How the Roomba started a robot revolution: the 25-year story of the world's first household robot

"If you had one of the early Roombas, you probably didn't think it was all that smart," begins The Verge's new Version History podcast. "It bumped around your carpet, got stuck under the couch, stopped when its tiny bin filled up. But you probably still gave it a name." iRobot co-founder and former chief executive Colin Angle joins hosts David Pierce and Jennifer Pattison Tuohy to trace the 25-year arc of how household robotics became a real category.

iRobot was founded in 1990 by Colin Angle, Helen Greiner and Rodney Brooks — then students at the MIT Artificial Intelligence Laboratory. Its first decade was not consumer at all: it built the PackBot for the US military, a tracked robot widely used by US forces in Iraq and Afghanistan for the disposal of improvised explosive devices.

The Roomba grew out of a hard constraint of military-robot design: "If you minimise algorithmic complexity, you minimise hardware cost," Angle says. The random-walk plus wall-following algorithms developed for PackBot were good enough for a vacuum. The change needed was a suction system and a much lower power draw — and a chassis low enough to fit under a sofa.

The first Roomba — a million-unit production run in disc shape — launched in autumn 2002 on Amazon and at big-box retailers for $199. Counter to expectations, 50,000 sold during the first Christmas season. Industry analysts had been writing that there was no consumer household-robot market; iRobot's response was to be the first market itself.

A key technical advance arrived in 2004: visual SLAM — camera-based simultaneous localisation and mapping — let the robot build a map of the house and clean in systematic stripes rather than wandering randomly. The 980 model in 2015 was the first fully mapping Roomba; the i7+ in 2018 added an automatic self-emptying base.

With complexity came price. The first Roomba cost $199; the j9+ Combo Auto-Fill launched in 2024 sits at $1,399. Like phones and tablets before it, the category developed a deeply premium tier while keeping an affordable entry level, with very different margin economics at each end.

Competition has hardened. Chinese makers — Roborock, Ecovacs, Dreame, Eufy — have pushed aggressively into Western markets over the past five years, riding scale advantages from the Chinese domestic market. They have rapidly squeezed iRobot's middle segment with lidar-based navigation, mopping functions and aggressive $600-1,000 pricing.

iRobot's position deteriorated in 2024. Amazon's planned $1.7 billion acquisition was abandoned in January 2024 in the face of EU antitrust scrutiny. The collapse was severe: the stock fell about 80 per cent, the company laid off some 350 staff, and Glenn Murphy took over as interim chief executive in 2024 before Tom Pierce stepped in in February 2025.

Angle is candid about the difficulty: "When we built the Roomba, if it had sold 50,000 units a year we would have called it a success. It has now sold more than 50 million. But the category we created has become the field where Chinese competitors are now attacking while we sleep." He argues the Roomba's real legacy is not the unit success but the normalisation of the very idea of a household robot.

The category keeps expanding. Robotic lawnmowers, robotic window-cleaners, robotic cooks, robotic gardeners — each is following the Roomba's "one narrow task at home, simple intelligence" template. Husqvarna's Automower, Eve's Aqua pool drone and various robotic cool-room sweepers fit the pattern. Without the Roomba, consumers would treat all of these with much greater scepticism. Angle's summary: "Once a robot has come inside your house, it does not really go back out again."