From Soros to Old Farm: How to Identify the Market’s Top Thematic Risk-Takers | Kieran Cavanna | Old Farm Partners

Episode Summary

Executive Summary: Kieran Kavanaugh argues that hedge funds still matter because they can pivot quickly, preserve capital, and exploit a narrow set of high-conviction themes—especially AI capex, power, infrastructure, and defense tech. He stresses manager selection based on passion, pattern recognition, and risk management rather than pedigree or benchmarks, and says co-invests are most compelling when managers have domain expertise and asymmetric upside.

Main Topics: How Kieran sources and underwrites managers (Priority: 5/5): Kavanaugh says he looks for passion, hard work, smart thinking, and the ability to anticipate change. He tracks managers over time to distinguish luck from repeatable skill, but notes he can spot exceptional talent quickly when references and pattern recognition align. Risk management as the first filter (Priority: 5/5): He repeatedly emphasizes that great risk managers come before great risk takers. In volatile regimes, he wants managers who can protect capital in bad markets and exploit opportunities when conditions improve. Why co-investing is central to Old Farm (Priority: 5/5): Kavanaugh explains that co-invests work best with intellectually curious managers who want to pursue a big, asymmetric idea. He prefers ideas with low downside and major upside over generic fund exposure. AI capex as the main current theme (Priority: 5/5): He says the central trade is AI capex and related infrastructure: power, uranium, optical networking, memory, semi-cap equipment, and data centers. He views the cycle as durable over the next several years but warns against overconcentration. Global and sector rotation opportunities (Priority: 4/5): Kavanaugh highlights opportunities outside the U.S. in Europe, Latin America, and selected emerging markets, while noting that local benchmarks can be misleading. He prefers bottom-up manager selection over top-down macro calls. The role of multi-manager and hedge fund structures (Priority: 4/5): He argues that multi-strats and hedge funds still have a strong role because incentives, drawdown discipline, and rapid repositioning matter. He also discusses fee compression, but says top managers can still justify premium pricing through performance. Client base, scale, and emerging manager selection (Priority: 4/5): Old Farm serves private wealth, pensions, family offices, and consultants. Kavanaugh likes emerging managers with strong references and meaningful personal capital at risk, but says pedigree alone is overvalued.

Key Arguments: Great hedge fund managers are identified less by pedigree and more by passion, work ethic, and the ability to recognize regime shifts before consensus does. Tracking managers over time is necessary, but decisive underwriting can sometimes happen quickly if references and idea quality are strong. Co-investing is most attractive when a manager has a concentrated, high-conviction idea with meaningful asymmetric upside and real domain expertise. AI capex is the dominant market theme right now, but exposure should be diversified across power, infrastructure, optical, uranium, memory, and defense tech rather than concentrated in one stock. The next several years still favor AI-related capital spending, but the trade has real downside if markets or policy shift abruptly. Hedge funds are valuable because they can react fast, go long and short, and protect capital in difficult markets; they are not benchmark-constrained. Multi-strategy firms have solved parts of the hedge fund problem through incentives and risk controls, but competition for capital remains intense and returns are harder for mid-tier players. Benchmarks matter only insofar as managers are transparent about what they actually do; style drift can be acceptable if it is deliberate and communicated. Personal capital and references from people who know a manager well are stronger signals than a famous prior employer or a large startup raise. There are substantial opportunities in non-U.S. markets and sectors, but they require the right manager and an understanding that local benchmarks may be poor proxies for opportunity.

Data Points: Old Farm Partners AUM: about $700 million - Kavanaugh says Old Farm manages roughly this amount today. Largest historical drawdown: 11% peak-to-trough - He says the firm’s worst drawdown over the last 10 years was about 11%. Prior-year performance: 21.5% - He cites Old Farm’s return last year as 21.5%. Optical strategy performance: 250% to 300% in one year - He says their optical networking thematic strategy had this approximate one-year gain. Single co-invest size growth: $20 million to over $100 million - A co-invest idea in optical networking grew from about $20 million to over $100 million in P&L/size over a year. AI-related stock moves: up 700% in some pockets - He warns that parts of the AI capex trade have seen extreme excess and huge gains. Multi-strat exposure move: 10 net short to 150 net long - He describes one macro manager’s range of positioning as an example of active risk-taking. Multi-strat fee example: 2 and 20; 1 and 10 - He notes some managers charge traditional 2/20 fees, while others are on 1/10 and perform very well. Private credit window: last 7-8 years - He says private credit has pulled capital away from hedge funds over roughly this period. Fund-to-fund and co-invest mix: 30% co-investments - He says co-investments are about 30% of their overall book, with the majority in hedge funds.

Pivotal Quotes: "Make sure the main thing is the main thing." — Kieran Kavanaugh: His central framework for current portfolio construction; he says AI capex is the main thing right now. "I want to make as much money as I can on a year. And when things are tougher, not lose a lot of money." — Kieran Kavanaugh: He explains Old Farm’s objective as absolute return with capital preservation, not benchmark tracking. "What I find is that the intellectually curious love doing these co-invest in the public market." — Kieran Kavanaugh: He describes the type of manager most likely to be a good co-invest partner.

Implications: Listeners should expect continued opportunity in AI-linked infrastructure and selective global themes, but success depends on fast, disciplined manager selection, active risk control, and avoiding benchmark-driven thinking.