The Sunday Robotics home robot Memo has achieved a 99.1% success rate folding laundry in homes it had never entered before, across garments it had not been specifically trained on, the Bay Area startup disclosed on Thursday alongside a new AI model called ACT-2.
‘I think 2026 will be when the timeline for autonomous home robots shrinks in a very drastic way because of this breakthrough,’ CEO and co-founder Tony Zhao told Business Insider.
What Sunday Robotics’ Home Robot Can Actually Do
ACT-2 is the model that now powers Memo, Sunday’s wheeled home robot. According to Sunday’s ACT-2 blog, the system was tested across garments ranging from baby clothes measuring 16 inches by 8 inches to 8XL oversized shirts at 38 inches by 42 inches, and large towels at 53 inches by 28 inches. Memo was able to reposition itself, adjust its height, and lean into the workspace to handle that variation.
The core advance is generalisation. A robot may master a task in a controlled setting and falter in a new room or when handed an unfamiliar object. Sunday says ACT-2 carries what it learns in one environment into the next, potentially removing the need to retrain Memo for every household.
Training happens in two stages. Memo first learns basic physical skills from people wearing sensor-equipped gloves developed in-house, which cost around $200 to make and mirror the shape and sensor layout of Memo’s own hands. A fleet of in-house robots then lets the model learn from its mistakes, with new skills refined within minutes. ‘Memo’s performance in our office almost perfectly transfers to its performance in the wild,’ Zhao said. The company has spent several months testing in employees’ homes and Airbnbs.
The ACT-2 blog describes the underlying method as ‘a repeatable scaling recipe: improve with limited in-house data, transfer the gains broadly, and build each new capability on a stronger foundation.’ Sunday considers laundry folding Memo’s first fully reliable capability under its own proposed standard, which it calls a ‘Solve.’ The framework would require companies to state clearly what a robot can do, where it was tested, and how much extra training or human assistance it needs in each new setting.
‘Each demo is a different robot, setting, and object, and it’s really hard to figure out if we’re progressing,’ Zhao said. ‘It’s like apples-to-oranges comparisons.’
An earlier model, ACT-1, demonstrated clearing a dinner table, loading a dishwasher, and pulling a shot of espresso. Sunday Robotics notes on its official website that in December 2024 Memo had only one arm and was learning to arrange shoes; by October 2025 it had progressed to folding socks, handling glassware, and making espresso. The company now says ACT-2 is learning to vacuum, organise toys, fasten zippers, and make coffee, though none of those tasks are yet reliable enough to qualify as a Solve.
Funding, the Beta Launch, and a Crowded Field
Sunday raised $165 million at a $1.15 billion valuation earlier this year. Sunday’s Series B announcement, dated 12 March 2026, described the round as oversubscribed and led by Coatue, with Coatue co-founder Thomas Laffont joining the board. Bain Capital Ventures, Fidelity Management & Research Company, Tiger Global, Benchmark, Conviction, and Xtal Ventures also participated. The announcement stated the capital would enable deployment to homes in 2026, with the beta programme intended to scale to thousands of households.
Conviction’s Sarah Guo, who wrote Sunday’s first cheque, previously told Business Insider that she backed the founders because they were ‘completely cracked researchers’ with ‘a real commitment to building products that got deployed and an Elon-style religion on delivering a mass-consumer product.’ Zhao left Stanford’s computer science PhD programme to build Sunday; co-founder Cheng Chi joined him in a Silicon Valley hacker house in 2024. The company has grown to more than 100 employees.
Sunday plans to place Memo in homes through a beta programme this autumn, though Zhao declined to say how many households would participate at launch. Memo will operate autonomously, with a remote operator stepping in only when customers request help, in a model Sunday compares to the remote assistance used by Waymo’s self-driving cars. Sunday says it will not use those interventions to collect training data from customers’ homes. ‘Being fully autonomous is the best way to build trust,’ Zhao said. ‘We are not trying to creep in a teleoperator into your home to collect your data so that we can get our model to be better.’
Several rivals are closing in. Y Combinator-backed Weave Robotics plans to begin shipping its Isaac 1 robot in California this autumn at $7,999 upfront, or $449 a month. The Next Web reports Isaac 1 is non-humanoid, rolls on a wheeled base, rises to 5 ft 9 in when in use, and runs for approximately eight hours per charge. 1X aims to begin delivering its $20,000 NEO humanoid around the same time, while Tesla is working toward placing Optimus in homes.
Sunday’s proposition rests on whether the 99.1% figure holds once the beta scales beyond controlled tests. The autumn launch will be the first real signal.