Norwegian robotics company 1X Technologies opened preorders for NEO, a humanoid robot designed for household use, at $20,000 for outright ownership or $499 per month on a subscription plan. The company says first units will ship to U.S. homes in 2026. On the same day preorders went live, 1X announced a new American factory with a target of producing 10,000 robots in its first year, setting up a direct test of whether consumers will pay a premium for a machine that has yet to prove itself in real kitchens and living rooms.
Why the $20,000 price tag and $499 subscription split matters right now
The two-tier pricing structure creates a clear financial fork for early buyers. A household that chooses the $499 monthly subscription would spend roughly $6,000 in the first year, well below the $20,000 ownership price. But over 40 months the subscription cost would exceed the purchase price, meaning the math depends heavily on how long a buyer expects to keep the robot and how fast the hardware improves.
That calculation gets more complicated once maintenance and software updates enter the picture. The subscription model shifts repair and upgrade risk to 1X, while outright owners absorb those costs themselves. If a subscription includes ongoing service, remote diagnostics, and automatic software upgrades, it effectively turns NEO into a robotics-as-a-service offering. In contrast, buyers who pay $20,000 upfront are betting that the first-generation hardware will remain capable and supported long enough to justify the premium.
As early 2026 reviews surface real-world reliability data, households are likely to weigh that risk transfer heavily. If breakdowns, sensor failures, or software gaps prove common, the subscription path could look far more attractive than locking in $20,000 upfront on a product with no track record in homes. Conversely, if NEO demonstrates durable performance and 1X quickly improves its software through over-the-air updates, long-term owners could end up with a capable home assistant at a lower total cost than multi‑year subscribers.
1X has not published detailed warranty terms, service-level commitments, or specifics on what happens to a subscriber’s robot if they cancel-whether the unit must be returned, disabled remotely, or can be converted to an ownership plan. Those gaps leave prospective buyers comparing two options without full information on the long-term cost or flexibility of either one. Until the company clarifies how repairs, accidental damage, and end-of-life recycling are handled, the headline prices only tell part of the financial story.
Factory scale, AI specs, and what 1X has put on the record
The production side of the equation took a concrete step forward when 1X opened a U.S. plant with plans to build 10,000 home robots in its first year of operation. That target would represent one of the largest initial manufacturing runs for any consumer humanoid robot to date, though no independent audit or third-party verification of the facility’s capacity has been published. The company describes the site as highly automated, with assembly lines tuned for humanoid form factors rather than traditional industrial arms.
In its own materials, 1X presents the factory as a cornerstone of a broader strategy to localize production and shorten supply chains for American buyers. A dedicated page on the company’s site frames the new facility as a specialized NEO factory focused on scaling up humanoid manufacturing while maintaining quality and safety checks. If the plant reaches its 10,000-unit goal, 1X will have to solve not only mechanical assembly at scale but also calibration, software flashing, and end-of-line testing for thousands of mobile, sensor-rich machines.
On the technical side, NEO runs on an onboard AI system called Redwood, a vision-language model with approximately 160 million parameters that processes information at around 5 Hz on the robot’s GPU, according to 1X’s technical overview. Those numbers are modest compared to the billion-parameter models powering cloud-based AI assistants, but the design choice reflects a deliberate tradeoff: running inference locally means the robot can function without a constant internet connection and avoids the latency of cloud round-trips during physical tasks like picking up objects or navigating a hallway.
Redwood’s training draws on both teleoperated and autonomous episodes, and 1X’s technical materials reference diffusion-policy research from Columbia University as part of the model’s development lineage. In practice, that suggests NEO learns from human demonstrations-people remotely piloting robots through tasks-and then refines those behaviors into policies it can execute on its own. The 5 Hz processing rate, while slower than typical game engines or high-frequency control loops, may be sufficient for deliberate household actions like folding laundry or unloading a dishwasher, where safety and precision matter more than speed.
CEO Bernt Bornich framed the preorder launch as a way to connect with early adopters while the company scales production. In an official announcement, 1X cast NEO as a step toward “redefining life at home,” emphasizing its potential to handle repetitive chores and to operate safely alongside people. Across its factory communications and order page, the company has repeated that first NEOs are scheduled to ship in 2026, tying the production ramp closely to that delivery window.
Unanswered questions before NEO arrives in homes
Several significant unknowns stand between the preorder page and a working robot in someone’s kitchen. No independent performance benchmarks exist for Redwood’s 5 Hz onboard inference on real household tasks. The only data available comes from 1X itself. Until third-party testers, consumer reviewers, or academic researchers run NEO through standardized manipulation and navigation tests, buyers are relying entirely on the manufacturer’s claims about what the robot can actually do.
That lack of external validation extends to edge cases that matter in homes: slippery floors, cluttered hallways, dim lighting, and interactions with pets or children. It is unclear how NEO handles failures, such as dropping objects, misclassifying items, or losing localization in a busy room. Without published logs or safety case studies, potential customers have no way to gauge how gracefully the robot recovers from mistakes-or how often those mistakes occur.
Regulatory and safety certification status is another blind spot. None of the available primary sources address whether NEO has cleared UL, FCC, or any other consumer safety standards required for a powered machine operating autonomously in a home with people, pets, and fragile objects. For a product shipping in 2026, the absence of any public certification timeline is notable. If approvals slip, the delivery schedule could move even if the factory hits its production targets.
Privacy and data handling also remain largely unaddressed in public materials. A humanoid robot with cameras and microphones in living spaces raises immediate questions about how video, audio, and sensor data are stored, who can access them, and whether any of that information is used to further train Redwood. 1X has not yet laid out a detailed privacy policy specific to NEO’s in-home operation, leaving early adopters to assume how their household data might be treated.
Demand signals are equally opaque. 1X has not disclosed how many preorders it has received, how the split between $20,000 purchases and $499 subscriptions is trending, or what its waitlist looks like. Without those numbers, the 10,000‑unit factory target is a supply-side ambition with no visible demand-side anchor. If orders fall short, the company may need to slow the production ramp, discount units, or lean more heavily on subscription offerings to fill capacity.
The practical question for anyone considering a preorder comes down to timing and risk tolerance. Placing a $20,000 order now means committing before any consumer reviews, safety certifications, or long-term reliability data exist. The subscription path at least limits the immediate financial exposure to $499 per month, but even that requires trusting that 1X will deliver on time, support the hardware for years, and keep improving Redwood’s capabilities. Until independent testing, clear warranty terms, and concrete certification updates emerge, NEO remains a bold promise: a full-size humanoid assistant priced like a car, asking households to decide how much of the future they are willing to prepay for today.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
