Serve's CEO on Why Roaches Beat Unicorns

Months before Covid-19 hit the United States, Ali Kashani wanted to be nothing more than a cockroach. Serve Robotics’ CEO chuckles knowingly, when I suggest it might not be the ideal analogy for a food delivery service.  

It was strictly an internal descriptor, he explains, before adding, “Cockroach versus unicorn. Unicorns die in the nuclear winter, cockroaches survive. So, we decided to be a cockroach at the peak.” 

For those among us pure enough to not speak venture capital, the term “unicorn” was coined in a 2013 TechCrunch article to refer to the statistical rarity of a startup valued at over $1 billion. The phenomenon has grown more common in the intervening decade-plus, but the industry still utilizes derivations of the term, as even well-funded startups tend to fail nearly all of the time. 

Kashani’s point is well taken. In addition to rarity, “unicorn” imparts a sense of fragility. If the one horned equines were made of something more robust, we’d likely not feel so inclined to comment on their scarcity. Roaches, on the other hand, well, the less said, the better. 

The household pests were a piece of an internal talk titled, “Winter is Coming.” In spite of the House of Starkness of it all, however, Kashani says the decision to fortify the company’s insectoid exoskeleton was a product of things having gone fairly well up to that point.  

“There are times where you're at the peak and the world looks like you're never going to, you’re never going to lose,” he says. “Everything you touch turns to gold and everybody's so excited. And then suddenly overnight it all goes away, and you can't raise capital and you're kind of in a situation where everybody tells you you're wrong and suddenly everybody's focus is somewhere else.” 

From the outside, at least, nothing about Serve’s nearly decade-long existence has been by the book. The company began life as Postmates X, the robotics division of the food delivery service. That was back in 2017, well before it was taken as read that every startup needed to be making their own robots on the side. Uber bought Postmates in 2020 (the year the aforementioned “Winter” talk was given), before spinning it out as Serve in March of the following year.  

Since then, Serve has inked deals with former owner Uber, top competitor, DoorDash, and convenience giant, 7-11. Last week, the company left many in the industry scratching their heads when it announced plans to acquire Austin-based healthcare robotics startup, Diligent.  

Kashani says the deal belies the idea that Serve’s goal is simply “moving food over sidewalks.”  

He adds, “It's not just a food database problem. It's the robots that move among humans in very complex, unstructured environments. Not a factory floor, also not the road. I think those are two domains that we are happy to let others focus on. There's a lot of work being done there, but pretty much everything in between is what we consider these unstructured, complex human environments. And if you solve that, I think there is just a long tail of applications, many of which we probably can't even think of right now, the last mile logistics.” 

Founded the same year Serve began life internally at Postmates, Diligent has made a name for itself in the healthcare industry. Both companies saw unique challenges with pandemic-fueled labor shortages. In Diligent’s case, several hospitals across the United States adopted the Moxi humanoid robot to aid nurses and other overburdened healthcare workers.  

In terms of sheer deployment, Diligent’s growth has been the more modest of the two. Kashani notes that the Austin-based startup currently has around 100 robots deployed — similar to where Serve was around this time a year ago. He adds, however, that its own figure has since scaled to around 2,000. Serve has learned how to build and deliver robots, with help from auto manufacturers like Magna — learnings it intends to use to help grow Dilgent’s fleet.  

Diligent, meanwhile, has plenty to teach the company about indoor navigation, after years of operating autonomously around hectic hospital floors. The Moxi robot has mastered some other handy tricks, as well, including the ability to ride up and down elevators. In both cases, the companies have been engaged in public-facing, real-world testing.  

“We've had folks come from other parts of the industry where they learned about the value of building in real life, or they really wanted to be out there, and we were able to provide that to them,” says Kashani. “Once you’re actually out there, when you're addicted to the real-world lessons, the set of choices you have gets limited when most people are building in the lab. I feel like there is a good DNA for this team as well and that's why it's been a good fit.”