The burrito that launched a robot army
He was ordering food online in Berkeley. The delivery fee was nearly as expensive as the food itself. Most people would have grumbled and moved on. Felipe Chávez Cortés built 500 robots instead.
That is the kind of founder Felipe is - someone who watches the same delivery fee everyone else pays, and instead of accepting it, starts reverse-engineering how to make it cost nothing. The gap between "that's annoying" and "that's a company" is where most people stop. Felipe set up a prototype on a Berkeley sidewalk.
Today he runs Robot.com - the company formerly known as Kiwibot, rebranded in May 2025 to match the scale of what it has actually become. Five hundred autonomous robots. More than a million completed delivery tasks. Operations in 30+ campuses across the United States, plus Dubai, Saudi Arabia, and pilot programs in Paris, Seoul, Berlin, and São Paulo. For a company that started by delivering burritos at walking speed on a single college campus, that is not an evolution. That is a category expansion.
"Robots aren't coming. They're here. And we're giving them work to do."- Felipe Chávez Cortés, CEO of Robot.com
Before the robots, there was fruit
The naming convention alone tells you something about the man. His two startups are both named after fruit - Lulo (a Colombian citrus) and Kiwi - because those happen to be his favorites. You can read that as whimsy, or you can read it as a CEO who refuses to take the obvious path even when naming a company.
Lulo came first, built in 2013 with his longtime collaborator Sergio Pachón (now Kiwibot's President). They were students in Colombia trying to solve a real problem: how do you get delivery to work for university students when the economics of human couriers don't scale? Their answer - an app-based delivery service - became Colombia's first app to accept credit card payments. They bootstrapped it, scaled it, and sold it to Merqueo, a Latin American logistics unicorn. Still in college.
After selling Lulo, Felipe did what most founders don't: he went back to finish his degree. Then got accepted into UC Berkeley's Global Founders Program in 2017. The culture shock of Bay Area delivery pricing was, apparently, productive.
The World of Warcraft origin story
The backstory goes further back than Lulo. In high school, Felipe set up an unofficial private server for World of Warcraft. Within a month, it had tens of thousands of players. He wasn't supposed to do it. It wasn't a business. But it was his first proof-of-concept for what the internet could do - and what happens when you build something people actually want. The WoW server didn't make him rich. It made him understand scale.
The robot that walks like a pedestrian
Kiwibot's design is deliberate and counterintuitive. The robots move at pedestrian speed - roughly 3 to 4 miles per hour. They carry up to 25 pounds. Each unit has six cameras and LIDAR sensors with AI-powered obstacle detection. They don't race through sidewalks; they share them. Felipe has always framed this as a feature, not a limitation. His line on pedestrian safety is unambiguous: it comes before operational efficiency.
The technical challenge he's most candid about? Crossing a street. "The most complex thing is crossing a street," he said in 2018, when the company was still a Berkeley-only operation. That specific honesty - calling out the hard part instead of papering over it with optimism - runs through how he talks about the business. He doesn't oversell the technology. He sells the model.
Robot.com by the Numbers
The math that makes it work
Felipe's clearest articulation of the value proposition is numerical: a human courier averages 2.5 deliveries per hour. With Kiwibot's platform, a single operator can manage 12 deliveries per hour. That's not a marginal efficiency gain. It's a structural change in the cost curve of last-mile delivery.
The business model - robots-as-a-service with enterprise partnerships through Sodexo and GrubHub - is designed to slot into existing infrastructure rather than replace it. That framing matters. It's how you get a food service giant like Sodexo to expand a contract, rather than fight the technology.
From 2021 to 2023, Kiwibot went from 3 campuses to 30. Not 30 total deployments - 30 active campus contracts. In two years. The milestone that changed the velocity was hitting 300,000 completed tasks, then quadrupling it to one million in the next seven months. The acceleration is the story.
From kiwi bots to Robot.com
The rebranding to Robot.com in May 2025 is the clearest signal of where Felipe is steering the company. Kiwibot was a great name for a campus food delivery startup. It is not the right name for a multi-sector automation company operating warehouse bots that carry 80 pounds, advertising robots generating 3,000 impressions per mile (via the Nickelytics acquisition), and AI navigation chips developed by its Taiwan-based subsidiary AUTO Mobility Solutions.
The acquisitions tell the diversification story more concisely than any press release. Nickelytics brought mobile advertising. AUTO Mobility Solutions brought AI chip capability and reduced supply chain dependency on China. The leadership hires followed: J. Kim Fennell as Chief Business Officer (formerly Uber), Eduardo Iniguez as CFO. Robot.com is not a startup anymore. It's a platform company with hardware, software, and media assets.
"My mission is to reimagine and democratize robotic delivery by integrating it safely into communities and giving access to everyone regardless of socio-economic background."- Felipe Chávez Cortés
The founder's operating philosophy
Felipe draws a distinction between what he calls the "war CEO" and the "peace CEO" - two different modes that the same leader has to inhabit depending on what the company needs at a given moment. In growth phases, you need the clarity of a wartime operator: fast decisions, tight accountability, explicit processes. In stable phases, you build culture and systems.
His more quotidian principle is the board game analogy: as a company grows, you have to set rules everyone can understand, even if the rules slow you down. Ambiguity scales badly. He bootstrapped Lulo, so he knows what zero-structure looks like. He also runs a 150-person operation across continents, so he knows what absence-of-process costs.
The through-line in how he talks about building is the emphasis on people who believe in the vision. Not just competent people - believers. His language about his co-founders and early team members reads less like corporate HR and more like someone who built something genuinely difficult with people he trusted. That matters, because Kiwibot is a hardware company, and hardware companies are where startup optimism goes to get tested.
MIT, Berkeley, and the road from Bogotá
In 2018, MIT Technology Review named Felipe to its Innovators Under 35 - Latin America list. At the time, Kiwibot had completed 18,000 deliveries on the Berkeley campus in its first year. The recognition landed because the innovation wasn't just technical - it was about applying AI-powered navigation to a problem with real social consequences: who gets affordable, fast delivery and who doesn't.
He started a courier company in Bogotá in 2015. Three years later, MIT was writing about him. Three years after that, his robots were in Dubai. The trajectory is fast by any standard, but especially striking for a founder whose first real business lesson was running a $0 World of Warcraft server for tens of thousands of strangers who had no reason to show up - except that the product worked.
"The objective is to move things between two points in the city with a cost close to zero."
"A person using the Kiwi Campus platform can manage 12 deliveries in one hour - compared to 2.5 for a human courier."
"The most complex thing is crossing a street."
"When a company grows, it's important to set processes and rules. Push a vision and create a team that believes in it."
What's next
The roadmap under the Robot.com banner is ambitious in scope: warehouse logistics, gated community delivery, smart city integrations, advertising robots in high-traffic urban environments. The Paris pilot and the Seoul deployment are early signals that this is genuinely a global play, not just a US campus story.
The most interesting bet is the advertising layer. A robot moving at pedestrian speed through a dense urban environment, generating 3,000 advertising impressions per mile, is a new kind of media surface. Felipe didn't set out to build a media company. He set out to cut delivery costs. The two are converging in ways that would have been implausible to explain in a 2017 SkyDeck pitch deck.
What makes this founder's story worth following is not the funding rounds or the delivery count. It's the consistency of the underlying move: find a problem that feels inevitable to most people, decide it doesn't have to be, and build the thing that proves it. He's done it twice. Robot.com suggests he's doing it again - just at a different order of magnitude.