He Took a Grocery Startup to $40M, Raised $35M, and Walked Away With Nothing

The buyers were at the table right up until they weren’t. Pavel Danilov had taken Fridge No More, an ultrafast grocery delivery company, from zero to $40,000,000 of revenue across 30 dark stores in New York and Boston between 2020 and 2022, and when the unit economics stopped working he did what a founder is supposed to do: he went looking for an acquirer. “We negotiated exit. We had a conversation with a few potential investors, but then at last minute, they walked.” No flash sale, no soft landing. “Basically, we shut it down.”

What came next is the part most shutdown stories skip. “It’s not just like we shut down. We shut down, and then I worked for almost a year on liquidation,” he told Nathan Latka. “And I paid all the salaries, all the taxes, but then I need to negotiate to all the suppliers and real estate.” Around $35,000,000 of investor money had gone into the company, by his own count. Nothing came back out of it.

Why the dark stores never reached breakeven

Latka opened on ghost kitchens, a category he had watched go off a cliff after two years of founders telling him on this show that they were crushing it. Danilov corrected the label and kept the diagnosis. “My business was not the technically speaking, Ghost Kitchen. It was a dark store, so we were selling groceries. But I think the trend is similar.”

“The main problem is courier related costs. So you pay them per hour, and it’s very high. So basically, it’s really hard to get to the breakeven.”

• Couriers bill by the hour — the largest cost line moves with time staffed, not with orders delivered, so volume alone never closes the gap.
• Every new store restarts the clock — “every time you start a new store, it’s a new market, you start from zero, you need to reach this breakeven.”
• Growth stacks the losses — grow fast and you end up, in his words, with a bunch of stores making losses and only a few already profitable.
• The only fix was more capital — hard to deal with, he said, without an extra capital injection. “And in ’21, you know, beginning ’22, it changed.”

That last clause is the whole cohort’s obituary. The model was never self-funding; it was a bet that cheap money would last long enough for enough stores to cross over. Thirty stores, two metro areas and $40,000,000 of revenue were not evidence the bet was working. They were the size of the position when the market moved.

A year of liquidation

Latka thanked him for the transparency and started to move on. Danilov didn’t. The shutdown was an event; the liquidation was the education. Salaries first, then taxes, then a year of negotiating with suppliers and landlords on behalf of a company that no longer sold anything. It was “quite a cumbersome process,” he said — not the kind of thing where you close one business and walk straight into the next.

He had spent eight years in investment banking before founding Fridge No More, which is presumably why the arithmetic of winding down a venture-backed company held no mystery for him. It still cost him a year of his life.

The scar became the spec

Bennudata reached him sideways, and he is precise about how little it has to do with groceries: “They are only connected through one of my investors. So one of my angel investors introduced me to” the cofounders. There are three of them, and they split the equity evenly at the start. What Danilov brought to the table was less a thesis than a constraint. He still carries a scar, he said, from a “human intensive business” with a lot of employees.

“So I wanted to build something which doesn’t have a lot of employees, just software and b to b.”

Backup is automated. Recovery still isn’t.

“The backup is a huge business. Right? How you back up your servers, your cloud resources, and it’s automated. But with the recovery, how you actually bring this backup back to life, back online is very manual still.”

That asymmetry is the whole opening. Danilov’s argument is that the cloud platforms have already shipped the primitives — cloud APIs, native backup — so the recovery plan itself can be assembled from data collected through those APIs and kept current automatically, rather than living in a runbook nobody has rehearsed. “So what it means when a disaster happens, I think ops or whoever is responsible needs to push this one button and the whole infrastructure, like applications will be recovered in a different environment.”

Who feels that pain first is the more interesting half of the answer. All three of his pilots are late-stage startups that have begun selling into enterprises and are now being asked for a disaster recovery process they either don’t have or have in a very simplified form. The pitch is that they don’t need “three engineers and twelve months” to build a custom one in-house. Midsize enterprises are the second segment, and he is candid that those are a longer conversation.

Pre-revenue, three pilots, none of them paid

The conversation was recorded in the first days of December 2023: Danilov is mid-programme at Berkeley SkyDeck, still pre-revenue, promising a first product release “by the year end” and first customers in Q1, three to five of them if the plan holds. The GetLatka record for Bennudata carries the rows that bracket that claim — revenue of $0 and a team of 3, both dated 1 January 2023, and a $400,000 pre-seed round dated 5 December 2023.

The funding line needs one caveat. Latka calls the $400,000 a seed round in his wrap; Danilov says it came from angels; GetLatka books it as a pre-seed. And it does not include the accelerator cheque: Berkeley SkyDeck invests $200,000 for around 7%, which Danilov describes as similar to others, and Bennudata is in the seventeenth batch — sixteen batches of later-stage alumni ahead of them, which is exactly where the pilots came from. Of the $400,000, $150,000 has gone out the door, mostly on the engineering team building the product, with the rest still in the bank.

• 2020–2022 · Fridge No More zero to $40,000,000 of revenue, 30 dark stores, then shutdown and a year of liquidation.
• Nov 2022 · Bennudata founded three cofounders, equity split evenly, introduced through one of Danilov’s angel investors.
• Jan 2023 · First GetLatka rows revenue $0 and team size 3, both as of 1 January 2023.
• Dec 2023 · $400,000 pre-seed recorded 5 December 2023; $150,000 spent, mostly on engineers.

Latka pushed on the obvious weak point: why aren’t the pilots paid? The answer was product maturity rather than pricing nerve. “I think people are not familiar with the product,” Danilov said, and the company is still building it while the pilots run — “half automation, half consulting” is his own description of the current mode. Q1 is when that is supposed to end.

Two prices, two channels

It is a conventional two-tier SaaS pricing structure with an unusually honest division of labour behind it: the marketplace tier is a self-serve wedge that costs nothing to sell, and the enterprise tier is where the annual contract value lives, at roughly five times the price and a much longer cycle. For a company that has spent $150,000 of the $400,000 it raised, that split is less a growth strategy than a way to keep the cheap channel earning while the expensive one matures.

Neither of the paid channels he tried is doing that job yet. “We used LinkedIn and Google. It’s not very effective so far,” he said — which, at three pilots and no revenue, is a verdict on customer acquisition cost before there is anything to acquire customers for. What works is the network: direct emails to the Berkeley and SkyDeck networks, advisers and alumni, which he calls pretty effective so far. The accelerator, in other words, is functioning as a distribution channel first and a $200,000 cheque second.

Danilov is 41, married with two children, his oldest daughter 13. He sleeps eight hours a night, names The Hard Thing About Hard Things as his business book, follows no CEO in particular, and builds websites in Softr. Asked what he wishes he had known at 20 — eight years of banking and one $40,000,000 shutdown ago — he gave the shortest answer of the interview.

“I wish I knew that tech was going to that big. And I, you know, jumped into that earlier than I did.”

Pavel Danilov, cofounder, Bennudata

Sources — Pavel Danilov, cofounder of Bennudata, interviewed by Nathan Latka, recorded in early December 2023. The Fridge No More figures ($40,000,000 of revenue, $35,000,000 raised, 30 dark stores in New York and Boston, 2020 to 2022), the Berkeley SkyDeck terms, the pilot count, the spend and both price bands are as stated on that tape; the $40,000,000 was introduced by Latka in his opening and closing summaries and not disputed by Danilov, who confirmed the raise, the collapsed exit and the shutdown in his own words. The GetLatka company record for Bennudata supplies the dated rows: revenue of $0 and team size of 3 as of 1 January 2023, and a $400,000 pre-seed round dated 5 December 2023.