I've spent more than two decades building, investing in, and advising businesses. For most of that time, the conventional wisdom was brutally simple: if your company touched physical things, moved slowly, or required human beings to deliver value, you were a dinosaur. The future belonged to software. Recurring revenue. Network effects. Infinite scalability.
That thesis just took a bullet from the people who manage nearly a trillion dollars.
Apollo's deputy global head of private equity, Antoine Munfakh, told CNBC this week that the industry's massive concentration in software represents "a systemic failure of risk management across the asset class." Software went from roughly 10% of global buyout volumes to 40%. Apollo's response? Zero exposure to growth software in its last two private equity vintages.
Their alternative has a name: HALO. Hard Assets, Low Obsolescence. Real-economy businesses in services, industrials, logistics, media, and business services where AI functions as a value creation lever, not an extinction-level event.
I think this is one of the more significant shifts in institutional thinking I've seen in a long time, and I think most founders and SMB owners are going to miss what it actually means for them.
The Reversal Nobody Saw Coming
For five years, every business owner I've talked to has heard some version of the same pitch: you need to become "tech-enabled." Slap an API on it. Build a dashboard. Get your SaaS multiples up. The implicit message was clear: if your revenue depends on people doing work in the physical world, you're the wrong kind of company.
Software revenue multiples have now collapsed roughly 38% in six months. Apollo's own research estimates that 85% of all technology deals from 2016 to 2025 were software-related, and exits haven't kept pace with deployment. The industry spent a decade writing checks to software companies at valuations that assumed perpetual growth, infinite retention, and near-zero competition. That last assumption is aging poorly (ahem, AI).
AI didn't create the problem, but it has accelerated a reckoning that leverage, crowding, and disciplinary laziness had been building toward for years. When barriers to entry drop, when a competitor can replicate your core product feature with a well-designed prompt, the moat you thought you had turns out to be a puddle.
What HALO Actually Tells You
Here's what I find interesting about Apollo's framing. They're not saying "avoid technology." They closed a $35 billion AI chip financing deal with Anthropic and Broadcom the same week Munfakh made these comments. They're saying the businesses that will benefit most from AI are the ones that were never in danger of being replaced by it.
A logistics company with 200 trucks and 15 years of route optimization data gains massive, massive leverage. AI doesn't threaten it. The trucks still have to roll. The warehouses still have to be staffed. But every dispatch decision, every maintenance cycle, every demand forecast gets a lot sharper.
Compare that to a mid-market SaaS company selling workflow automation to SMBs. The core product does something a foundation model can now approximate and the threat is only growing. The switching costs that justified the valuation are eroding. The "recurring" in recurring revenue starts to look less certain (we'll save my beef with people's obsession with Annual Recurring Revenue, ARR, for another day).
I think Apollo's framework, whether they intended it this way or not, is a diagnostic tool for any business owner thinking about what AI means for their company's value. Three questions worth asking yourself honestly:
- Does your revenue depend on a human relationship that a customer would actually miss?
- Can a competitor replicate the core of what you deliver with a well-crafted prompt and a good data pipeline?
- Would your customers notice if you disappeared for a week?
If the answers are yes, no, and absolutely, you're probably sitting on a HALO business whether you knew the term or not. If the answers go the other way, I'd start thinking hard about what your moat actually looks like in 2027. It's not the end of the world, yet, but start looking for tweeks.
The Boring-Business Revenge Tour
I spent a chunk of my early career surrounded by people who treated service businesses, distribution companies, and anything involving physical infrastructure as consolation prizes. The real money, the real prestige, was in software. Build something once, sell it forever, watch the margins expand to infinity.
The irony is rich. The businesses that got dismissed as boring, the ones with real customers who pick up the phone, real assets that depreciate on a schedule you can model, and margins that depend on execution rather than network effects, those are exactly what's getting bought now. Apollo put it plainly: "the enduring winners may not be AI businesses, but rather businesses in the traditional economy that harness AI to drive durable efficiency gains."
I've been saying a version of this to founders for a while now. In the investment space, as market leaders start to get frothy, I'm looking to broad equal-weight index ETFs like Invesco's S&P 500 Equal Weight ETF (symbol: RSP) as long-term winners, and at least a counter-balance to chip and infrastructure stocks.
The question was never "are you a tech company?" The question was always "do you have something AI can make better without making you replaceable?" Most businesses I work with do. They just never got credit for it because they couldn't put "SaaS" in an investor deck.
That credit is arriving. Late, and from an unexpected direction, but it's arriving.
Keep building,
-- JW
