# Why Are Billion-Dollar Humanoid Startups All Starting With Laundry?

Three well-funded humanoid robotics companies — [Sunday Robotics](https://humanoidintel.ai/companies/sunday-robot), Weave Robotics, and [1X Technologies](https://humanoidintel.ai/companies/1x-technologies) — have independently converged on the same first task: folding clothes. Together with [Figure AI](https://humanoidintel.ai/companies/figure-ai), which described laundry-folding in a blog post as "one of the most challenging [dexterous manipulation](https://humanoidintel.ai/glossary/dexterous-manipulation) tasks for a humanoid robot," these companies collectively represent billions of dollars in venture capital — and they've all chosen your laundry hamper as their beachhead into the home.

The strategic logic is clear: laundry is hard enough to be technically meaningful, safe enough that failure carries no real consequences, and universal enough to anchor a consumer value proposition. When a robot drops a sock, nothing breaks. That asymmetry — high technical signal, low failure cost — makes it an ideal benchmark for a generation of humanoids that need to prove themselves outside the lab before the end of 2026.

Sunday plans to deploy its Memo robot into homes through a beta program this fall. Weave Robotics and 1X both expect to begin customer deliveries later this year.

---

## The Technical Case: Deformable Objects Are a Genuine Research Problem

Laundry isn't a PR stunt. Handling deformable objects — garments that bunch, fold, slip, and wrinkle unpredictably — remains one of the genuinely unsolved problems in robotic manipulation. Roboticist and Spelman College president Ayanna Howard has described teaching a robot to fold clothing as a longstanding "holy grail" of the field.

The challenge is multi-layered. Unlike a rigid box or a metal component, a t-shirt presents a different configuration every single time it's picked up. The robot's perception system must classify the garment's current state, plan a manipulation sequence, and execute it with sufficient precision — all while the object's geometry is actively changing in the gripper. This is a stress test for [physical AI](https://humanoidintel.ai/glossary/physical-ai) stacks in a way that picking a fixed-shape item simply is not.

That's also why previous attempts failed on technical grounds before they failed commercially. The $16,000 Laundroid, made by a Japanese company, filed for bankruptcy in 2019. FoldiMate generated repeated Consumer Electronics Show coverage in the late 2010s and never shipped. Both were undone by high costs and poor reliability — not lack of demand.

Weave Robotics CEO Kaan Dogrusoz told Business Insider the difference this time is the convergence of better AI, cheaper hardware, and an influx of capital and talent into physical AI. "For the longest time, the technology just wasn't there," he said. "But then, everything changed."

That claim deserves scrutiny. Today's systems are demonstrably faster than their predecessors — but per the source reporting, a single garment can still take minutes to fold, and unfamiliar clothing items can leave the robots stumped. Zero-shot generalization across the full distribution of consumer garments — different fabrics, sizes, and initial configurations — remains an open problem. The beta deployments this fall will be the first real-world data point on whether lab performance translates.

---

## The Strategic Case: Bounded Task, Unstructured Environment

Columbia University mechanical engineering professor Matei Ciocarlie framed the home environment bluntly: "The home is the ultimate unstructured environment." Pets, toddlers, irregular floor plans, variable lighting — every variable that makes factory-floor robotics tractable disappears.

Laundry sidesteps most of that complexity. Sunday CEO Tony Zhao told Business Insider the company "looks for hard things to do" — but laundry also happens to be a spatially bounded task. A robot can stay in one corner while a hamper is brought to it, eliminating the need for whole-body navigation through a cluttered living space. Weave's Dogrusoz cited this containment as a core reason his company chose it as a first task.

This is a deliberate sequencing strategy, not a limitation. The near-term goal isn't a robot that roams a home autonomously; it's a robot that does one useful thing reliably enough that consumers will pay for it and generate the real-world training data needed to expand the task set. Laundry is the wedge.

The commercial framing also matters for investor relations. Consumers don't need to understand vision-language-action models or sim-to-real transfer gaps to grasp the value of a machine that converts a crumpled pile of clothes into a neat stack. That legibility — immediate, domestic, emotionally resonant — is worth something in a category that has historically struggled to close the gap between lab demos and consumer trust.

---

## What the Convergence Signals for the Industry

The fact that Sunday, Weave, and 1X arrived at the same first task independently is itself a signal. It suggests the field has quietly converged on a shared theory of how to enter the home market: start with a task that is technically hard but operationally contained, establish a reliability baseline outside the lab, then expand. That's a more disciplined go-to-market thesis than the "general-purpose household robot" framing that plagued earlier generations of home robotics companies.

The risk is that the beta programs reveal the same gap that sank Laundroid and FoldiMate: the distance between a controlled demo and consistent performance in real homes with real garments is larger than the lab data suggests. Per the source, reliability in actual homes "remains unknown because few have been deployed." That caveat carries real weight heading into fall beta programs.

If the deployments hold up, laundry becomes the data flywheel — each folded garment generates manipulation training data that improves performance on the next unfamiliar item. If they don't, the industry will have run the same hype cycle for a third time.

---

## Key Takeaways

- **Sunday Robotics, Weave Robotics, and 1X Technologies** have all independently chosen laundry folding as their primary home-use case, alongside Figure AI.
- **Laundry is technically demanding** — handling deformable objects is a recognized open problem in dexterous manipulation — but low-stakes enough that failure doesn't erode consumer trust catastrophically.
- **Previous attempts failed:** Laundroid's manufacturer filed for bankruptcy in 2019; FoldiMate never shipped despite years of trade show coverage.
- **Current limitations are real:** single-item fold times can run to minutes, and unfamiliar garments can still stump today's systems.
- **Deployment timelines are imminent:** Sunday's beta program is targeting fall 2026; Weave and 1X expect customer deliveries before year-end.
- **The convergence on laundry** reflects a shared industry theory: use a bounded, high-difficulty task to establish reliability outside the lab before expanding to whole-home navigation.

---

## Frequently Asked Questions

**Why are so many humanoid robot companies starting with laundry folding?**
Laundry is strategically ideal as a first task: it's technically challenging (deformable objects are a genuine unsolved problem in robotic manipulation), spatially bounded (a robot can stay in one corner rather than navigating an entire home), low-stakes on failure, and universally relatable to consumers. Sunday Robotics CEO Tony Zhao confirmed to Business Insider that difficulty was a deliberate selection criterion.

**Have laundry-folding robots been tried before?**
Yes, and they failed. Laundroid, a $16,000 Japanese laundry-folding robot, saw its manufacturer file for bankruptcy in 2019. FoldiMate generated significant Consumer Electronics Show attention in the late 2010s but never delivered a product. Both were undone by high costs and poor reliability. Today's companies argue that advances in AI and cheaper hardware make this generation different — a claim that will be tested in real homes later in 2026.

**When will these robots be available to consumers?**
Sunday Robotics plans a home beta program for fall 2026. Weave Robotics and 1X Technologies both expect to begin delivering robots to customers later in 2026, per Business Insider's reporting.

**How fast can current humanoid robots fold laundry?**
Today's systems are faster than their predecessors, but per the source reporting, a single garment can still take minutes to fold. Unfamiliar clothing items can also leave the machines unable to complete the task — meaning zero-shot generalization across the full range of consumer garments remains an open technical problem.

**What's the broader significance of humanoid robots starting with home tasks?**
Laundry is positioned as a data-collection wedge, not an end goal. Consistent performance on a bounded domestic task generates real-world training data that can improve capability across a wider range of home tasks over time. The underlying bet is that the home market — not warehouses or factories — is where humanoid robots achieve scale, and laundry is the entry point.