How Chef Robotics found its success by turning away their original customers
Chef Robotics faced death a few years ago.
“There was a lot dark periods when I thought of giving up,” founder Rajat Bhageria told TechCrunch about his six-year old company. But his friends and investors encouraged, so he persisted. Chef Robotics is one of the few food technology robotic companies that are thriving. The startup, which raised $23 million in Series A funding, has 40 employees, and customers such as Amy’s Kitchen, Chef Bombay, and others. Bhageria claims that dozens of robots have been installed across the U.S. and have produced 45 million meals.
Compare this to the graveyard of failed robotics food companies, such as Chowbotics and its salad-making robot Sally, pizza delivery robot Zume, food kiosk robot
Karakuriand, more recent, agtech
Small Robot Company
Bhageria claims he saved his business by doing what early-stage founders are afraid to do: turning down signed customers and millions in revenue.
The grasping Problem
Bhageria’s master’s in robotics began at the University of California, Berkeley.
UPenn’s GRASP Lab. He imagined a sci-fi world where robots would do our housework, mow our lawns, cook us five-star meals, and even mow our lawns.
This world does not exist yet, because engineers haven’t fully solved the robotic dilemma.
grasping problem. It is difficult to train the same robot how to wash a cast iron pot without dropping it, or a wineglass without crushing it.
“Nobody has built a dataset on how to pick up a cheese block without clumping it up or a blueberry without crushing it,” he says.
His initial idea with Chef Robotics is similar to the list of robotics startups who failed: a robotic line designed for fast casual restaurants. This is a huge industry
With a chronic shortage of employees.
We had contracts signed. We had signed contracts worth millions of dollars. We’re no longer doing this. “So what happened?” asked he. “We couldn’t solve the technical issue.”
An employee in these types of businesses completes orders by assembling the various ingredients for each meal. These restaurants want robots that can replicate this process, because it would be more expensive to have dozens or robots calibrated and dedicated to a single ingredient. Some of these ingredients may only be used on occasion (we’re talking about you, anchovies).
But Bhageria’s team was unable to build a robot that could pick up anything because there were no training data. He asked his customers if they would let him install robots to pick up one or two ingredients. He would then collect training data and build from there. They said no.
Bhageria then had an epiphany.
Maybe he didn’t need to go bankrupt trying to satisfy existing customers. He needed new ones. “It was a real drag, because I spent a year and a quarter of my life trying convince these people, fast casual companies, that we should work together,” he said.
Saying no leads you to yes
The fact that fundraising after 2020 was brutal didn’t help. VCs also looked at the graveyard. Bhageria stated that they had spoken to dozens of funds. “We were rejected over and again.”
Bhageria thought of giving up. “You go home and wonder, what are you doing with your life? Am I doing the right thing? “Should I quit?” he asked himself.
He stayed the course and in March 2023 raised $11.2 million in a seed round with Construct Capital. Promus Ventures and Kleiner Perkins also contributed.
Bhageria, along with his team, also found the perfect market in a part of food manufacturing known as “high-mix manufacturing.”
This is a food maker that has many, many recipes and makes thousands of servings but usually as meals or meal tray. Salads, sandwiches, main dishes and side dishes are all examples. These are meals that are used by hospitals and airlines, or frozen food meals that are sold to consumers.
Instead of one employee picking up all the ingredients for each dish, “high-mix” employees form an assembly. Each employee adds his or her individual ingredient to the tray until the order has been completed. They then assemble the recipe.
He says, “It is actually hundreds of people standing in a room at 34 degrees Fahrenheit for eight hours each day. They are essentially scooping out food.” “It’s a horrible job.”
This industry also has chronic labor shortages.
Robotics was not economically feasible in the past due to the wide variety of ingredients. A startup that builds flexible-ingredient robots in partnership with food makers is successful.
Even better, “as soon as we learn how to make this chorizo or we learn peas or we learn this sauce or these zucchinis”, the bots will get the real-world data they need to serve fast-casual restaurant. Bhageria confirms that this is still a part of his roadmap. Bhageria says that, thanks to the renewed interest of VCs in all things AI this time, fundraising was “weirdly easy”.
Avataar Venture Partners was co-founded by former Norwest VC Mohan Kumar. They were specifically looking to fund startups that applied “AI in the real world”, and they pursued Chef Robotics. He closed the round in less than one month. Avataar was the leader, with existing investors Construct Capital and Bloomberg Beta, as well as Promus Ventures, also contributing.
With the new funding, Chef’s total amount raised now stands at $38,8 million. He also signed a loan of $26.75 from Silicon Valley Bank to finance equipment.
This time, the process was “exhilarating,” said he.
