On this page
How You'll Pay for Your Robot
I remember photographing my first car. A black Mercedes A-class that I bought just after my second kid. The keys in my hand represented more than...
I remember photographing my first car. A black Mercedes A-class that I bought just after my second kid. The keys in my hand represented more than transportation. It represented freedom, or more practically, it meant that I was able to move with my family whenever we wanted without having to deal with public transportation. For many people, buying their first car represents independence, a threshold crossed into adulthood. People still take pictures of their first car. It is an event to remember.
Will you photograph your first humanoid robot?
Within a decade, acquiring a humanoid might feel as significant as buying your first car or your first home. But the path to ownership will look different. The previous chapter asked who might operate humanoid fleets. But there's a more personal question: how will you pay for these machines? When humanoids arrive at scale, most people will not buy them outright. They'll lease them, subscribe to them, or pay for them in ways that don't exist yet.
In October 2025, California-based 1X Technologies opened pre-orders for NEO Gamma, the first consumer humanoid designed for homes. Based on the NEO Gamma design unveiled in February 2025, the robot costs $20,000 to purchase outright or $499 per month on subscription, with deliveries beginning in 2026. Within weeks of the announcement, hundreds of people put down $200 deposits. The robots have not shipped yet, but the business models are already being tested in the market.
Cars went from outright purchases to leasing and ride-sharing. Smartphones went from upfront payments to carrier subscriptions. Humanoid robots are following the same arc, but faster.
The question is how you'll pay for them, and who will actually own them.
The Subscription Economy Comes to Robotics#
Robot-as-a-Service, or RaaS, is not a new concept. Industrial robots have been leased for years. But applying this model to humanoids changes the game. According to Grand View Research, the global RaaS market was valued at $1.05 billion in 2022. Precedence Research projects it will reach $8.72 billion by 2034, growing at 17% annually. In more optimistic scenarios, if adoption accelerates faster than expected, the market could hit $125 billion by 2034, growing at over 25% per year.
These numbers reflect a fundamental shift. Businesses and consumers no longer need to sink tens of thousands of dollars into a machine that might be obsolete in three years. Instead, they can subscribe, much like they do for software, cloud computing, or streaming services. The robot arrives, performs tasks, receives software updates, and scales with demand. When it breaks, the provider fixes or replaces it. When a better model arrives, subscribers can upgrade.
This model lowers barriers to entry. A warehouse does not need $100,000 in upfront capital to deploy a humanoid. It pays $5,000 per month and cancels the subscription if the economics do not work. A hospital testing robots for elder care does not need board approval for a major capital expenditure. It runs a three-month pilot on a subscription and expands from there.
1X Technologies is already testing this model with NEO Gamma. The $20,000 purchase option exists for early adopters willing to own the hardware outright. But the $499 monthly subscription is designed for everyone else: individuals who want to try the robot without commitment, small businesses testing automation, and institutions that need flexibility. The subscription includes software updates, maintenance support, and access to 1X's "Expert Mode," where human operators remotely guide the robot through tasks it has not yet learned.
This is not a perfect product. In fact, Wall Street Journal testing revealed that NEO Gamma cannot yet perform most household tasks autonomously. Every chore requires a remote human teleoperator. But 1X is transparent about this. The robot learns by watching humans. Over time, as thousands of NEO Gamma units collect data from real homes, the AI improves. Early subscribers are not buying a finished product. They are funding the training data that will make the robot autonomous.
This is what RaaS looks like in practice: you pay for access, not ownership. The robot improves continuously, and the whole fleet learns together.
The App Store for Robot Skills#
Germany's NEURA Robotics is taking a different approach. In January 2025, the company raised €120 million in Series B funding, led by Lingotto Investment Management, with participation from Volvo Cars Tech Fund, Delta Electronics, and others. The investment validated NEURA's position as Europe's leading cognitive robotics company, with a €1 billion order book and over 600 employees. Five months later, at the Automatica trade show in Munich, NEURA unveiled the full scope of its ambition: the Neuraverse ecosystem.
Imagine an app store, but instead of downloading games or productivity tools, you download skills for your robot. Need your humanoid to vacuum? Download the vacuuming skill. Want it to unload the dishwasher? There is an app for that. Health monitoring, tidying rooms, welding specific parts in a factory: all available as modular software that partners develop and sell.
Traditionally, robots are sold as complete systems. If you buy a welding robot, it welds. If you buy a vacuum robot, it vacuums. Reprogramming requires specialized expertise. NEURA's Neuraverse changes that. Robots become platforms. Third-party developers can create new capabilities, and robot owners can purchase and install them on demand.
What makes Neuraverse powerful is not just the app store model. It is the shared learning architecture. NEURA is building physical AI training centers called "NEURA Gyms" in partnership with NVIDIA, where data from real application scenarios is generated and combined with synthetic data from simulations. When one NEURA robot masters a new skill, every other NEURA robot on the network can learn it too. A MiPA robot in a German nursing home learns to assist with mobility exercises. That knowledge transfers instantly to MiPA robots in Japanese hospitals and American care facilities. The network grows smarter with every deployment.
MiPA, NEURA's household and service robot, is already available for reservation, with deliveries beginning in late 2025. The robot's cognitive abilities are built on NEURA's industrial AI architecture and powered by NVIDIA's GR00T N1 model for advanced cognition. Partners can develop skills for MiPA ranging from vacuuming to health monitoring to elderly care support.
This model protects intellectual property while encouraging collaboration. Developers can monetize their skills without giving away proprietary code. Companies can innovate without competing directly on hardware. NEURA's goal is to deliver 5 million robots by 2030, and the Neuraverse is the engine that makes that scale possible.
For buyers, this changes the math. A humanoid purchased in 2026 could learn skills in 2028 that didn't exist at the time of purchase, all without hardware upgrades. You're not buying a fixed set of capabilities. You're buying access to an evolving ecosystem where new skills appear constantly, priced individually or bundled into subscriptions.
Ownership, Leasing, and Everything In Between#
Here's what the market looks like in 2025:
Outright Purchase: 1X NEO Gamma costs $20,000 for early access buyers. Unitree's R1 humanoid costs $5,900. Figure AI announced Figure 03 in October 2025, designed for home use with its Helix AI model, with pricing estimated between $30,000 and $150,000 depending on configuration. Tesla's Optimus is targeting $20,000 to $30,000 at scale, with initial units in 2025 likely costing $40,000 to $50,000 before manufacturing economies kick in. Tesla plans to deploy 5,000 Optimus units in its own factories in 2025, scaling toward 50,000 to 100,000 units by 2026. Outright ownership appeals to early adopters, researchers, and businesses with capital to deploy.
Monthly Subscription: 1X NEO Gamma offers a $499 monthly subscription with a six-month minimum commitment and a $200 refundable deposit. This model mirrors smartphone financing, spreading costs over time while retaining the option to cancel or upgrade.
Hourly or Task-Based Leasing: Agility Robotics' Digit, priced around $250,000 per unit, is offered to logistics companies on an hourly payment basis. GXO Logistics, a major customer, does not pay upfront. It pays per hour of robot operation, aligning costs with productivity.
Skill-Based Marketplace: NEURA's Neuraverse allows businesses to pay for specific capabilities rather than entire robots. A factory might subscribe to a welding skill for six months during a production run, then cancel when the project ends.
Hybrid Models: Some providers are experimenting with combinations. A business might lease the robot hardware while purchasing skills a la carte. Or it might own the robot but subscribe to cloud-based AI updates and remote support.
Humanoid robots will serve wildly different markets. A university researching locomotion has different needs than an Amazon warehouse optimizing logistics, which has different needs than a family caring for an aging parent. No single business model will dominate. Instead, the market will fragment, with providers tailoring pricing to customer segments.
Why This Fragmentation is Good#
RaaS lowers barriers to entry, which means more competitors can enter the market. A startup does not need to manufacture robots itself. It can lease hardware from Unitree or NEURA, layer its own software on top, and offer specialized services to niche markets. One company might focus on elder care robots, another on agricultural automation, a third on entertainment and performance.
Regional fragmentation also plays a role. Europe's GDPR (General Data Protection Regulation) and China's Data Security Law create regulatory moats that favor local providers. A Chinese company offering RaaS in Europe must navigate strict data privacy rules, limiting its competitive advantage. Similarly, American companies face barriers in China. This regulatory patchwork ensures no single provider captures the global market.
The Neuraverse model accelerates fragmentation even further. If skills are modular and transferable, then competition shifts from who builds the best robot to who builds the best ecosystem. NEURA might dominate hardware in Europe while American developers create the most popular skills. Chinese manufacturers might win on price while Japanese firms lead in precision. The market becomes a network of interdependent players rather than a winner-take-all race.
This is fundamentally different from the smartphone market, where Apple and Samsung dominate. Humanoids are not standardized devices. A factory floor and a living room have nothing in common. No single design or business model will fit both.
The result is a competitive, pluralistic market where innovation happens at every layer: hardware, software, skills, pricing, and services.
What Happens When Robots Get Cheap#
Unitree's R1, priced at $5,900, is a signal of what is coming. In July 2025, the Chinese company shocked the industry by releasing a fully functional humanoid robot for less than the cost of a high-end laptop. The R1 is not a toy. It performs backflips, runs downhill, responds to voice commands, and navigates 3D environments. It lacks hands in the base model, but an upgraded version with dexterous hands is available for a few thousand dollars more.
This price point changes the calculus. Suddenly, universities, hobbyists, and small businesses can afford to experiment. A developer in Vietnam can order an R1, write custom software, and deploy it in a local business. A research lab in Brazil can purchase 10 units and run large-scale experiments on robot learning. A startup in Poland can lease R1s to test new business models without securing venture capital.
Cheap robots do not just expand the market. They accelerate the pace of innovation. More units in the field means more data. More data means faster learning. Faster learning means better robots. This creates a compounding effect. The first 10,000 robots will teach the next 100,000 how to work better. Those 100,000 will teach the next million.
Pricing is also a geopolitical weapon. China's manufacturing scale allows companies like Unitree to undercut Western competitors by 70% or more. Figure AI's Figure 02 costs around $50,000. Tesla's Optimus will cost under $20,000 at scale. Unitree's R1 costs $5,900 today. This forces American and European companies to either match Chinese prices or differentiate on quality, software, or services.
The race to the bottom on price is inevitable. Within a decade, humanoid robots could cost $10,000 or less. At that price point, they become accessible to middle-class households, not just businesses and institutions. And when that happens, RaaS models will shift again. Instead of replacing ownership, subscriptions will coexist with it. Some people will buy robots outright. Others will lease them. Still others will pay per task or subscribe to ecosystems like Neuraverse.
The Road Ahead#
By 2030, the business models for humanoid robots will be unrecognizable from today's. RaaS, ownership, skill marketplaces, pay-per-use — they'll all run in parallel, just as the car industry eventually settled into a mix of buying, leasing, and ride-sharing. The difference is speed. What took cars decades is happening in years.
The diversity of models matters more than which one wins. It means no single company can monopolize the market, and businesses of all sizes can access humanoid robots on their own terms.
The age of the humanoid robot is also an age of new economics — new ways to own, access, and pay for machines that didn't exist five years ago.
These business models are not emerging in a vacuum. They're shaped by a global competition — and nowhere is that competition fiercer than in China.