The Uncrowded Trade
In July I claimed the capital left food and the buyers didn't. Here is the work behind that.
A few weeks ago I made a brief post arguing that early capital has walked away from food while the companies that buy food brands never went anywhere, and that the distance between those two facts creates a mispricing. It was a short LinkedIn post, so here is the longer version.
Where the money actually went.
Everyone has seen a version of the concentration number by now. The one I cannot get past is newer and more extreme than the figures that were circulating in the Spring. In the second quarter of this year, by PitchBook's count, 87.5% of every venture dollar invested in the United States went to AI companies. That is the most lopsided split they have ever recorded. Nearly nine dollars in ten, in the largest venture market in the world, in a single quarter.
For scale, the first quarter was the one that made headlines: roughly $300 billion into startups globally, about $242 billion of it into AI. The second quarter was smaller in absolute terms and more concentrated in relative ones.
I want to be careful here, because there is a version of this essay that is just complaining, and that’s neither helpful nor productive. I spent five years building my own venture-backed company and I understand why capital behaves this way (though I also see the folly in it). I don’t care to wager a view on whether any of this lopsided landscape is overpriced, either.
What I have is an interest in the other side of the ledger. Because a crowded trade does two things at once: it raises the price of what everyone is buying, and it lowers the price of everything else. Only one of those gets written about.
What that did to the cost of backing a brand.
In the second quarter, foodtech recorded 142 venture deals, its fewest in eight years, at an average round size around $11 million. Agtech had its weakest quarter in eight years in the same window. The longer trend is worse: US venture funding into companies at the intersection of e-commerce and consumer products fell about 97% from its 2021 peak, from more than $5 billion to a little over $130 million by late 2023, and it has not meaningfully recovered.
The consumer capital that remains has mostly bunched up around brands that are already working, where the velocity is proven and the diligence is easy. That is a rational place to put money. It is also an expensive place to put money.
Who is on the other end?
Now look at where food brands go when they succeed.
Strategic acquirers drove about 88% of food and beverage deal flow last year. Food M&A is up roughly 67% so far this year, and deal activity in branded food specifically has more than tripled. Last month Ferrero agreed to buy Purely Elizabeth, a granola brand founded in Boulder in 2009 and now doing around $250 million in sales, for a reported $850 million. Less than three weeks after that, Barilla agreed to buy Goodles, a boxed mac and cheese brand launched in 2020, on undisclosed terms. Two Italian family companies, two American better-for-you brands, nineteen days apart. And in the two years before those: Mars closed on Kellanova ($35.9B), Intersnack agreed to buy Utz ($2.9B), PepsiCo bought Poppi ($1.95B, $1.65B net of tax benefits), and Hershey bought LesserEvil (reported $750M).
Line those two facts up. Backing an American food brand costs less today than it has in a decade. Strategic buyers have rarely been more active. Very few investors work the distance between the two, and that distance is where I spend my time.
What the revenue is made of.
The quality of the underlying numbers provides a second reason I find this side of the trade more comfortable.
A meaningful share of the money moving through the AI economy moves in circles. OpenAI has contracted to buy $250 billion of cloud services from Microsoft, which owns roughly a quarter of OpenAI. This month Nvidia agreed to guarantee something over $100 billion of payments on a data center built to run Nvidia chips. Though perhaps curious, none of this is necessarily improper, and vendor financing in the 1990s helped build infrastructure we all still use. It does mean, however, that a serious portion of the revenue underneath the boom represents money passing between companies that own pieces of each other.
It has also not yet arrived where you would expect. In February, a survey of nearly 6,000 senior executives across four countries found that about nine in ten reported no measurable effect on their firm's productivity or employment from AI over the previous three years. That will likely change, though it has not yet.
Again, I say this not to gripe, but to juxtapose the fact that food revenue has none of this texture. Somebody bought the thing, ate it, and decided whether to buy it again. There is no structure to unwind, no counterparty who is also an investor, and no ambiguity about whether the demand is real. The numbers are extremely legible and the operations entirely tangible.
A note from the dotcom era.
From the Nasdaq's peak in March 2000 to its low in October 2002, the index fell about 78%. Consumer staples offered the only sector of the S&P 500 to finish that stretch with a positive return. I don’t mean to make a forecast with that, and anyone who tells you they know how this cycle resolves can only guess. The point is narrower: food demand does not need the cycle to cooperate.
Food is hard, and that’s the point.
Food is not a magical, leisurely financial safe haven, of course. You live or die on margins. Cold chain punishes a single mistake. One bad shelf reset undoes a year of work. Retail has been genuinely difficult this year, too: grocery dollars rose while volumes fell about a percent, and roughly half of shoppers say they are cutting impulse purchases. A cheap entry price makes none of that easier.
It does something better. Difficulty keeps the category cheap and keeps it uncrowded. Easy businesses attract capital until the return disappears. This one has attracted very little since 2021, which means the founders still building in it chose the work rather than the funding. When nobody will hand you a big round, you price for margin, you learn your loaded unit cost, and you earn a second purchase from someone who owes you nothing. Those skills are unfashionable. They are also the ones that survive a bad year. So the claim is narrow, and I will state it plainly.
The entry price has fallen a long way. The buyers have kept buying through every cycle anyone can name. And the discipline this drought imposes selects hard for the people worth backing.
That is the trade. It is uncrowded for reasons that are mostly about hype, and I would rather be early and unglamorous than late and correct.
If you are building a food brand in this stretch, I would like to hear about it: firstruncpg.com/pitch
