The Dirty Word
eCommerce has a bad rep but it's a prime investment for the AI age and rollups
eCommerce is a dirty word in private equity right now. I get it.
Act 1
Between 2020 and 2023, investors poured over $16 billion into eCommerce aggregators. Less than $1 billion of that capital is still available for deployment today. The graveyard includes: Thrasio, Perch, Berin Brands, Elevate, Razor Group, Branded, Moonshot, and Stryze. Each raised millions and 100s of millions of dollars, and in Thrasio’s example $3.4 billion at a $10 billion valuation, and filed for Chapter 11. Overall, we saw a 99% drawdown in two years; funding to the sector went from $6 billion in 2021 to $68 million through May 2023. Let that sink in.
That’s the body count. And that’s why most investors I talk to have an aversion to the word “eCommerce.”
But here’s why that’s the wrong signal.
What collapsed was not the eCommerce thesis or the market - it was a specific type of operator buying eCommerce businesses with too much debt, at peak multiples, without scalable systems to operate them. The asset class didn’t break. Instead, the cycle did exactly what cycles are supposed to do; it cleansed the market of the operators who shouldn’t have been there in the first place. Yielding in a smaller pool of buyers, with equally motivated and more realistic sellers, and a better-priced asset class.
Aggregators raised capital fast, most of it debt at 70-80% leverage, and used it to buy Amazon brands in volume. The margin was supposed to come from centralizing ops across the portfolio. It worked on paper, but every assumption broke in practice.
Thrasio at its peak was acquiring up to one Amazon brand per week. They built a portfolio of 100s of brands in a few years. Think about what that means operationally. Every brand they bought came with its own manufacturers, freight contracts, SKU history, and ad account. Multiply that by a few hundred and you have the operational footprint Thrasio was trying to run from a centralized team. The thesis was that you could plug all of it into a centralized ops function and unlock margin. No team could have run that operational variance without purpose-built software, and the software didn’t exist in 2021. The AI workflows that make multi-brand operating tractable today are the system we’ve been building at Pilot Wave for seven years.
Three things hit at once: rates went from zero to five, the operator talent pool contracted, and capital-deployment pressure pushed buyers and multiples. None of these were predicted by the underwriting models running in 2020. And the strategy itself had a structural flaw nobody priced in: aggregators bought competing brands in the same categories, thinking they were diversifying, but on Amazon, two brands selling similar products in the same category compete with each other for ad placement and organic ranking. On Amazon, two brands selling similar products in the same category compete with each other for ad placement and organic ranking. Aggregator portfolios cannibalized themselves, and a lot of the brands got worse.
Act 2
The reality is that this is not an uncommon pattern with a new investment thesis. The first wave overcapitalizes, overpays, and implodes, but the opportunity is still alive. The dotcom bust didn’t kill the internet; it cleared the runway for the operators who built on it after 2003. The aggregator wave was the first rush. It’s up to the next wave to take advantage of the thesis.
Now look at what’s actually true about eCommerce today.
eCommerce hit 25% of total US retail sales in Q4 2025. That’s the highest penetration since the Commerce Department started tracking online sales in 1999. That share has not declined in a single year since the data series began. This is a fundamental part of the thesis that hasn’t changed and shows no sign of reversing.
Inside this trend, Amazon keeps climbing. Amazon and its sellers moved $830 billion of goods in 2025. Third-party marketplace sales (brands like the ones we buy) were $575 billion of that. Since 2019, Amazon GMV has grown by roughly $495 billion. 76% of that incremental growth came from the marketplace, not from Amazon’s own retail operation. Third-party sellers now drive 69% of total GMV, up from 60% in 2019.
Read that again. The largest retailer in the world is increasingly a platform on which independent brands sell, and those independent brands are taking share faster than Amazon itself. These are our targets.
To put things into numbers: fewer than 8,000 sellers generate half of Amazon’s $300 billion in US third-party GMV. That’s 1.6% of the active seller base. In 2023, it took 15,000 sellers to hit that same threshold. The concentration has nearly doubled in less than three years. This is brand-level natural selection in real time. The best brands survived the last five years and got bigger. Most of the rest got compressed out. What that means for us is simple: the asset pool we’re buying from has never been higher quality, and it has never been more concentrated in the hands of a small group of operators who are ready to sell.
And the price is finally fair. When aggregators were deploying $6 billion a year, they pushed Amazon FBA multiples to 6-8x EBITDA. Sellers got rich while buyers got overpaid.
Act 3
Today, healthy growing FBA brands trade at 3-5x EBITDA. Omni-channel brands in the growing categories and with healthy margins can still exit at 7-8x. That’s the entry-to-exit math.
And nothing has changed about the supply side of this trade. Founders still want to sell, but not for the reasons people assume. The best operators in this space are 0-to-1 builders. They enjoy the thrill of building a brand from nothing until it gets to a stage that a solo founder can no longer run alone, and now they’re itching for the next idea. The buyer pool has dramatically shrunk, so this is exactly the right moment for disciplined teams to deploy capital.
I’ll keep the proof short. We have grown revenue at 100% of the brands we have acquired. Our largest portfolio brand was at breakeven EBITDA when we acquired it; it is meaningfully profitable today. At another, we have more than doubled EBITDA under our ownership. We have done this through the worst macro environment the eCommerce sector has ever faced: COVID demand swings, freight cost spikes, the collapse of the aggregator buyer pool, tariffs that have fundamentally reset landed costs, and now changes in search because of AI. Our largest brand grew topline through all of it, in significant part because the AI workflows let us react to supply chain disruption faster than incumbents could.
That’s the proof, and not what we plan to do. The facts of what we have overcome in the toughest possible conditions.
The risks are the ones everyone already knows, and they are the reason “eCommerce” is dirty in the first place: tariffs and trade policy, Amazon platform risk, AI commoditization, consumer pullback in a downturn, Chinese sellers going direct to US consumers, and rising customer acquisition costs across every paid channel.
None of these are wrong. They are also not new, and they are not the reason aggregators failed.
eCommerce got dirty for a reason. $16 billion in capital was largely mispent in three years. That history is real, and the investors avoiding the word have a defensible reason for doing so.
The argument against eCommerce is built on risks that the underlying asset class has absorbed and grown through anyway. US eCommerce hit 25% of total retail in Q4 2025, despite all the challenges, and Amazon’s third-party marketplace generated 76% of incremental GMV over the same window. The demand kept compounding through every risk the bears named. Multiples have reset. Founders still want exits. The buyer pool is largely empty. And the operating tools that didn’t exist in 2021 exist now, so it’s time to retire the old definition and acknowledge that eCommerce is back on the menu.
The trade is the same one the first aggregators (including us) saw. The conditions are now what they should have been the first time around.
And remember once bitten, twice shy is not a strategy.
Links:
Pilot Wave Holdings: https://www.pilotwaveholdings.com
Pilot Wave on X: https://x.com/pilotwaveai
Pilot Wave on LinkedIn: https://www.linkedin.com/company/pilot-wave-holdings/
Arik Oganesian on LinkedIn: https://www.linkedin.com/in/arik-oganesian/
Thanks for reading Pilot Wave ! Subscribe for free to receive new posts and support my work.