Half Baked Thoughts: Rethinking Search Funds in an AI-Disrupted World

I've spent the last two weeks building AI agents for deal sourcing, scraping databases, finding owners, gathering contact information. The experience has been clarifying in ways I didn't expect, and it's fundamentally changed how I'm thinking about search funds and private equity investing. Let me be direct: I'm at a seven out of ten conviction that we're facing a structural challenge in how we evaluate long-term equity value. This post is me thinking out loud about a problem I don't have fully solved yet. *The Core Problem: We're Making a Seven-Year Bet A traditional search fund is a two-year search plus a five-year hold. That's a seven-year bet on equity value. In venture capital, this uncertainty is already showing up. Even if a business model isn't disrupted today, there's dramatically less confidence it won't be at risk in five years. Valuations reflect this. I believe this same uncertainty is coming to private equity markets. Here's why: if globalization destroyed America's working class, AI will destroy the middle class. We're already seeing this in white-collar hiring. Recent graduates aren't getting job offers, first-year analysts can be replaced. If I can replace my first-year investment banking analyst, how do I build equity value around a middle-class economy that may not exist in its current form? *From Thinking Economy to Physical Economy We moved from a muscle economy (industrial revolution) to a thinking economy (knowledge work). AI is replacing the thinking. So what's left that creates value? Two things: capital and scarcity. Capital will become scarce because AI infrastructure is capital-intensive. But that doesn't help us buy businesses. It just means the supply of labor (even thinking labor) goes up while its value goes down. What remains valuable is what's scarce: physical goods, land, locations, commodities. This leads me toward businesses that deal with the physical world rather than the intellectual world. The more physical something is, the more archaic the process, the better. These aren't classic high-margin search fund businesses, but they may be the ones that survive. *Revolution vs. Evolution In my battery business, we used to say: "Batteries don't go through revolutions, they go through evolutions." Lithium-ion batteries were invented in the 1980s but took###-###-#### years to reach cars. Physical constraints—safety, testing, regulation—created slow adoption curves. That's what I'm looking for now: businesses that move slowly. In a world accelerating toward uncertainty, where do things change slowest? My answer: government-facing businesses. I'm looking at environmental testing required by government mandate, businesses selling to municipalities, anything where regulation creates friction. Even if an AI-native solution exists, getting government to adopt it takes forever. Government will be busy dealing with rapid change elsewhere. They won't have bandwidth to modernize slower-moving regulated activities. The irony: as a businessman, I hate highly regulated businesses. But regulation makes things go slower, and in this world, slower is safer. *The Discount Rate Problem Here's the financial framework shift: we typically use a constant discount rate in our DCF models—20%, 30%, whatever. But I think we need to model discount rates that increase over time. The further out you go, the higher the discount rate should be. Maybe it approaches 90% for cash flows seven years out. This has profound implications: it means you should value initial free cash flow dramatically more than exit value. I ran my business for eight years without seeing a dividend. All my value was on paper until exit. If I were searching today, I'd structure it differently. I'd look for businesses where I could dividend recap early. Higher leverage, lower multiples, stable cash flow. I'd rather experience $100K per year in profit sharing for five years than wait for a big exit that might not materialize. Even if it means less total equity at exit, money now is worth dramatically more than uncertain money later. *The Offshore Staffing Example I built an offshore staffing operation within my company and was certain I'd buy or build an offshore staffing business after my exit. Then I started talking to companies in the space. Their valuations dropped from 8x multiples to 4x multiples. But here's what's interesting: they still have free cash flow. It's going to take time for AI to replace them on a cash flow basis, even as their equity value evaporates. This is the disconnect: businesses can still generate cash while their exit multiples collapse. This is why I'm shifting focus from equity value at exit to free cash flow generation throughout the hold period. *Two Possible Strategies I see two paths forward, and I don't know which is right: 1. Buy the disruption: Find businesses that will be disrupted by AI and use AI to accelerate that disruption yourself. Ride the tsunami wave while it's building. The challenge: knowing when to exit before it crashes. High risk, potentially high reward, shorter time horizon required. 2. Buy the slow movers: Focus on physical-world businesses, government-facing companies, highly regulated industries. Accept lower growth and lower multiples in exchange for predictability. Structure for early cash extraction rather than exit value. My risk tolerance points me toward option two, but I can see the case for option one if you have the timing right and a clear exit strategy. *The Uncomfortable Questions Even my old battery business, fundamentally a physical manufacturing operation, won new customers through technical expertise. Engineers called us because we could design solutions. What happens when engineers can prompt ChatGPT and get 80% of the answer? Our differentiation was the thinking layer on top of the physical product. You have to be careful even with "physical" businesses. Where is the actual value creation happening? If it's in knowledge, skill, or expertise, you're at risk. *Where I'm At I'm at conviction level seven out of ten on this thesis. I might be wrong. In a few months, I might think differently. But right now, this is keeping me up at night. The search fund model worked brilliantly in a world where you could make seven-year bets with confidence. I'm asking myself: does it still work when the seven-year time horizon is fundamentally uncertain? How is everyone else in the community thinking about this? Am I overthinking it, or are others wrestling with the same questions?