Get ready for a game-changer in the world of robotics! The future of household chores is here, and it’s not what you’d expect.
Sunday Robotics, a forward-thinking startup, is about to revolutionize the way we tackle daily tasks. Imagine a robot, named Memo, that can make your morning espresso, clear tables, and even load the dishwasher – all without breaking a sweat (or a cup!).
But here’s where it gets controversial: Memo isn’t just any robot. It’s a fully autonomous, humanoid-like machine with a unique ability to adapt and learn. And it’s about to enter our homes, changing the way we live and work.
Memo’s design is inspired by the beloved Wall-E, with a sleek white body, two arms, and a friendly face. Instead of legs, it moves on a wheeled platform, adjusting its height with ease. This robot’s movements are a sight to behold, especially when it’s making coffee.
Tony Zhao, CEO of Sunday Robotics, envisions a future where robots free us from mundane chores. “We want to build robots that take care of the laundry, the dishes, and all the tedious tasks,” he says. But making a cup of espresso is no small feat for a robot, especially in a real kitchen.
It requires an incredible level of dexterity and understanding. Memo has to identify objects, grasp them precisely, and use them correctly. And Sunday Robotics is not just building the hardware; they’re also training the models that make this possible. “We believe the key to a successful home robot is full-stack integration,” Zhao explains, “and that’s an ambitious goal.”
Most robots today are confined to controlled environments, performing precise, repetitive tasks. Unlike humans, they struggle with adapting to changes or unfamiliar situations. But Sunday Robotics is pushing the boundaries, aiming to create robots that can operate in the messy, unpredictable environment of a real home.
And this is the part most people miss: robot demos can be misleading. The real test is how well Memo performs in various homes without Sunday’s engineers. Can it adapt to different kitchens, with their unique layouts and challenges?
I witnessed Memo’s skills first-hand. Besides making coffee, it cleared glasses from a table and loaded them into a dishwasher. The most impressive part? Figuring out how to grasp two glasses in one hand. Memo’s dexterity is a result of Sunday’s innovative training method, which involves remote workers wearing special gloves that mimic Memo’s hands.
Ken Goldberg, a roboticist at UC Berkeley, calls this approach “a beautiful design and a smarter kind of data capture.” He believes Sunday’s variant on home robots is exciting and unique.
The optimism surrounding home robots is sky-high, and for good reason. Researchers have demonstrated that robots can tap into large language models, like those powering today’s chatbots, to respond to commands and understand scenes. Some even hope to develop a more general robotic intelligence by gathering vast amounts of data on different actions.
Zhao and his co-founder, Cheng Chi, have contributed significantly to this field. Zhao worked on Mobile ALOHA, a project that trained robots using a low-cost teleoperation system, while Chi was part of a team that developed a claw-like device to gather data from humans doing tasks like cleaning dishes.
“The most powerful AIs are trained on the entire internet,” Zhao says. “In robotics, we don’t have that luxury.”
Several startups are racing to develop capable robots for ordinary homes. Physical Intelligence, Skild, Generalist, and 1x are all working on robot models that can adapt to new situations using similar approaches. But Sunday Robotics stands out with its practical, full-stack approach, backed by veterans from Tesla and Google DeepMind.
“Tony and Cheng are incredibly talented,” says Sarah Guo, founder of Conviction. “They’ve assembled an all-star team capable of unique hardware-AI co-design and full-stack product delivery.”
Eric Vishria, a partner at Benchmark, agrees. “The promise of AI robotics is not backflips or dancing demos, but robots that work in messy, real-world situations. Sunday’s breakthroughs mark the start of an exponential curve towards a future where robots are an everyday reality.”
Sunday plans to give Memo to beta testers next year. The pilot program will reveal how people adapt to having a home robot, albeit one that might be slow and imperfect at times. A key question is how well Memo can handle the chaos of real homes, with kids, pets, and mess.
After beta testing, Sunday will roll Memo out to early adopters. Just like early home computers, Memo might initially appeal to enthusiasts willing to tolerate some imperfections. Users might even teach their robots new tricks. “I believe people should be able to teach their own robots,” Zhao says.
The era of truly capable home robots might be closer than we think. For now, I’ll settle for a perfectly brewed espresso, served by a friendly robot.
Updated 12:23 pm ET, November 19, 2025: To clarify, the gloves used for robotic training data collection cost approximately $400 per pair.
Updated 1:26 pm ET, November 19, 2025: Added comments from Sarah Guo and additional details about the Sunday team.
Starlink devices are allegedly being used at scam compounds.
Will Knight is a senior writer for WIRED, covering artificial intelligence. He writes the AI Lab newsletter, a weekly glimpse beyond the cutting edge of AI. Previously, he was a senior editor at MIT Technology Review, where he wrote about fundamental advances in AI and China’s AI scene.