Robot-as-a-Service was supposed to be the answer to automation's oldest problem: the terrifying upfront check. Instead of $150,000 for a cell you might mis-spec, you pay a monthly fee and let someone else own the hardware risk. It is a genuinely good idea, it is growing fast — and it is quietly repricing the entire industry in ways that create very clear winners and losers. Our view: the model wins for buyers and loses for most of the companies trying to provide it, and the 2026 evidence already shows which side is which.
The number: a $32B market growing 20% a year
The RaaS market is not a niche. Per The Business Research Company's 2026 market report, it grew from about $26.72 billion in 2025 to $32.08 billion in 2026 — a 20.1% CAGR — and is forecast to reach $67.85 billion by 2030. That growth is real. What it obscures is that "RaaS" bundles three very different pricing mechanics, and the choice between them is where the winners and losers separate:
- Time-based — a flat monthly or annual rate per robot (the 1X NEO consumer humanoid, for instance, is offered at roughly $499/month).
- Usage-based — billed by the hour, shift, or square foot covered.
- Outcome-based — charged per result: orders picked, shelves scanned, pallets moved.
The further you move down that list — from renting a robot to selling a result — the more risk the provider absorbs, and the more the buyer likes it. That is the whole tension in one sentence.
Who wins: small manufacturers and asset-light aggregators
Buyers, especially small ones, win clearly. RaaS converts capex into opex, removes the mis-spec risk, and folds maintenance, uptime guarantees and consumables into one predictable line. For a shop that can't underwrite a six-figure purchase — or justify it against uncertain demand — this is the difference between automating and not. It also sidesteps the trap we've written about before, where the hidden costs of a cheap cobot blow up a naive purchase budget; with RaaS, integration and service are the provider's problem.
On the provider side, the winners are the asset-light aggregators. Formic is the clearest example: rather than manufacturing robots, it deploys existing hardware on an hourly model — publicly citing rates in the $8–$30/hour range depending on application, with guaranteed uptime and minimum annual commitments around $75,000 (these are the vendor's own stated terms). That structure — standardize the financing and service layer, ride other people's hardware — is the version of RaaS that scales without drowning in balance-sheet risk.
Who loses: the pure-play startups renting their own hardware
Here is where the repricing gets brutal. If you build the robot *and* finance it *and* service it, every deployment is a capital outflow you recover slowly over a multi-year contract. Growth consumes cash instead of generating it. That model has a body count.
Rapid Robotics — one of the earliest and best-funded RaaS pioneers — is the cautionary tale. It pivoted away from pure RaaS toward a picking-acceleration product (per its later leadership), and ultimately shut down as a US firm, with its assets acquired by RobCo in September 2025. It was one of more than a dozen robotics companies to fold or be absorbed in that stretch. The pattern is consistent: the pricing model that most delights customers is the one that most strains an under-capitalized provider.
Even the strong players are hedging. Geek+, which became the world's first listed pure-AMR company in its July 2025 Hong Kong IPO and has delivered on the order of 56,000 AMRs for roughly a 9% global warehouse-fulfillment AMR share, derives its revenue mainly from selling robotics solutions, with only a small portion from RaaS. The volume leader in warehouse robots is not betting the company on renting them. When the market's biggest AMR vendor treats RaaS as a side dish, that tells you something about the economics of the main course.
Our view
RaaS is repricing automation the way SaaS repriced software — by moving the cost from a wall of upfront capital to a stream of operating expense. For buyers that is almost pure upside, and it will keep pulling smaller manufacturers into automation who could never have written the purchase check. But the risk did not disappear; it moved onto the provider's balance sheet. The survivors will be asset-light aggregators (Formic's model), OEMs with real financing partners, and outcome-based specialists in narrow, high-repeatability tasks — not startups that build, own, and rent their own fleets on venture money. If you are a buyer, that asymmetry is a gift: negotiate hard, prefer outcome- or usage-based terms, and remember the provider needs your multi-year contract more than you need their robot.
One practical caution: RaaS math only works if you'd otherwise run the robot for years. On a long horizon, monthly fees can quietly exceed an outright purchase — the same breakeven logic that governs a cobot's ROI by application. Rent to de-risk and to pilot; buy when the use case is proven and durable. For the mechanics of how these contracts are structured, see our Robot-as-a-Service guide, and for the hardware itself, the warehouse robot category.
Sources
- The Business Research Company via TechTimes — Robotics-as-a-Service business models (market size $26.72B→$32.08B, 20.1% CAGR, $67.85B by 2030)
- Layer3 Labs — Robotics as a Service: 2026 Pricing Guide (time/usage/outcome models; 1X NEO ~$499/mo)
- Formic — Robots-as-a-Service (hourly pricing model, uptime guarantees; vendor-stated terms)
- The Robot Report — Rapid Robotics coverage and CMRA — robotics company failures 2024–25 (Rapid Robotics pivot, shutdown, RobCo acquisition)
- Longbridge — Geek+ HK IPO analysis (revenue mainly from solution sales, small RaaS portion; ~56,000 AMRs; ~9% AMR share)
