The Food Delivery Death Star

VCs still give billions to food delivery startups. But what’s their endgame?

“We’re math geeks. We don’t know anything about making food.

It’s not the first thing you’d expect to hear from the CEO of America’s hottest food delivery startup — but then again, maybe it should be. Tony Xu is the 31-year-old CEO of DoorDash, a company whose fleet of on-demand drivers bring restaurant food to hungry customers. Xu grew up in the Bay Area, where his parents ran a small Chinese restaurant. Today, Xu serves far more customers than his parents — he serves 1-in-3 households in Silicon Valley.

By all accounts, DoorDash (a company founded in 2013) is a sterling success. With more than 700 employees — not including its contractor-drivers — , $180M+ in funding, and impressive month-over-month growth, Tony Xu is the poster-child for the sort of rapid Zuckerberg-esque ascent that today’s technology makes possible. The CEO of a sexy consumer-facing company with a sky-high valuation, Xu has even won comparisons to Amazon chief Jeff Bezos.

As they say in Silicon Valley, Xu and his team are “crushing it”. They’ve been invited to lecture at Y Combinator. They dole out startup advice on industry podcasts. They’re in the midsts of an aggressive international expansion.

And yet, unlike his parent’s Chinese restaurant, DoorDash loses money on almost every meal it serves.

DoorDash CEO Tony Xu. Image credit: Steve Jennings (Getty Images)

In the last five years, investors have poured billions into food delivery startups like DoorDash, hoping that they (or one of their competitors) become large enough to form a lucrative monopoly — like Uber has for taxis, Google for search, and Amazon for e-commerce.

Most of the new delivery companies offer a similar service: you order food on an app, and within 10 to 100 minutes, your food arrives.

Some companies (like DoorDash, Postmates and Caviar) deliver food from restaurants that don’t have their own drivers. Some (like Sprig and Maple) deliver hot meals which they make themselves. Some (like Munchery) deliver cold food you have to reheat. Some, like Zume, deliver with the help of mechanical robots. Others, like Forkable, use artificial intelligence.

10X growth: In 2010, 20 venture capital investors had funded an on-demand company. By 2015, more than 200 had. Source: CBinsights

There are dozens (if not hundreds) of new startups delivering food: À La Carte Express, Bento, Deliveroo, Deliverd, DeliveryHero, EatFirst, Farm Hill, Favor, Fluc, Foodora, FreshMint, Gourmaleo, Homer, PepperTap, Radish, Spicy Radish, SpoonRocket, SupperBell, TinyOwl, Thistle, Wizrd, Zomato, Zoomer, Zume. Legacy food ordering websites (like, and now have their own drivers, while titans like Uber, Amazon, and Google have also entered the fray.

Everyone, it seems, wants a piece of the pie delivery fee.

And yet…

While food delivery was “possibly the most saturated industry in 2014/2015”, competition thinned out in 2016. As more and more delivery startups cropped up, a “ruinous marketing war” ensued, with startups offering lower and lower prices to undercut competitors for marketshare.

It wasn’t long before the space saw its first casualties: Sprig (funding: $57M) ceased operations in Palo Alto and Chicago, SpoonRocket (funding: $13.5M) shut down entirely, DoorDash (funding: $186.7M) lost a good chunk of its near-billion dollar valuation, Square tried and failed to sell its delivery service, Caviar (which it bought in 2014 for $90M) to, Munchery’s CEO is on the outs while the company prepares for a massive devaluation, and Postmates was rumoured to have tried to sell its business before raising $141M this week from Peter Thiel’s Founders Fund.

To Postmates’ credit — and to its competitors’ chagrin — it managed to raise a “flat round” which did not lower the $600M valuation it received in 2015. However, Quartz reported that Postmates’ latest deal was laden with investor-protecting “deal sweeteners” that may have inflated the real value of Postmates stock.

Image credit: Bloomberg + CB Insights

2016’s chilly investment climate was a result of “extensive expectation resetting” among food delivery investors. The New York Times said it was “the end of the on-demand dream”. Pando, more memorably, called it the “foodpocalypse”.

But what happened? How did such a frothy industry go frosty so quickly? Are investors bi-polar? Were books being cooked? Was delivery a fad?

The answer is a complicated one. And it starts with a San Franciscan company.

Its name is Bento.

Bento and the Curse of Growth


We’ve raised $2 million dollars and have been featured in TechCrunch, CNBC, and the San Francisco Chronicle. We now have $350,000 left in the bank, which will last til the end of April.

We need to raise more money — or we will die.”

For a CEO, Jason Demant is unusually candid. Perhaps it’s his lack of experience. Or perhaps it’s because he’s run out of options.

Demant made the decision to tape-record his conversations (both with his cofounder, and his wife) when his company, Bento, was starting to fail. He published them on StartUp, a popular technology podcast.

In 2015, despite not having ever worked at restaurant, Demant and his cofounder Vincent Cardillo decided to create their own food app. The app let customers order customizable “bento boxes”, which would be delivered anywhere in San Francisco.

They launched their app-restaurant hybrid at the 2015 LAUNCH festival, which is organized by Uber investor Jason Calacanis. Bento’s on-stage pitch was received with rave reviews. The team soon raised $2M in seed funding.

However, five-and-a-half months after formally launching, Demant was contacted by his accountant, who informed him of Bento’s staggering losses.

“We burned almost $70,000 more than we expected — almost 30–40% more — and it was like ‘woah, what happened?’’’

— Bento CEO Jason Demant

As it turns out, the company was paying $32 to make and deliver bento boxes being sold for $12.

Between ingredients, kitchen equipment, chefs, drivers, and the costs of developing its app and logistics software, Bento was losing up to $20 per meal. Worse yet, the company was growing 15% per week.

“I mean, it’s almost a little embarrassing — because I should have been watching [operation costs], especially in an operationally-intensive business like this.”

While revenue growth is the usual benchmark for startup success, in Bento’s case, it was accelerating their losses.

Rigorous cost-cutting (included the layoff of Bento’s kitchen staff) and the introduction of higher delivery charges would eventually get Bento closer to profitability: “We got to, like, positive a penny,” recounts Cardillo, Bento’s CTO.

Nevertheless, the company would eventually shutter its app and the on-demand ordering part of its business.

“What we have learned in the last year basically tells us the way that we started the business was f***ing stupid. But if we had been in restaurants — or had tried a similar business or worked at SpoonRocket [another food delivery startup] for 3 years — we probably would have just known all this.”

— Bento CEO Jason Demant

Despite the struggles, Demant and his team would successfully raise another $100,000 to help them pivot to towards a straightforward catering model. However, the disappointment in Demant’s voice while recounting the Bento story is palpable:

“We had smart people encouraging us, I dunno… It felt like we were doing all the right things.”

Dancing on Razor Thin Margins

Why food delivery is harder than people delivery — and why investors are still interested

Why would an investor ever pay $20 to subsidize your lunch? On its face, it seems absurd. And yet, it’s the same strategy investors have pursued with Amazon, Uber, and now Postmates: they provide so much cash to these companies that they’re able to drown their competition. Billion dollar war chests allow companies like Uber to run at a loss, and offer products below their actual cost. That helps them gain customers and kill competitors. Once they’ve established a monopoly, they can then raise prices to improve their margins.

“Uber’s main mode is this Death Star thing… they’re using their balance sheet as a weapon very well, and I think it’s actually unprecedented.” — Postmates CEO Bastian Lehmann

However, the strategy is significantly riskier for food delivery startups, and investors know this: in 2001, delivery startup WebVan went bankrupt after raising $396 million (despite cumulative revenues of only $395,000.) Notably, three of WebVan’s investors are also investors in DoorDash.

So what makes food delivery so hard? Delivery companies face special challenges that other on-demand startups (like Uber) don’t. Transporting people is a simple problem of connecting drivers and customers; food delivery startups have the additional problem of connecting drivers and customers and kitchens.

The extra step not only introduces extra costs, but it makes margins thinner, too. Customers are used to paying $30 to Uber across the city — but very few will pay $30 to Uber their sandwich. So while driver costs are comparable, food delivery services are unable to charge as much for their services.

The psychological aversion to high delivery fees has forced delivery startups to find creative income streams. DoorDash edits the menus of participating restaurants to include a hidden 20% markup fee — in addition to its standard delivery charges. (Xu defends the practice as helping reduce “sticker shock” on DoorDash prices). Most other companies simply charge a flat delivery fee to the customer (and a commission to the restaurant). Some, allegedly, go as far as to take their driver’s tips.

But despite its penny-squeezing manoeuvres, DoorDash has yet to turn a profit. So why are investors taking the risk?

Robot Saviours?

The on-demand economy now employs more than a million drivers. Even if they aren’t technically classified as “employees”, they’re the largest expense for on-demand companies like Uber, Lyft, DoorDash and Postmates. But what if those drivers could all be fired?

Uber, who employs a majority of those drivers, is trying its best to replace them with self-driving cars. Last year, the company poached 50 robotics researchers from Carnegie Mellon in its attempt to build the world’s biggest fleet of autonomous vehicles. Without drivers, DoorDash (and other food delivery companies) would stand to gain even more than Uber: since they make less money per ride, a driverless fleet would help them pad their margins, free them from expensive liability lawsuits, and liberate them from an another big capital expense: recruitment.

DoorDash spends more than $200 recruiting each delivery driver, and according to the New York Times, those drivers usually only work three to six months for the company. Multiply $200 by 25,000 drivers (a conservative estimate), and that’s a $5,000,000 expense, every few months, just to get DoorDashers on the road. That doesn’t include the money that DoorDash and its ilk pay their drivers for deliveries.

As on-demand companies scale, these costs don’t go down, either. Lyft reportedly loses $50M a month, while Uber loses about $700M a quarter. On a per unit basis, food delivery companies lose even more. So how do these companies expect to reach profitability if older, higher-margin, and higher-volume services aren’t able to?

Increasingly, it looks as though the long-term sustainability of on-demand companies hinges on a self-driving future. Investors are letting their companies lose money in the short-term — ravenously acquiring customers and marketshare — so that once the robots arrive, their companies will be positioned for a big payday.

That payday can’t come soon enough. Delivery companies have invested tremendous amounts of money in labor-saving contraptions, from autonomous “taco copters” and “burrito bombers” to robots that can make and deliver pizza.

Nevertheless, the driverless future is still a long way off: ground-based robots have a highly restricted range. Flying drones are beset with technical problems that make them unsuitable for use in urban environments. As for self-driving cars? Elon Musk believes that semi-autonomous cars could be as little as 2 years away. But fully driverless, road-legal cars? BMW predicts that the technology will take until at least 2021 to develop. Uber CEO Travis Kalanick thinks 2030 is a more realistic target for a full, legal roll-out.

That’s not to say that robots aren’t making significant inroads in the delivery industry: Forkable, a San Franciscan startup, uses an AI robot to orchestrate bulk food orders (which reduce its per-unit delivery costs). The Atlantic recently mocked the company for its hyperbolic claim of “reinventing lunch” — but Forkable might have the last laugh: it could be the tech industry’s first truly profitable delivery service.

Rather than processing orders from individual consumers — like DoorDash and Postmates — Forkable delivers large batch orders to multiple people at once. The company partners with coworking spaces and offices (where large lunch orders are already common), and its auto-ordering “lunchbot” means that office managers don’t have to referee fights on what food to order. As for the meals the bot chooses? They’re surprisingly well-received, largely because Forkable delivers personalized restaurant meals — not big trays of sandwiches, carrots and dip like traditional catering companies.

However, the company’s “AI” is likely just a clever algorithm that learns what you like to eat via pre- and post-meal surveys. What is more striking than Forkable’s tech is Forkable’s business savvy. Forkable wins the loyalty of eaters since their meals are highly-personalized. And by helping office managers avoid the hassle of corralling orders and calling restaurants, Forkable is able to secure long-term food delivery contracts.

The end result is big, predictable, recurring orders — a trifecta few delivery startups have replicated. And while Forkable’s funding history is curiously opaque, it is already beating Postmates and DoorDash to the Asian and European markets.

The Future of Delivery

Many investors believed (naïvely) that creating an “Uber for Food” was as simple as taking Uber, and adding a few sandwiches. Clearly that is not the case.

Despite the setbacks, many people are optimistic about online food delivery. Morgan Stanley predicts the market will be worth $210 billion a year, citing rapid growth: pizza chains now take more than 50% of their orders online, South Korea orders 75% of its takeout online, and food delivery startups are signing up new customers every day.

While there’s no question about the demand for these services, how (and which) companies will capture that $210 billion is much less clear.

Will self-driving cars arrive in time for DoorDash or Postmates? Or will investors eventually tire of waiting? Will “death stars” like Uber and Amazon double-down on delivery and outlast their younger competitors? Or will they give up on delivery and return to their fatter-margin homesteads? Will a breakthrough in drone technology reduce the need for self-driving cars? Or will the breakthrough come in the form of better AI software, like Forkable?

In the emerging food delivery wars, the stakes are high, the margins are slim, and the competition is vast. Some will fail. Others will quit. But a few of them might just end up delivering the billion-dollar goods.

The California Review is an independent tech publication founded in 2016. We pick one industry every month, and publish our highly-researched, uncensored findings in a story format like this. You can join our 1000+ newsletter subscribers here. Or you can just recommend this post.

The California Review

Written by

Longform criticism on startups, tech, and venture capital.