Is AI Lighting Money on Fire? Protein Bars Say No.

Anyone looking at the AI landscape right now could conclude that the Startup ecosystem has gone crazy. Many startups have no obvious moats and are raising at 50x revenue. Acqui-hires are happening that create generational wealth with no strings attached. We’re seeing billion-dollar rounds come together for companies that might not exist ten years from now.

The case that too much money is being poured into the system is easy to make. And it’s almost certainly true.

But there is a counterargument that’s worth thinking through. One playbook could justify the frenzy is sitting in your pantry.

Laugh if you’d like, but the consumer packaged goods industry might be the best analogy for understanding AI’s current moment. Here’s why:

Both industries are dominated by a handful of kingmakers. In CPG, it’s P&G, Unilever, Nestlé, Mars, PepsiCo, and Kellogg. In AI, it’s OpenAI, Google, Microsoft, Meta, and Amazon. These giants sit at the top of their ecosystems, and everyone else is either building around them or building to be bought by them.

The analogy is interesting because the infrastructure to launch a CPG company has become so commoditized (contract manufacturing, third-party logistics, Amazon as a distribution channel) that the barrier to entry has collapsed.

The result? The CPG category has become a petri dish for Entrepreneurs. The number of experiments being run in the industry has skyrocketed. Some are wacky. Some are interesting but safe. Prebiotic sodas. Grain-free tortillas. Oat milk that froths like dairy. New categories are being invented every year. Old categories are getting tweaked thousands of ways. Then Darwin shows up and does his thing by letting consumers speak with their wallets.

And guess what happens? Darwin’s survivors end up operating as independent companies or get bought by one of the giants.

Sound familiar?

The AI ecosystem works the same way. Cloud infrastructure, open-source models, and API-first development have made it trivially easy to spin up an AI company. So Founders are running thousands of experiments. Most will fail. But the ones that get traction and are doing something new or different? An exit is almost certainly available.

Why? For CPG companies, the economics of their product are terrible while they’re sub-scale and instantly better when acquired. The giant buyer brings distribution muscle and manufacturing cost advantages that the Startup could never achieve alone. The acquirer makes the business better, which is why they can pay a premium.

RXBar sold to Kellogg for $600MM (about 5x revenue). Siete just went to PepsiCo for $1.2B. Campbell paid $2.7B for Rao’s. These aren’t lottery tickets. They’re rational transactions where both sides win.

So is AI different? Yes and no.

The parallel holds for the concept: Kingmakers at the top, infrastructure enabling experimentation, acquirers who can extract more value than founders can alone.

But here’s where the analogy breaks down.

CPG acquisitions happen at 3-5x revenue. AI companies are getting funded at 20x, 50x, sometimes 100x revenue. The math only works if crazy growth rates are sustained OR if the multiples at funding roughly match the multiples at exit. And right now, growth rates are crazy and the big AI acquirers are valued at nosebleed levels themselves which allows them to use their currency to pay premium prices.

But multiples compress. They always do. And when the giants’ valuations come back to earth, their appetite for expensive acquisitions should shrink with them. The currency they’re using to buy Startups is cheap right now, but when it isn’t, the music might just stop.

So is everyone lighting money on fire? Probably, in aggregate. But for the best companies that survive Darwin’s filter, the logic might actually hold, at least while the current multiple environment persists. The protein bar playbook should work until it doesn’t.