iRobot: Colin Angle — How the Roomba Became a Household Icon

This episode traces the 35-year arc of iRobot, from a three-person MIT robotics lab with no business model to the company that created the consumer robotics category with Roomba, and ultimately to its dissolution after a blocked $1.7 billion Amazon acquisition. Colin Angle tells Guy Raz a story of survival-mode engineering, accidental marketing breakthroughs, and the painful reality that even a world-defining product couldn't protect the company from global competition and regulatory headwinds. The conversation moves from the thrill of building bomb-disarming robots for the military to the absurdity of a doll that supposedly said "bite my butt," capturing both the idealism and the chaos of building a company that never quite knew who its customer would be.

7:25"What Have You Built?" — The MIT Lab That Started Everything

Colin Angle grew up in upstate New York as a self-described "smart jock" who balanced wilderness guiding with a compulsion to build things. At MIT in 1985, he was walking back from class when a fraternity brother mentioned he was going to apply for a summer job at a robotics lab. Angle tagged along, and the application process was revealing: each candidate received a blank piece of paper and was told simply, "What have you built? Just write it on the paper." While other students finished quickly, Angle kept writing for an hour — listing a Lego gun that could fire a piece of track through plywood, a canoe he had built, and countless other projects. He realized, "This might be the place for me."

The lab was run by Rodney Brooks, a pioneer in robotics who would become a co-founder of iRobot. Angle's first assignment was to enable a robot named Seymour to open doors. Rather than building a complex robotic arm, Angle took a characteristically unconventional approach: he mounted a candy machine on the robot and programmed it to offer candy to anyone who would open the door for it. This early lesson — that "robotics isn't a thing, it's a toolkit" — shaped his entire career. Brooks was also building Genghis, an insect-like six-legged robot that could climb rugged terrain using only an 8-bit microprocessor with 256 bytes of RAM (not kilobytes — bytes). When Brooks decided to start a company based on behavior control, Angle immediately said, "I'm in. I'll run it."

12:25No Business Model, No Problem — Living Hand to Mouth

iRobot launched in 1990 with three people — Angle, Brooks, and fellow MIT student Helen Greiner — and no clear application. The company built three types of robots (wheeled, tracked, and legged) and sold them for $5,000 to $10,000 each to companies like Boeing and Mitsubishi, who wanted them for AI research. The business model was primitive: customers paid half upfront, Angle used that money to buy parts, and he built the robots out of bent sheet metal using an origami-like technique he invented. The company did not take venture capital until year eight. For six and a half years, Angle never started a month with enough money in the bank to make payroll.

The dream, as Angle put it, was simple: "We were promised robots." But from the very first day, whenever he told people he built robots, they asked the same question: "When are you going to clean my floor?" They wanted Rosie from *The Jetsons*. Angle would get snarky and ask if they'd pay $5,000 for a robot vacuum, and they'd say no — they'd rather hire a human cleaner. The company was full of engineers who "didn't respect" marketing; the joke was that you couldn't say "marketing" without adding "weenie" immediately afterward. Yet Angle admits that the dream was always to "build the robots for the rest of us" — the military and industrial contracts were just how they paid the bills while waiting for their moment.

18:36The Value-Sharing Model and a Doll That Almost Said Something Terrible

The breakthrough business model came when Angle realized he could work at cost for large companies in exchange for splitting the value created. He would charge nearly nothing for MIT engineers to work on a problem, and if something good came out of it, both sides shared the IP. SC Johnson (Johnson Wax) was a key example: they wanted to develop cleaning chemicals specifically designed for robots, creating a differentiated value proposition for commercial floor cleaning. The robot would use the chemicals, the whole system would save shopping malls money, and SC Johnson would own more of the solution. It never commercialized, but it kept the lights on.

In the late 1990s, iRobot signed a three-year deal with Hasbro to make robotic toys. The first product was "My Real Baby," a doll that used artificial intelligence to detect how it was being played with — rocked, fed, or put to sleep — and respond accordingly. It sold 150,000 units in 1999 and taught iRobot the fundamentals of low-cost manufacturing and consumer products. But it also taught them about the perils of consumer feedback. Early prototypes had eyes that moved left and right instead of opening and closing, which testers found "really creepy." More dramatically, just before launch, a Hasbro executive became convinced that the doll's random babbling was saying "baby, baby, bite my butt." The company had to launch a crisis software strategy to "provably keep My Real Baby from babbling in that particular pattern."

36:10The Cheerios Demo That Changed Everything

The Roomba's origin traces back to an early iRobot employee named Joe Jones, who had built a crude round robot at MIT that could suck up crumpled paper balls. Twelve years later, as the Hasbro contract was ending, Jones approached Angle and said, "Give us 15 grand in parts, and I'll have a prototype in two weeks." The first version was round — designed so it could never drive into a space it couldn't escape — and dragged an electrostatic cloth behind it like a Swiffer. But consumer research revealed a critical insight: people would only pay $40-50 for a cloth-dragging robot. When asked what they'd pay for a vacuum, they said "hundreds of dollars." Angle realized, "Time to make a vacuum."

The engineering challenge was battery efficiency. Angle's team invented a two-step vacuuming system: counter-rotating brushes picked up large debris, while a "squeegee vac" — two rubber blades with a narrow gap between them — used accelerated air velocity to pick up dust, like putting your thumb on a garden hose. This made the Roomba incredibly efficient, but selling it was another matter. Sharper Image backed out. Angle cold-called Brookstone and reached a woman on her first day who hadn't yet been trained not to talk to unsolicited callers. She agreed to see the demo.

Angle's secret weapon was a Ziploc bag of Cheerios. He would sprinkle them on the conference room carpet, stomp and grind them in dramatically, then set the Roomba loose. Cheerios were ideal — bright against dark carpet, visible proof the robot worked on both big chunks and dust. The demo moved from the lowest to the highest level of Brookstone in 40 minutes. They launched at $199. Two weeks later, Brookstone called to order more — as many as iRobot could build. The reason: Brookstone discovered that demoing the Roomba in front of their mall stores drew traffic. Combined with 150 articles from outlets like the *Wall Street Journal* and *Newsweek* — generated by a PR budget of just $5,000 a month — iRobot sold 70,000 Roombas in the first three months of 2002.

44:20The Crash, the Pepsi Miracle, and the Replacement Crisis

Euphoria led to overconfidence. Based on the 70,000-unit quarter, iRobot ordered 300,000 Roombas for 2003. But sales flatlined. The company's first commercial — made "by engineers for engineers" — showed only the robot driving around with the slogan "If it's down there, we'll get it." It failed completely. After Black Friday and Cyber Monday 2003, iRobot had sold only 50,000 units for the entire year and was sitting on 250,000 unsold robots. It was an existential crisis.

Then, inexplicably, sales tripled in a single day. An intern mentioned seeing a commercial. It was a Pepsi ad starring Dave Chappelle. In the spot, Chappelle walks into a beautiful house, grabs a Pepsi and a potato chip, and a Roomba appears. He throws the chip on the ground; the Roomba cleans it up, then locks onto the Pepsi and chases Chappelle. Cut to Chappelle on his back with his pants ripped off, saying, "Your vacuum cleaner ate my pants." The tagline: "Nothing chases thirst like a Pepsi." Pepsi had made the ad without telling iRobot. In the six weeks after it aired, iRobot sold 250,000 Roombas. Angle's epiphany: "We knew nothing about marketing."

But success brought a new crisis. The Roomba had been designed to last 150 hours — the European standard for upright vacuum cleaners, which are used infrequently. But a robot vacuum used half an hour daily would hit that limit in under a year. Customers' Roombas were breaking after six months. iRobot made a painful decision: free replacement for every broken unit, no questions asked. It cost a fortune, but Angle argues it generated more customer loyalty than if the product had never failed. "If a company stands behind its product, you generate more customer loyalty than if that customer never had a problem at all."

55:55The Plateau, Navigation, and the Search for a Second Act

By 2008-2009, iRobot had plateaued. The early adopters were saturated, and the company faced a "significant fallacy" in its technical direction. Angle had believed the perfect Roomba was one you never saw, never touched, and came home to a clean house. But the early majority of customers — the ones needed to scale — didn't trust that. As Angle put it, imagine a cleaning person who says, "You can't talk to me, you can't give me guidance, just trust me." That's creepy, not compelling. Customers wanted control: they wanted to tell the Roomba where to clean, and they wanted confidence it wouldn't knock over a vase.

This drove the shift to systematic navigation, object recognition, and user control — the features that opened the next phase of growth. By 2016, iRobot had sold 10 million Roombas. But the company needed a second act. They tried a mopping robot called Scooba, a lawn mowing robot called Terra, and considered dishwashing, laundry folding, and bed-changing robots. None succeeded. Angle realized that the Roomba itself was becoming a smart home device — it could map rooms and understand where other smart devices were. This led to the idea that combining iRobot's mapping with a voice interface like Alexa could create a new smart home solution. That thinking made Amazon's eventual interest feel natural.

58:00The Amazon Deal, the FTC, and the End

By 2020, iRobot had sold 30 million robots and peaked at over $1.5 billion in revenue in 2021. But the perfect storm was building: US tariffs, COVID shutdowns, shipping costs soaring, component shortages (iRobot was spending $22 on a half-cent resistor just to keep production running), and Chinese competitors finally putting products on the market that customers liked. Market share began dropping.

In 2022, Amazon bid $1.7 billion to acquire iRobot. Angle saw it as a win-win: investors would win, consumers would win, and iRobot employees would "get to be innovative again." But the deal required regulatory approval from both the US Federal Trade Commission and the European Commission. The FTC's argument, as Angle describes it, was that Amazon acquiring another robotics company would create a monopoly — a claim Angle calls nonsensical given Amazon's existing robotics holdings. In 2024, iRobot and Amazon abandoned the deal. Angle says they made the decision when it became clear both regulators were blocking it.

Angle is blunt about the outcome: "The FTC blocked this deal, knowing that they were effectively putting iRobot in a box and handing it to somebody else." He describes being deposed and walking through FTC offices where employees had "badges of honor on the doors" listing deals they had blocked. "The biggest loser was the consumer," he argues, pointing to the hundreds of innovative products that could have come from the merger. Shortly after the deal collapsed, Angle stepped down as CEO after 34 years. Within a year, iRobot was sold to a Chinese company. "We used to lead the consumer robotic floor care industry," Angle says. "We don't anymore. China does. And we chose to give it to them."

Conclusion

What stays with the listener is the tension between the idealism of the early years — "we were promised robots" — and the brutal reality of what it takes to build a lasting company. Angle's story is not a simple rise-and-fall narrative; it's a meditation on how a company can create an entirely new category, dominate it for two decades, and still lose everything to forces beyond its control. The Roomba became a cultural icon — parodied by Dave Chappelle, featured on *Parks and Recreation*, ridden by cats in endless viral videos — but that cultural resonance couldn't protect iRobot from Chinese competition, supply chain chaos, or regulatory decisions that Angle believes were misguided. The episode matters because it shows that even a hero product is not enough: you need a second act, you need to understand marketing (not just engineering), and you need to navigate a geopolitical landscape that can undo decades of work in a matter of months. Angle's final word — "Onward" — captures the resilience of a builder who, even after losing the company he spent 35 years building, is already back in stealth mode, using generative AI to build consumer robots again.

Key takeaways

• iRobot survived its first six and a half years without ever having enough money in the bank to make payroll, funding itself through government contracts and corporate research deals with no venture capital until year eight.
• The company's early business model involved working at cost for large companies in exchange for shared IP on jointly developed products — a risky but necessary approach that kept the lights on.
• The Roomba's round shape was a deliberate engineering choice: it could never drive into a space it couldn't turn around in and escape, eliminating the need for complex navigation.
• The breakthrough marketing moment came not from iRobot but from a Pepsi ad starring Dave Chappelle, which sold 250,000 Roombas in six weeks and taught the company that engineers don't know how to sell to consumers.
• When early Roombas began failing after six months due to a design flaw (150-hour lifespan designed for infrequent upright vacuum use), iRobot offered free replacements to preserve the brand — a decision Angle says generated more loyalty than if the product had never failed.
• The company's attempts to diversify into mopping (Scooba) and lawn mowing (Terra) failed, leaving iRobot dependent on the Roomba as a single hero product.
• The $1.7 billion Amazon acquisition was blocked by US and European regulators, and Angle argues the FTC's antitrust concerns were misguided — the deal's collapse ultimately led to iRobot's technology being sold to a Chinese company.