1X NEO Gamma at $499 per Month: The First Robot You Might Actually Rent

Every other humanoid robot company is chasing factory floors. Tesla wants Optimus assembling cars. Figure AI is deploying at BMW. AgiBot is flooding Chinese warehouses. The entire industry has converged on a single thesis: the first viable market for humanoid robots is industrial.

1X Technologies disagrees. The Norwegian-American company, backed by OpenAI and with nearly a billion dollars in commitments, is making the most audacious bet in robotics. Their NEO Gamma is not designed for a factory. It is designed for your living room. And you do not have to buy it. You can rent it for $499 a month.

This is either a stroke of genius or a spectacular way to burn through investor capital. Let us figure out which.

The basics: what is NEO Gamma?

NEO Gamma is a fully bipedal humanoid robot built explicitly for homes. It stands between 165 and 170 cm tall, roughly the height of an average adult woman, and weighs just 30 kg. For context, that is lighter than a golden retriever. It is lighter than most carry-on luggage limits for international flights. It is about half the weight of Figure 02.

The weight matters enormously. A 30 kg robot that falls on a child is a very different safety event than a 60 kg robot that falls on a child. 1X made weight the central design constraint, and every other engineering decision flows from that choice.

NEO uses tendon-driven actuation rather than the rigid electric motors found in virtually every other humanoid robot. Think of how human muscles work: tendons pull on bones through flexible connections that inherently absorb shock and provide compliance. When you bump into a human arm, it gives way. When you bump into an industrial robot arm, it does not. NEO’s actuators are designed to behave more like the former. The robot moves quietly, smoothly, and with a kind of organic softness that rigid systems cannot replicate.

The trade-off is strength. A tendon-driven system cannot exert the same force as a high-torque electric motor. NEO is not going to carry heavy boxes or perform industrial assembly tasks. 1X is betting that for household applications, you do not need that kind of force. You need something that can pick up a coffee mug, fold a towel, and push a vacuum without being a danger to the people around it.

The rental model: why $499 per month changes everything

The purchase price for NEO Gamma is $20,000 for early adopters. That is aggressive compared to the broader humanoid market, where Figure 02 is estimated at $150,000 or more and even the budget-oriented Unitree G1 starts at $16,000. But $20,000 is still a lot of money for a consumer product that is essentially a first-generation device.

This is where the $499 per month rental becomes genuinely interesting.

A rental model solves three problems that have killed every previous consumer robot. First, it eliminates the massive upfront cost barrier. $499 per month is expensive, roughly the cost of a car payment, but it is psychologically easier for a household to absorb than a $20,000 lump sum. Second, it shifts the risk from the buyer to the manufacturer. If NEO does not deliver enough value in month one, you cancel. You are not stuck with a $20,000 paperweight. Third, and most importantly for 1X, it creates a recurring revenue stream that funds continuous improvement of the AI and software stack.

The rental model also lets 1X iterate rapidly. When a rented NEO receives a software update that makes it 20% better at folding laundry, every subscriber benefits immediately. When a hardware revision improves the actuators, 1X can swap units during a maintenance visit. The subscriber always has the latest version. The purchaser is stuck with whatever they bought.

This approach mirrors what worked for other technology categories. Nobody bought their first smartphone outright. Carrier subsidies and installment plans lowered the barrier to adoption. Nobody buys a Tesla Powerwall when they can lease a solar panel system. The subscription model is how expensive new technology crosses the chasm from early adopter to mainstream, and 1X is the first humanoid company to apply it seriously.

The Norwegian roots: how a small country built a big robot

1X Technologies was not always called 1X Technologies. It started as Halodi Robotics, founded in 2014 in the small coastal town of Moss, Norway by Bernt Oivind Bornich. For its first several years, Halodi was a quiet Nordic robotics research company, far from the hype and funding frenzy of Silicon Valley.

The company rebranded to 1X Technologies and everything changed when the OpenAI Startup Fund led its Series A2 round. OpenAI’s involvement was not just about money. It signaled that the most important AI company in the world saw something in 1X’s approach that was worth betting on. That signal brought in Samsung NEXT, NVIDIA, SoftBank, and Jeff Bezos as investors. The Series B, led by EQT Ventures, raised $100 million.

In July 2025, 1X relocated its headquarters to Palo Alto, California, the center of the AI universe. But manufacturing stayed in Moss, Norway. This is a deliberate and telling choice. Norway has some of the highest labor costs in the world, but it also has an extremely skilled manufacturing workforce, strong worker protections that create operational stability, and access to cheap renewable energy. Building in Norway says something about 1X’s confidence in its margins.

The company now employs roughly 960 people, split between the California headquarters (business, AI research, partnerships) and the Norwegian manufacturing facility (hardware, production, testing). Additional operations run out of Hayward, California.

The production roadmap is ambitious. 1X targeted thousands of units in 2025, tens of thousands in 2026, and has publicly stated a goal of millions by 2028. As of early 2026, approximately 20 units have shipped. That gap between aspiration and reality is worth examining closely.

The EQT deal: 10,000 robots walk into a warehouse

In December 2025, 1X struck a landmark deal with EQT, the Swedish investment firm that led its Series B. The agreement calls for up to 10,000 NEO units to be deployed across EQT’s portfolio of more than 300 companies for manufacturing, warehousing, and logistics tasks.

This deal is important for several reasons that go beyond the headline number.

First, it provides 1X with a guaranteed customer base as it scales production. Building a factory is expensive. Building a factory without confirmed orders is terrifying. The EQT deal de-risks the production ramp.

Second, it reveals something about 1X’s actual strategy. Despite all the home-robot messaging, the first large-scale commercial deployment of NEO will be in industrial settings. This is pragmatic. Factories and warehouses are controlled environments where robots can operate with fewer safety constraints and where the ROI calculation is straightforward: if the robot costs less than a human worker per hour of output, it makes economic sense.

Third, the deal validates that serious investors see NEO’s design as commercially viable. EQT is not a venture capital fund chasing hype. It is a $130 billion AUM investment firm that deploys capital based on operational value. If EQT is willing to put 10,000 NEO units into its portfolio companies, it has done the math on unit economics and productivity gains.

What makes NEO different: the tendon-driven bet

To understand why NEO Gamma matters, you have to understand how it moves, because 1X made a fundamental engineering decision that sets it apart from every major competitor.

Almost every humanoid robot on the market uses rigid electric actuators. These are essentially high-torque motors connected to gearboxes that drive joints directly. They are powerful, precise, and well-understood. They are also loud, potentially dangerous in close human contact, and heavy.

1X chose a different path. NEO uses a tendon-driven actuation system, sometimes described as a musculoskeletal design. Instead of motors at each joint, NEO routes flexible tendons from centralized motor units through the body to the limbs. The tendons pull on joints the way biological tendons pull on bones.

The advantages for a home robot are clear. Tendon-driven actuation makes NEO quiet. This is not a trivial point. If you have ever heard an industrial robot arm move, you know the high-pitched whine of electric motors under load. Living with that sound in your home, all day, would be maddening. NEO moves with a softness and quietness that rigid systems cannot match.

The compliance advantage is equally important. When NEO’s arm bumps into a person or a pet, the tendon system naturally absorbs the impact and gives way. A rigid electric actuator does not do this unless complex force-limiting software intervenes in time, which it might not always do. For a robot sharing space with toddlers and cats, inherent compliance is not a nice-to-have. It is a safety requirement.

But the disadvantages are real, and they define the limits of what NEO can do. The soft actuator approach makes NEO weaker, less precise, and harder to maintain than rigid alternatives. If you need a robot to pick up a 15 kg box, NEO probably cannot do it. If you need a robot to insert a screw with sub-millimeter accuracy, NEO definitely cannot do it. If a tendon wears out or snaps, replacing it is more complex than swapping an electric motor.

The AI brain: OpenAI in the background, teleoperators in the foreground

NEO Gamma runs a proprietary AI stack built on an NVIDIA Jetson AGX Orin processor, paired with a world model and a Redwood AI vision-language model. The system handles basic navigation, object recognition, and learned behaviors on-device.

But here is the part that most coverage glosses over: NEO currently relies on human teleoperators for complex tasks.

1X employs a team they call “Turing operators” who can remotely control NEO when the onboard AI encounters a task it cannot handle autonomously. This means a human somewhere is watching through NEO’s cameras and guiding its limbs through difficult manipulation tasks. The customer can decide when a 1X employee is allowed to view their surroundings, and teleoperators cannot connect without explicit owner approval.

This is honest engineering. Rather than pretending NEO can do everything autonomously and failing badly, 1X built a human-in-the-loop fallback. The teleoperator approach also generates training data. Every time a human guides NEO through a task, that demonstration becomes a data point for improving the autonomous AI. Over time, the system should need teleoperators less and less.

Still, the teleoperation dependency raises questions. How many simultaneous teleoperators does 1X need per deployed robot? At scale, if 10,000 NEO units are in homes and each one needs an operator for even 10% of its active time, that is a massive human labor cost. The unit economics of the $499 monthly rental start to look much less attractive if each robot needs meaningful human oversight.

1X has not disclosed the current ratio of teleoperator hours to autonomous hours. Until they do, the true cost structure of the rental model remains opaque.

Privacy in your living room: the uncomfortable reality

Putting a camera-equipped, internet-connected, partially teleoperated robot in someone’s home creates privacy concerns that dwarf anything raised by smart speakers or security cameras.

NEO has cameras. It needs them to navigate and manipulate objects. Those cameras can see everything in your home. When a teleoperator connects, a human being at a 1X facility is looking through those cameras in real-time. They can see your furniture, your family, your personal belongings, and anything else in the robot’s field of view.

1X has implemented several privacy controls. Teleoperators cannot connect without explicit owner approval each time. Owners can designate no-go zones, rooms or areas where NEO is not allowed to go and where teleoperators have no visual access. Time-based restrictions let owners block teleoperation during certain hours.

These are reasonable first steps, but they may not be sufficient. Consider the scenario where NEO is tidying a bedroom and encounters a task it cannot handle autonomously. It requests teleoperator access. The owner approves. Now a 1X employee is looking at your bedroom through a robot’s camera while guiding it to pick up clothes from the floor. This is a level of intimate access that goes far beyond what any consumer technology has asked for before.

The privacy framework will need to evolve significantly before mainstream consumers are comfortable with this arrangement. One possible solution is a privacy mode where NEO can only attempt tasks autonomously, with teleoperation completely disabled. The robot would be less capable, but the privacy trade-off would be clear and controllable.

How NEO compares to the competition

NEO Gamma occupies a unique position in the humanoid robot landscape. It is not competing directly with industrial humanoids like Figure 02 or Tesla Optimus. It is competing with the idea of a home robot, a category that does not really exist yet.

Still, comparing NEO to other humanoids reveals how different 1X’s design choices are.

The most important number in that comparison is production volume. NEO has shipped approximately 20 units as of early 2026. Unitree has shipped over 3,200 G1 units. AgiBot has shipped over 5,000 units. Even Figure AI, which is notoriously cautious about scaling, has deployed roughly 200 units.

Twenty units is a prototype, not a product. 1X has preorders open and the EQT deal in hand, but until those translate into thousands of deployed units, NEO Gamma is still in the “promise” phase of its development. The gap between 20 units and the stated goal of millions by 2028 is staggering.

The manufacturing challenge: from Moss to millions

1X manufactures NEO vertically in-house at its facility in Moss, Norway. Vertical integration means 1X controls the production of most major components rather than sourcing them from suppliers. This gives the company tighter quality control and potentially better margins at scale, but it also means the company has to build manufacturing capability from scratch.

The production roadmap is the most ambitious part of the entire 1X story.

Thousands in 2025. Tens of thousands in 2026. Millions by 2028.

To put this in perspective, Tesla, the world’s most celebrated manufacturing company, took roughly five years to scale from producing thousands of Model 3s per year to producing a million. Tesla had existing factory infrastructure, a proven supply chain, and unlimited access to capital markets. 1X has a single factory in Moss and $125 million in funding.

The gap between 20 shipped units and a target of millions within two years is not a rounding error. It is a chasm that requires either massive additional capital investment, a manufacturing partnership with an established hardware company, or a significant scaling of the Moss facility. More likely, it requires all three.

This is where the EQT relationship could prove critical beyond the 10,000-unit order. EQT is an infrastructure investor with deep experience in scaling industrial operations. If EQT provides not just orders but also operational expertise and capital for manufacturing expansion, the production roadmap becomes more plausible.

But “more plausible” is not the same as “likely.” The honest assessment is that 1X will probably ship hundreds of NEO units in 2026 and perhaps low thousands in 2027. Millions by 2028 would require a manufacturing miracle.

The investment case: why serious money is flowing in

Despite the gap between ambition and current reality, 1X has attracted some of the most sophisticated investors in technology.

The OpenAI Startup Fund leading the Series A2 is the most significant signal. OpenAI does not invest casually. Their involvement suggests that 1X’s AI approach and data collection strategy align with OpenAI’s broader vision for embodied AI. The connection also gives 1X potential access to OpenAI’s models and research, though the specifics of any technical collaboration have not been disclosed.

Samsung NEXT brings consumer electronics expertise and potential integration with the Samsung smart home ecosystem. NVIDIA brings AI compute expertise and preferential access to hardware (the Jetson AGX Orin in NEO is an NVIDIA product). SoftBank brings a portfolio of robotics investments and distribution networks in Asia. Jeff Bezos brings… well, Jeff Bezos bought the company that makes Roomba, so he clearly sees something in home robotics.

The bear case on 1X is straightforward. The company has shipped 20 robots and is valued in the hundreds of millions. Its production targets are wildly aggressive. The home robot market it is targeting does not yet exist. And the tendon-driven actuator design, while elegant for safety, may prove too limited for the range of tasks consumers actually need.

The bull case is equally straightforward. 1X has the most thoughtful approach to home robot safety in the industry. The rental model could crack the consumer adoption problem that has killed every previous home robot. The OpenAI backing provides an AI moat that no competitor can match. And the EQT deal provides industrial revenue while the consumer market develops.

What NEO can actually do today

Based on demonstrations and home trial reports, here is a realistic assessment of NEO Gamma’s current capabilities and limitations.

The honest summary: NEO Gamma today is a robot that can do light tidying, fetch items, and follow basic routines in a home environment. It does this more safely and quietly than any other humanoid on the market. But it cannot do the hard stuff. It cannot cook dinner. It cannot do laundry end-to-end. It cannot handle the chaotic reality of a home with kids. Not yet.

The question for a potential $499 per month subscriber is: is light tidying and fetching worth $499 per month? For most households, the answer is probably not yet. For households with specific accessibility needs, elderly care requirements, or simply a desire to be among the first to live with a humanoid robot, it might be.

The bigger picture: is home-first the right bet?

1X Technologies is making a contrarian bet that goes against the entire industry’s current direction. Every market analyst, every investment bank report, and every competitor’s strategy deck says the same thing: industrial first, consumer later. Build robots for factories, prove the technology, generate revenue, and eventually bring a consumer version to market when costs come down and capabilities go up.

1X is saying: no. Build for the home first. The home generates better training data because it is more complex and varied than a factory. The home market is larger (there are 130 million households in the US alone vs. approximately 300,000 factories). The home market has better unit economics at scale because consumer products achieve economies of scale that enterprise products never do.

The counterargument is timing. The industrial market exists today. Companies are paying for robots today. BMW is writing checks to Figure AI today. The consumer market is a projection. It depends on robots being good enough, cheap enough, and safe enough for people to invite them into their homes and pay $499 per month to keep them there. Those conditions are not met today, and it is genuinely uncertain when they will be.

The most successful technology companies often created markets rather than entered existing ones. The iPhone created the smartphone market. Tesla created the premium EV market. Airbnb created the home-sharing market. Each of them looked like a crazy bet at the time. 1X is hoping to be in that category: the company that created the home robot market.