lifted the market for a decade and a half is ending.

Coming IPOs from massive AI companies like Anthropic and OpenAI, combined with a $2 trillion SpaceX valuation already in the system, will accelerate this shift. Huge blocks of venture-backed wealth will seek exits and diversification. At the same time, fewer traditional companies are left to provide the steady buyback support that masked underlying weakness.

The math was easy when one thing—corporate support—dominated. Now multiple large forces are working against it.

Multiple compression in the tech sector reflects genuine worry about AI safety. Concerns about models going rogue or malicious actors weaponizing AI are real. But Jones points to a new pressure: regulatory threat. The Administration's rhetoric makes clear that companies will be held responsible if they make mistakes with AI deployment. That threat has prompted boards to reassess risk and cost.

Anthropic is well-positioned to benefit from these concerns—it was founded explicitly on safety principles and may be seen as a solution to the regulatory problem. OpenAI's performance, by contrast, will rest heavily on what Anthropic demonstrates: whether safety and scale can coexist. If Anthropic stumbles, OpenAI's valuation could face additional pressure.

Jones runs three quantitative research teams and a sports analytics business that have used neural networks and other AI forms for years. His concern does not stem from inexperience; it comes from it. The core problem remains unchanged: you know what the network outputs, but you cannot trace why or how it arrived at that answer.

With billions and trillions of permutations flowing through the system, the path to any given prediction is opaque. No amount of scale or accuracy changes that fact. A system that works 99% of the time but fails catastrophically 1% of the time on a mission-critical decision—and you cannot explain why it failed—is a systemic risk. That opacity is why regulators are right to worry and why boards are right to reassess.