For decades, the vision of a helpful robot quietly taking care of everyday household chores has lived somewhere between science fiction and carefully staged tech demos.

But a robotics startup called Sunday believes that moment may finally be approaching.

The Mountain View–based company has just raised $165 million in a Series B funding round, bringing its valuation to $1.15 billion. The funding will accelerate the development and deployment of Sunday HomeBot, the company’s autonomous household robot designed to operate inside real homes rather than laboratory environments.

Unlike many robotics projects that remain confined to controlled demonstrations, Sunday is preparing to send its robots directly into everyday living spaces.

And that decision could be the key to solving one of robotics’ biggest unsolved challenges.

The Hardest Environment for Robots

Factories are predictable. Warehouses are structured.

Homes are neither.

Every apartment, house and living space is filled with unpredictable layouts, cluttered surfaces, moving humans, pets and constantly changing environments.

That complexity has long prevented robots from operating reliably in domestic settings.

Sunday’s strategy is simple but ambitious: teach robots directly from real human behavior inside real homes.

Instead of relying solely on synthetic training data or simulated environments, the company has built a system that collects large volumes of real-world data and uses it to improve its AI models.

At the center of this approach is a training device known as the Skill Capture Glove.

Teaching Robots by Watching Humans

The Skill Capture Glove allows humans to record their own movements while performing everyday tasks.

These movements are translated into data that can be used to train robotic models more efficiently.

Tasks such as picking up objects, loading dishwashers, folding laundry or organizing surfaces become training episodes for the robot’s learning system.

According to Sunday, the system has already collected tens of millions of movement examples.

That data feeds directly into what the company calls a robotics data flywheel.

The more robots are deployed, the more data they collect.

The more data collected, the better the AI models become.

And the better the models, the more useful the robots become in real homes.

A Full-Stack Robotics Approach

Sunday’s founders, Tony Zhao and Cheng Chi, are researchers known for their work in robotics AI systems, including projects such as ALOHA and diffusion-based learning techniques.

Their philosophy is that robotics cannot succeed with isolated breakthroughs in either hardware or software.

Instead, the company is building a full-stack robotics platform combining:

• robotic hardware
• perception and navigation systems
• machine learning models
• data collection pipelines
• cloud infrastructure

This approach allows Sunday to iterate rapidly.

In one recent robotics demonstration, the company reported major improvements in robotic capabilities achieved within just three months of training cycles.

From Demos to Real Deployment

Perhaps the most striking part of Sunday’s strategy is its commitment to real-world deployment.

The company is launching a Beta program later this year, sending early versions of the HomeBot robot into selected households.

Thousands of applicants have already signed up.

These early deployments will provide valuable operational feedback and generate even more real-world data.

CEO Tony Zhao describes the shift as a turning point for the company.

“We raised our Series B to stop giving demos,” Zhao said. “Now we’re focusing entirely on deployment.”

For investors, the strategy signals a bold attempt to solve robotics’ long-standing data problem.

Thomas Laffont, co-founder of Coatue, the lead investor in the round, believes the company’s speed of development could give it an edge.

“Sunday’s velocity is the best signal we have that they will be the first to ship truly helpful autonomous home robots at scale,” he said.

The Race for the Robot Home

The dream of a robot assistant has attracted enormous interest from both startups and tech giants.

Companies like Tesla, Amazon and several robotics startups are exploring similar ideas.

But success in the home environment may ultimately depend on one factor above all: data from real human environments.

By turning actual homes into training grounds for robotic intelligence, Sunday hopes to build the first truly useful household robot.

If the strategy works, Sunday HomeBot may become one of the first robots to move from tech demo to everyday appliance.

And when that happens, the idea of a robot helping with daily chores may finally become reality.

‍