NEXT SHIFT

Robot-as-a-Service and the End of Owning Your Automation

You have been planning your first automation cell for eighteen months. You priced the cobot. You priced the end-of-arm tooling. You priced the safety fencing, the integration, the controls, the programming, and the operator training. The quote sits at $180,000. The justification memo is still unsigned. Your CFO wants to see an ROI model. Your floor manager wants to see it run a part first.

Here is what neither of them knows yet. The shop down the street skipped every one of those steps. They deployed a machine-tending robot last Tuesday. No capital purchase. No integrator on site for six months. No seven-figure sign-off. They pay by the hour, and if the robot misses its cycle-time targets, they do not pay at all. Their operator learned the handoff procedure in an afternoon.

Robot-as-a-Service is the quietest shift in manufacturing automation, and by the time most shops hear the term out loud, the economics will have already rewritten who competes for what work. This is a blog about what happens next, why it matters for big shops and small shops differently, and why the conversation Resell CNC customers should be having right now sounds nothing like the automation conversation the industry has been having for twenty years.

The Problem RaaS Actually Solves

The standard automation pitch goes like this. A robot costs $60,000 to $80,000. Integration, tooling, fencing, controls, and programming add another $80,000 to $120,000. The total lands somewhere between $140,000 and $200,000 for a single-cell deployment. The shop then depreciates that asset over five to seven years, assuming the job runs that long, assuming the part does not change, assuming the robot does not become obsolete, and assuming the shop can find the talent to program and maintain it.

None of those assumptions hold reliably in 2026. Jobs change. Parts revise. Robots improve. Programming talent costs more than the robot did. And smaller shops, which account for the majority of American machining capacity, almost never clear the capital expenditure threshold in the first place.

An MIT report cited by Formic Technologies puts the number starkly: only 10% of U.S. manufacturers currently leverage automation in their production facilities. Not because they do not want to. Because the acquisition model is wrong for the problem.

Robot-as-a-Service flips the model. A third party owns the robot, handles the programming, guarantees uptime, and provides 24/7 monitoring. The shop pays a monthly or hourly subscription. Integration cost at the start: zero. Maintenance cost over the contract life: zero. Capital requirement: zero. The shop pays only when the robot is running.

The Real Numbers

Pricing in the RaaS market is no longer theoretical. The leading U.S. vendor in the space, Formic Technologies, rents out robots at rates between $8 and $30 per hour depending on application complexity. Formic currently operates the largest independent robot fleet in the country, having surpassed 400,000 production hours across 60-plus facilities. Their reported uptime over the last twelve months is 99.43%. Customer renewal rate: 97%.

A Formic robotic palletizing cell recently ran a promotional rate of $4,750 per month, fully installed, shipped in 24 hours, with all maintenance included. For comparison, a comparable purchased system including integration would run $180,000 to $250,000 and take 16 to 24 weeks to deploy.

Hirebotics runs a similar model in welding. Their arc welding cobot subscription lands around $1,499 per month for the base configuration, with pay-per-arc-hour variants available. Rapid Robotics, Path Robotics, and several other vendors operate adjacent models. The common thread is that the vendor absorbs the capital risk, the deployment risk, and the uptime risk. The shop absorbs only the obligation to have work for the robot to do.

The math shifts from a depreciation question to a throughput question. At $15 per hour and 2,000 productive hours a year, the shop pays $30,000 annually for the equivalent of a $250,000 capital asset, with no depreciation schedule, no service contract, and no vendor lock-in beyond the term of the subscription.

What This Means for Small Shops

The single biggest barrier to automation in shops under 20 employees has never been the robot itself. It has been everything around the robot. Integration, fencing, controls, programming, and ongoing support. A $60,000 robot becomes a $150,000 project before the first part runs, and the project requires specialty engineering talent that most small shops do not have on staff and cannot justify hiring.

Formic's own vice president of sales has been direct about this on record. Most small manufacturers lack the programming and specialty engineering talent to get a robot ready for its job. Without that expertise, a purchased robot too often ends up in what the industry has started calling a robot junkyard, a very expensive paperweight.

RaaS collapses that risk to zero at the start. A one-machine job shop can deploy a machine-tending cobot next week, with full vendor accountability for programming, deployment, uptime, and maintenance. If the job ends, the subscription ends with it. The shop never owns an asset it cannot redeploy. It never pays for engineering it cannot use. It never inherits a robot it cannot reprogram.

For a shop running one or two turning centers that currently sit idle eight hours a night because there is no one to tend them, a $1,500-per-month machine-tending subscription turns those idle hours into billable hours. The payback math rarely requires a spreadsheet. The operator can see it in one shift.

What This Means for Big Shops

RaaS is not a small-shop-only story. Large manufacturers are using it in ways the small-shop conversation often misses.

The first use case is surge capacity. A tier-1 aerospace supplier that wins a 36-month program can deploy rented robots for the program duration and return them when the program sunsets. No stranded asset. No depreciation on a cell that will not have work in year four.

The second use case is pilot deployment. A large manufacturer evaluating whether to automate a particular process can run a 90-day pilot on a RaaS contract before committing to a purchase or a broader rollout. The cost of being wrong drops from $250,000 to a month-to-month subscription fee. That reduction in trial cost changes how aggressively the shop can experiment.

The third use case is labor risk hedging. With U.S. manufacturing projected to need 3.8 million new employees by 2033 and 1.9 million of those jobs likely going unfilled, large manufacturers are treating RaaS as a way to lock in production capacity that does not depend on hiring. A shop that cannot find welders can subscribe to welding capacity. That is not a future-looking argument. Shops are doing it right now.

The largest RaaS fleets in the country are not being deployed at 8-person job shops. They are being deployed at companies with hundreds of employees who already own plenty of robots and are using RaaS to expand without expanding their CapEx budget.

The Used-Iron Angle

Here is where the math gets uncomfortable for the buy-new-and-automated narrative, and where the Resell CNC thesis lands with the most weight.

A 2014 Okuma LB3000 with a fresh tooling package can be bought used for roughly $85,000. A new equivalent with OEM-integrated automation from the factory can cost $450,000 or more. The used machine runs the same part to the same tolerance. The new machine runs it with a factory-integrated robot that only the OEM's service team is authorized to touch.

Pair that used Okuma with a rented Formic machine-tending cell, and the total deployed cost of an automated turning center lands at roughly $85,000 in equipment plus $15 to $20 per hour for the robot. The new-and-automated equivalent costs $450,000 upfront plus depreciation plus service contracts plus the reality that the shop has created a single-vendor dependency for the next ten years.

The math is not close. For shops producing under 4,000 robot-hours per year, the used-plus-rented stack beats the new-and-automated stack on total cost of ownership for the entire useful life of the asset. And it does so while leaving the shop free to swap vendors, swap robots, or retire the cell without a capital loss.

This is the sentence that did not make sense two years ago and makes uncomfortable sense now. The lowest-CapEx path to an automated shop floor in 2026 is no longer a new machine. It is a used machine paired with a rented robot.

What the Big OEMs Are Doing About It

The machine tool OEMs have noticed. Haas, Mazak, Okuma, and DMG Mori have all announced partnerships with robotics vendors over the last eighteen months, and most of them position their own automation offerings as the answer to the RaaS threat. The question worth watching is whether they open their controls to third-party RaaS vendors or keep them closed.

An open control allows a Formic or Hirebotics cell to plug directly into the machine's M-code handoff without an integrator in between. A closed control forces the shop to either buy the OEM's own robot or hire a third-party integrator to bridge the gap. The posture each OEM takes on this over the next twelve months will determine whether used iron remains competitive with new automated iron, or whether RaaS becomes the only path to cost-effective automation on older machines.

Early signals suggest the OEMs are splitting. Some are opening their APIs and publishing integration guides. Others are tightening control access and making third-party robotics more expensive to deploy. Shops buying used machines in 2026 should factor control openness into their purchase decisions the same way they factor in spindle hours or tooling packages. A used machine with an open control is a different asset than a used machine with a closed one.

The Risks Nobody Is Talking About Yet

No emerging model is without risk, and RaaS has three the industry is still working through.

The first is vendor survivability. Several RaaS companies have raised hundreds of millions in venture capital. A few have already folded. Consolidation is coming. A shop subscribed to a vendor that goes bankrupt mid-contract faces a real continuity problem: who maintains the robot, who owns the data, who takes responsibility for uptime guarantees. Shops evaluating RaaS should weight vendor financial health as heavily as they weight pricing.

The second is insurance treatment. Most shop insurance policies were written before RaaS existed. Whether a rented robot is treated as a shop asset or a vendor asset for liability purposes is still an open question with most carriers. Shops should confirm in writing how their carrier treats rented automation before signing a subscription.

The third is data ownership. RaaS vendors collect enormous amounts of operational data, cycle times, uptime, production rates, rejection rates, tool changes. That data has real strategic value. Shops should read subscription contracts carefully to understand who owns that data and what the vendor can do with it. A shop that lets a vendor aggregate its production data into a competitive benchmark may be handing over more than the contract price implies.

The Reframe

The automation conversation for the last two decades has been a capital conversation. Can the shop afford it. Can the shop justify it. Can the shop depreciate it fast enough to matter before the technology changes.

Robot-as-a-Service does not answer any of those questions. It retires them. The only question that remains is whether the shop has a problem worth automating. If the answer is yes, the capital question no longer sits between the shop and the solution. The deployment question does not sit there either. The talent question does not sit there. Every barrier that kept the bottom 90% of U.S. manufacturers out of automation has been restructured into someone else's problem.

That does not mean every shop should subscribe instead of buy. Shops running high-volume, stable, long-horizon programs will still want owned automation for the economics and the control. But the number of shops that fall into that category is smaller than the industry has pretended, and the number of shops that should be running rented automation against used machines is larger than anyone has yet admitted.

Automation used to be something a shop bought. Soon it will be something most shops subscribe to. The shops that move first will not own more equipment than their competitors. They will just produce more with less of it, and they will have spent a fraction of the capital to get there.

The next shift in manufacturing is not about owning better machines. It is about owning fewer machines and subscribing to more capability. That future is not twenty years out. It is already running on 400,000 production hours at one vendor and counting, and the shops that understand it first are the shops that will be quoting faster, bidding lower, and turning more work than the shops still waiting for their CFO to sign the CapEx memo.