SerpApi.Org
The two Apple veterans behind Weave are hoping their $7,999 robot is helpful in homes, even if it can only do two things.
When Isaac 1 finishes tidying a room, it lowers itself from human height to around three feet, closes a set of physical shutters over its cameras, and rolls behind an upholstered partition that matches its fabric panels. A machine capable of making beds and folding laundry spends most of its day trying to look like a piece of furniture.
Weave Robotics, the San Francisco-based startup behind it, opened pre-orders for Isaac 1 in early July at $7,999 upfront or $449 a month, with first deliveries planned for California homes this fall and the rest of the United States in 2027. It was founded in 2024 by two Apple veterans: Kaan Dogrusoz, who worked on machine learning and robotics, and Evan Wineland, a former lead AI product manager on Apple Intelligence. Their robot arrives in an industry that has spent the past three years—and several billion dollars of venture capital—working to recreate the human body.
Weave’s competitors include Tesla’s Optimus, which is being developed in its factories, with the home discussed as an eventual destination. Figure has deployed its humanoids in manufacturing environments, including work with BMW. And 1X, the closest thing Weave has to a direct competitor, opened pre-orders for Neo at the end of 2025—it’s a $20,000 bipedal humanoid with five-fingered hands, with American deliveries promised this year.
Goldman Sachs projects the humanoid market could reach $38 billion by 2035, with 1.4 million units shipped. Morgan Stanley has sketched a scenario in which the broader humanoid economy passes $5 trillion by 2050.
So far, demos powering that enthusiasm tend to involve walking, dancing, and backflips: proof of what’s technically possible. Wineland’s argument is that the home rewards something else entirely. “Since our founding, we’ve felt that to successfully bring robots into the world, they shouldn’t feel like they require accommodation from us—they should fit naturally into our lives and spaces,” he tells me.
The cost of looking human
That design philosophy is as much about economics as aesthetics, and it explains most of what makes Isaac 1 unusual. The robot moves on wheels because most homes have flat floors, and legs are among the most expensive problems in robotics due to balance, falls, and the actuators needed to prevent both.
The trade-off is stairs. And how will it do on carpet? Wineland says Isaac 1 “navigates confidently across rugs, carpets, hardwood, tile, and the transitions and bumps between them” because Weave designed its wheel modules in-house. Whether the simpler base proves more dependable in real homes than legs or thicker rugs prove it wrong is something the deliveries this fall will start to settle. Its hands are simple grippers rather than articulated fingers, because folding a shirt does not require the same dexterity as playing the piano.
Those hands point at the deeper split between Weave and its rivals. Neo is pitched as a general assistant, a machine that can, in principle, be taught anything a person does around the house. Isaac 1 launches with exactly two capabilities, handling laundry and a daily tidying routine, and Wineland treats the narrowness as the strategy. “We don’t believe the path to a general home robot starts by trying to do everything, or tasks that people actually enjoy doing,” he says. “It starts by solving a few valuable, tedious problems exceptionally well, learning from experience in real homes, and expanding capability over time.”
Isaac 1’s whole body runs on 21 degrees of freedom, a fraction of what more humanlike rivals carry. It can reach 80 inches (over 6 feet) vertically, high enough for the top shelf. “Every motor, sensor, degree of freedom, and feature must earn its place by making the robot meaningfully better at valuable work,” Wineland says. “Any company adding unnecessary complexity is adding cost and difficulty on top of an already extremely hard problem.”
The same discipline runs down into the supply chain. Weave designed its own actuators and safety systems rather than assembling off-the-shelf parts because, as Wineland puts it, “if too many critical components come from somewhere else, every improvement depends on someone else’s roadmap, costs, and lead times.”
All of that subtraction is what lets Wave sell Isaac 1 for less than half the price of 1X’s Neo, a machine that recreates far more of the human body. And the parts of Isaac that the designers kept are there for reasons that have little to do with engineering. The fabric shells come in different colors and can be swapped or removed, a choice with numbers behind it. In a November 2025 survey by the consultancy Altman Solon, softer, cushioned robots edged out full-size humanoids as the form Americans most wanted in their homes.
The camera shutters exist so that anyone in the room can easily tell whether the robot is sensing anything at all, with no trust in an on-screen privacy indicator required. “A mobile robot earns trust by behaving predictably,” Wineland says. “It should do what you expect when you ask it to and make its state obvious.”
He says showing the robot to families taught the team things engineers would have missed. “Non-roboticists react to things that might have otherwise been overlooked—not just whether a robot is safe but whether it feels safe, whether kids are comfortable around it, whether it could fit in a non-enthusiast’s home, whether it’s cute.”
What the robot vacuum has already proved
The appetite for all of this is unusually well-documented for a product category that barely exists. The International Federation of Robotics counted close to 20 million consumer service robots sold in 2024, up 11 percent year-over-year, and robots for domestic tasks such as floor cleaning made up 97 percent of them.
Consumers have spent two decades proving they will pay machines to do chores. “Robot vacuums proved something incredibly important: People will gladly adopt a robot if it reliably removes a recurring household chore,” Wineland says. The academic record supports his emphasis on reliability. Researchers at Ben-Gurion University who classified 10,072 Amazon reviews of domestic robots found that failures of task completion and robustness did the most damage to customer ratings, far more than flaws in how the robot interacted.
In that same Altman Solon survey, 65 percent of Americans said they were interested in owning an advanced home robot but nearly 70 percent also said they would pay no more than $5,000 for one, a ceiling that sits uncomfortably below both Isaac 1 and Neo.
There’s a respectable argument that the general-purpose machine wins this market in the end. A household that solves its chores one robot at a time ends up with a mower in the shed, a vacuum under the sofa, and a laundry robot by the washing machine, each with its own app, dock, and failure modes. One capable humanoid replaces the fleet, and that is exactly the product 1X is building toward, with real engineering conviction behind it.
The public, for now, leans the other way. When University of Washington researchers surveyed people on humanoids in the home, participants generally preferred special-purpose robots as safer, more private, and more comfortable to live with, though most still found humanoids acceptable, and some preferred them for tasks no dedicated robot yet handles.
Wineland’s answer is that the fleet and the generalist are a false choice. “Isaac 1 isn’t a robot built for one task,” he said. “It’s a general-purpose home embodiment designed to perform many different household tasks over time. That means future capabilities are increasingly a software problem rather than requiring customers to buy an entirely new robot.”
That pitch carries a familiar catch: A $7,999 purchase made today rests partly on software that doesn’t exist yet, and anyone who’s bought hardware on the promise of future updates knows how unevenly those promises age. It also, for now, remains a pitch. Neo and Isaac 1 are both pre-order products, with deliveries promised rather than made.
A vision that already works
Buyers punish a robot far more severely for failing at its job than for being awkward, which is the strongest case for shipping two chores that work rather than 20 that mostly do. “We don’t think ‘good enough’ means attempting everything imperfectly,” Wineland says.
Weave has some evidence its version works. The company has been shipping Isaac 0, a stationary laundry-folding robot, to California homes and businesses since February. It says the fleet has logged more than 2,000 field hours and folds over 1,000 pounds of laundry a week; figures, it’s worth noting, that are Weave’s own. Wineland recalls one customer, a mother in a family of five, telling him the robot had her household caught up on laundry for the first time in a year.
Behind the autonomy sits a familiar asterisk. Isaac 1 is autonomous by default, but remote human operators step in when it gets stuck, the same arrangement Neo depends on. In 2026, even the robots built to disappear occasionally need a person watching through their eyes. Weave says operators see only what is necessary for the task the customer has requested, and only when they have requested it.
Pressed on the mechanics, how the queue works, and how many operators the model demands as thousands of robots ship, the company’s answers are noticeably vaguer than its answers about fabric panels. Wineland is unapologetic about shipping before the autonomy is finished. “Robots are, in a sense, like any other product: to matter, they have to ship,” he says. Real homes generate the failures and recoveries that training videos cannot.
The first Isaac 1s are due in California living rooms this fall. The hardware, Wineland notes, “is physically capable of more than its initial software and model enables today,” so the robot that arrives may be less than the robot its owners eventually live with.
What it should never become is louder. “A robot should not demand attention when it is not actively working for you,” he says. In a category that’s spent three years auditioning for the world’s attention, Weave is betting on the machine that knows when to disappear.
This post originally appeared at inc.com.
“Click here to subscribe to the Inc. newsletter: inc.com/newsletters"
