Colin Angle's New Robot Startup: From Roombas to AI Companions

Colin Angle built iRobot and spent three decades running the company behind the Roomba vacuum. When Amazon's deal to buy iRobot fell apart, he moved on to something different. Now he has unveiled Familiar Machines & Magic, a startup building home robots designed to be companions rather than just cleaning tools.

The company showed its first prototype at The Wall Street Journal's Future of Everything conference. The robot, called Familiar, is dog-sized with moving eyebrows, ears, and expressive eyes. It looks like a cross between a bear, barn owl, and golden retriever, and it can move around your home on its own. The key difference from his Roomba: it's built to talk with you and develop what the company calls distinct personalities over time.

A Sharp Change in Direction

Roomba solved a specific problem — vacuuming — with minimal conversation needed. Familiar Machines & Magic is aiming at something broader. The company describes its robot as a "physically embodied AI system designed to perceive, adapt, and interact with people in ways that feel natural."

The startup is based in Woburn, Massachusetts. Angle brought along several former colleagues from iRobot, including his old CTO Chris Jones and engineer Ira Renfrew, who had been working on Amazon's now-cancelled Scout delivery robot. The team also includes software engineer Colin Soguero and creative director Morgan Pope, according to AP News.

How It Works: What We Know

The robot runs AI software directly on its own hardware rather than relying on cloud servers — think of it like having a smart assistant that doesn't need to phone home to think. This approach also addresses privacy concerns. When a device sends all your home data to the cloud, you lose control over where that data goes. By processing information locally, the Familiar avoids that problem.

The robot combines several established technologies: sensors that help it navigate and understand what's around it, motors and joints that let it move and express emotions through its face, and a generative AI model (similar to ChatGPT, but smaller and running on the device itself) that lets it have conversations and learn your preferences over time.

The company is described as "well-funded" and is seeking to raise $30 million in additional capital, TechCrunch reported. That amount reflects the real costs of building physical robots — hardware development, manufacturing, testing, and supply chain work is expensive compared to pure software startups.

Why This Matters: Patterns in Tech

Across technology history, successful people often take what they learned in one era and apply it to the next. Angle built a company in the era of mechanical automation — making smart devices that did one job well. Now he is betting that combining advanced AI with robotics can create something new: machines that feel like companions rather than tools.

Previous attempts at companion robots haven't taken off commercially. Sony's AIBO robot dog has existed since the early 2000s, but it costs thousands of dollars, so few people own one. Boston Dynamics created impressive robots that can move and climb, but focused on commercial uses. Startups like Anki and Jibo tried to build affordable companion robots but ran into problems with cost, technical challenges, or simply poor timing. The market just wasn't ready.

The timing may finally be different. AI has improved dramatically in the past few years. Battery technology is better. Robot manufacturing has become cheaper. At the same time, people have grown comfortable talking to AI assistants like Siri and Alexa, and using smart home devices. That cultural shift might create an audience for a robot that actually lives in your home and engages with you over time.

The target markets the company mentioned — companionship, entertainment, eldercare, parental support — also align with real demographic trends. Aging populations and increased attention to mental health mean there may be genuine demand for machines that provide interaction and engagement.

The Real Challenges Ahead

The harder part isn't robotics engineering. Angle's team knows how to build robots that move and work reliably. The harder part is behavioral design — making the robot feel like a real companion, not just a cute toy.

Unlike a vacuum, where success is simple (clean floor or no clean floor), a companion robot has to navigate complicated social interactions. It needs to remember things about you, adapt its personality to match your moods, and do this consistently over months or years. It needs to justify your emotional investment in it.

There are practical obstacles too. Sophisticated robots with moving parts, sensors, and batteries cost a lot to manufacture. Consumer willingness to spend $3,000, $5,000, or more on a robot companion has not been proven. Battery life, durability, and keeping the hardware reliable in a home environment all present engineering problems.

Data and privacy also matter. A robot living in your home sees and hears more than any other device. The company's choice to process information locally is a good sign, but we will need to see exactly how the robot handles data, who owns what the robot learns about you, and what happens if the company goes out of business.

The broader context here involves significant technical and market challenges that differ from the problems Angle solved with the Roomba. Manufacturing discipline and engineering excellence got him where he is. Whether behavioral AI design and sustainable business models for ongoing service delivery will carry him further remains to be seen.

What Happens Next

Familiar Machines & Magic is an early test of whether today's AI technology and robotics can finally deliver on promises people have heard for decades: helpful home robots that do more than one specific chore.

If this venture works, it could validate an entirely new category of consumer products — physical AI companions that blend robot bodies with generative intelligence. That success would likely influence how other companies invest in robotics and what features they prioritize.

For the robotics industry more broadly, Angle's shift from task-focused robots to companion robots might signal something important: that the fundamental engineering problems of robot movement, navigation, and reliability are solved enough that attention can move to interaction design. That could reshape where money and talent flow across the sector.

As the prototype moves into development and eventually the market, it will provide real data on whether consumers actually want this category of product, and what technical hurdles remain. We've seen this pattern before, across multiple computing shifts, where the most experienced players from an earlier era have applied their expertise to the next one. Some transitions succeed cleanly. Others teach hard lessons about what has truly changed and what has not.