The Robotics Subscription Cash Crunch

Zach Martin

Here's a number that explains most of this episode. One hundred thousand dollars. That's roughly what it can cost a robotics company to build and deploy a single robot into a customer's facility once you add up the arm, the sensors, the compute, and the integration work. Now the second number. Somewhere between fifteen hundred and five thousand dollars a month. That's about what the customer pays to use that robot. Treat both as illustrative ranges, not one company's books. They still set up the whole puzzle. The company spends the hundred thousand on day one. It gets it back in small monthly pieces, and the payback runs thirty six to forty eight months. Welcome to ROBOT CAPITAL. I'm Zach Martin, and today we're looking at the other side of the robotics subscription: the vendor's balance sheet.

Zach Martin

Last time we looked at this from the customer's seat. We talked about how warehouse operators turn a multi million dollar capital outlay into a predictable operating expense with a three year commitment. That's a great deal for the customer. Their CFO gets a clean line item and no depreciation schedule. But accounting is a bit like a seesaw. When one side goes up, the other comes down. When the customer pushes the capital expense off their books, it doesn't vanish. It lands on somebody else's. And that somebody is a startup that was probably founded a few years ago, with a few dozen engineers and a lot of ambition. Today we follow that equipment onto the startup's balance sheet, and then follow the money that lets the startup carry it without giving away the company.

Zach Martin

Let's start with how robotics was sold the old way, because the old way explains why the new way exists. Traditionally, industrial robotics was capital equipment. A customer bought a workcell. Depending on the application, that could run seventy five thousand dollars to two hundred fifty thousand or more, and it can reach five hundred thousand once you count custom integration, tooling, and commissioning. The customer paid cash or took a commercial bank loan. After that, the customer owned everything. They carried the depreciation. They carried the risk the robot would sit idle half the week. They carried the risk the technology would look dated in five years. The vendor sent an invoice, collected, and moved on to the next sale. For the vendor, that's clean. The cash comes in around delivery.

Zach Martin

The trouble was getting to the signature. The brief for this episode describes enterprise sales cycles of twelve to eighteen months. During that time, a budget freeze or a nervous CFO can undo everything. The result has a name in the industry: pilot purgatory. A customer agrees to a pilot. The robot works. Everybody smiles. And then the purchase order never comes, because buying a robot is a big, scary, upfront decision with a payback the customer can't be sure of. Meanwhile the startup is burning cash on engineers and field support, waiting for a hardware signature that keeps slipping to next quarter. If you've ever run a small business and waited on a big client's payment, multiply that feeling by a very large number.

Zach Martin

So the industry came up with a fix. If the customer won't take the capital risk, the vendor takes it. Robots as a Service, or RaaS, and subscription leasing do exactly that. The customer pays by the month, or by the hour, or by the task, and the vendor owns the machine. Formic, a company we'll spend real time on, describes the pitch plainly. In its Series A announcement, the CEO wrote that customers pay nothing until the system is fully operational on their factory floor. After that, they pay only to use the robots by the hour, only when the robots are running, and only when they hit the performance Formic promised. That's about as friendly to the customer as a contract can be. And that's the problem. Every bit of friendliness on the customer's side is a cost somewhere on the vendor's side.

Zach Martin

Here's the arithmetic of what I'd call the working capital trap. Say a startup deploys one hundred robots in a year, at about a hundred thousand dollars each. That's ten million dollars going out the door for steel, wiring, lidar, compute, and installation. Now look at what comes back in that first year. At an illustrative twenty five hundred dollars a month per robot, each robot generates thirty thousand dollars a year. Even if all one hundred were running on January first, which they won't be, because they get deployed through the year, that's three million dollars at most. So the company is out ten million and has collected well under three. That's before we talk about technicians, software, or the sales team. And notice what I said: receipts, not profit. The seven million dollar gap is only the beginning of the story.

Zach Martin

Think of it as a house you build to rent out. You pay for the lumber, the labor, and the permits this year. The tenants show up later and pay in monthly installments. If you want to build a hundred houses, you can't cover the whole thing from the rent on the first few. The more successfully you grow, the bigger the hole gets. That's the cruel part of this model. In most businesses, growth eventually generates cash. Here, growth consumes cash first, and for a long time. A RaaS company that signs customers faster can run out of money faster. The coach in me wants to say this plainly: fast growth is not the same as healthy growth when every new customer costs you a hundred thousand dollars before they pay you a nickel.

Zach Martin

So the obvious question: who pays for those robots? The first instinct of any startup is venture capital. That's the money they know how to raise. And for a few early machines it can work. But the brief lays out why it breaks at scale. Venture equity carries an implied cost of capital of twenty to forty percent a year. That's not an interest rate anyone writes down. It's the return venture investors need to make their fund math work, because most of their bets fail and a few have to return everything. Now think about what you're asking that money to do. You're asking it to buy steel and wiring that depreciate over time, and to earn something like a modest, steady yield. It's a bit like paying for your delivery vans by selling pieces of your secret recipe. You get the vans, sure. But you've paid with the most valuable thing you own.

Zach Martin

Put some illustrative dilution numbers on that. Suppose a startup is valued at a hundred million dollars after a funding round. That's about the valuation Formic announced in January 2022. If it raised ten million in equity just to fund a hundred robots, that cash would cost it roughly ten percent of the company. Do that a few times and the founders and early employees own a shrinking slice of a company whose main product is a fleet of depreciating machines. And there's a second problem on the investor side. Venture investors are used to underwriting software, with gross margins of seventy to eighty percent and revenue multiples of ten to twenty times. A robot fleet financed on the balance sheet looks nothing like that. So the same dollars that would have funded the software and the artificial intelligence, the real intellectual property, are being spent on hardware.

Zach Martin

That brings us to a number you've probably seen, and I want to handle it carefully. S2G Investments, in a November 2025 report on Hardware as a Service, writes that companies with this model attract valuation multiples fifty nine percent higher than non HaaS frontier tech peers. S2G attributes that figure to Silicon Valley Bank. The report's explanation is that investors value the combination of hardware growth potential with the stability of contracted, infrastructure like cash flows. That's the source's claim. Here's my interpretation, and I'm flagging it as mine. A premium for recurring contracts doesn't tell you whether the company has solved the cash drain of building the fleet. A premium on paper is only worth something if the company survives the capital hungry years to collect on it. Which is why the financing structure matters as much as the contract.

Zach Martin

There's a market backdrop that makes all of this more pressing. S2G describes a world where equity capital got scarcer after 2022. Its report says global equity activity has fallen sharply since 2022, with capital concentrated in fewer, larger deals, and it describes a missing middle of growth stage companies that can't raise enough equity to fund capital heavy models. Its own example comes from clean energy, citing PitchBook: deal value of 23.5 billion dollars in both 2022 and 2023, falling to 17.5 billion in 2024, with 2025 at 7.8 billion in the data the report shows. That's one sector, not robotics, so don't over read it. But the line in the report that stuck with me is this: expensive equity capital is no longer the default tool for financing heavy capital expenditure business models. When the default tool gets harder to use, people go find better tools. That's what this whole episode is about.

Zach Martin

So what does the better toolbox look like? The way the brief frames it, and the way I'd frame it, is a three tier capital stack. Each tier has a job, a price, and a limit. The whole art of financing a robotics company is matching the right dollar to the right job. Think of it like packing a toolbox. You don't use a sledgehammer to hang a picture, and you don't use a tiny screwdriver to knock down a wall. A lot of robotics founders have learned this the hard way, and the ones who learn it early are the ones who get to keep more of their company.

Zach Martin

Tier one, at the top of the stack, is venture equity. It's the most expensive and the riskiest money, with the most dilution, so you reserve it for the things only equity should pay for. That means building the intellectual property: the physical AI models, the software that orchestrates a fleet, the proprietary research, and the sales and marketing machine. Think of equity as paying for the recipe, not the ovens. It's where a company takes its biggest, least certain bets, because that's the only place investors who expect twenty to forty percent returns can reasonably get paid. It's also the money that makes everything below it possible, because lenders want to see serious equity investors standing behind the company before they write a check.

Zach Martin

Tier two is venture debt and corporate credit lines. This is the middle of the stack. The brief puts the typical cost at roughly eight to fifteen percent interest, often with one to two percent warrant coverage. A warrant is a small right to buy equity later, a sweetener for the lender. So it's debt, but with a little bit of upside attached. Think of it as a bridge loan that connects one equity round to the next. Trinity Capital, a lender in this space, describes it as capital that supplements equity financing, extends runway to the next milestone, and minimizes dilution for shareholders. Trinity notes that the best time to raise venture debt is concurrent with or right after an equity raise, when momentum is strong, the diligence materials are in hand, and the cash is in the bank. That's a pretty human observation. You borrow when you don't look like you need to.

Zach Martin

How big is that middle tier? Runway Growth Capital's Venture Debt Review for 2025 and 2026 reports that U.S. venture debt hit a record 68.8 billion dollars in 2025, across roughly one thousand transactions a year. Notice the combination: dollars going up, deal count flat. Runway reads that as a market that is scaling, not overheating, driven by larger and repeat financings. The median deal size rose to 5.5 million dollars, and the seventy fifth percentile reached 27.7 million. For context, Runway says total U.S. venture equity in 2025 was 321.6 billion dollars, but half of all those dollars went into just 0.05 percent of transactions, and AI took 63.5 percent of deal value. So equity is abundant at the very top and picky everywhere else. Debt, in Runway's telling, is where the broader ecosystem is finding room to grow. Keep in mind that Runway is a venture lender. It has a reason to say this. The numbers are still useful.

Zach Martin

One passage from that review speaks directly to our topic. Lisa Chai, a general partner at Interwoven Ventures, is quoted saying that more business models are becoming financeable as lenders get better at underwriting nontraditional indicators of quality, including Robotics as a Service contracts, equipment backed structures, and contracted revenue streams. Read that again. A robotics service contract, something that looks strange to a traditional banker, is starting to look like an asset a lender can price. The same review points to Palmetto and its 1.2 billion dollars in capital commitments for residential solar leases as a different kind of example, where debt is built around long term customer contracts and asset backed cash flows, closer to project finance than traditional venture lending. That's solar, not robots. But it's the template robotics is borrowing.

Zach Martin

Here's a concrete robotics example of the middle tier at work. J.P. Morgan published a case study on Gecko Robotics, the company that builds climbing robots for industrial inspection. J.P. Morgan says it backed Gecko early with venture debt when it needed capital to keep building. The brief adds that this came before Gecko's 125 million dollar Series D round. I'd take the Series D figure as the brief's claim rather than something I've verified in the source text. What matters is the logic. A company building physical machines needed money to keep building between equity milestones, and a bank provided debt instead of the founders selling more of the company early. That's the bridge function in action. It buys time, so the next equity round happens when the company has more to show, and therefore at a better price.

Zach Martin

There's a warning label that goes with this tier, and I like how Steve Brotman of Alpha Partners puts it in the Runway review. He says that used from a position of strength, debt can be a highly rational tool for extending runway, preserving ownership, and reaching a value creating milestone. Used from a position of weakness, it can add complexity. The distinction, he says, comes down to whether the company is using debt to amplify momentum or to mask the absence of it. I think that's the most practical sentence in the whole report. Debt doesn't fix a weak business. It's a magnifying glass. If the business is working, the glass makes it bigger. If it isn't, the glass shows you the problem in high definition, and the lender wants to be paid on schedule anyway.

Zach Martin

Now tier three, and this is the one that matters most for our story: equipment financing, asset backed lending, and project finance. This is the asset tier. The brief puts the cost at roughly six to twelve percent, the cheapest in the stack, and the reason it's cheaper is the key idea of the episode. The lender isn't betting on the startup's future. The lender is lending against a specific thing: the physical robots and the contracted cash flows from the customers using them. The capital is tied to hardware that's out in the field, earning money. If the startup stumbles, the lender has a claim on collateral that generates cash. That's a very different risk from lending against a promise.

Zach Martin

Trinity Capital's equipment financing page describes this clearly. It says the product originated to let revenue ramping startups acquire capital intensive operating equipment while preserving equity and reducing the overall cost of capital. Trinity says it can provide up to one hundred million dollars of equipment capital, and it lists robotics and hardware as a service among the equipment types it finances. Two features stand out. The first is just in time financing: you draw down the money when you need it, as you deploy, rather than borrowing a lump sum up front and paying interest on cash sitting in the bank. The second is that the lien covers only the financed equipment. Trinity says that equipment specific liens do not impact the credit position of other lenders. In plain English, the robots secure the robot loan, and the company's other assets, including its software and its intellectual property, stay out of it.

Zach Martin

Let me show you why this tier changes the math, using a simple illustration that is mine, not from any source. Suppose that ten million dollar fleet is funded entirely with equity at a thirty percent expected return. The cost of that capital is three million dollars a year, in expectation. Now suppose eighty percent of the fleet, eight million dollars, is financed with asset debt at ten percent, costing eight hundred thousand a year. The remaining two million comes from equity at thirty percent, costing six hundred thousand. The total is one million four hundred thousand, a blended cost of fourteen percent instead of thirty. And the equity you raised is one fifth as large, so the founders give up a fifth as much of the company. The numbers are simplified, and real deals have fees and reserves. But the direction is the whole point. Put the right capital against the right asset and the same fleet becomes dramatically cheaper to own.

Zach Martin

There's a nice discipline hiding in this. Each dollar of the stack has to answer a question: what is this money going to do, and what will it be paid from? Equity answers: build something nobody has built, and get paid from a big, uncertain outcome. Venture debt answers: keep the lights on until the next milestone, and get paid from the company's future funding and revenue. Asset debt answers: buy a robot that has a signed customer contract on it, and get paid from that contract's monthly cash. That kind of clarity is what I'd call good financial hygiene, a small habit that compounds. Companies that mix these up, using expensive equity to buy steel or short term bridge debt to fund a five year fleet, tend to find out why the layers exist.

Zach Martin

The S2G report adds an interesting underwriting philosophy here. S2G's Special Opportunities team says it emphasizes the durability of lease and offtake cash flows as primary security, rather than relying solely on equipment value. That is a subtle but important point, and we'll return to it when we talk about what happens when things go wrong. For now, notice what it implies. The lender isn't mostly counting on being able to seize and resell the robot. The lender is counting on the customer's payments. The robot is the backup. S2G also says it has mapped nearly two hundred Hardware as a Service companies across renewable energy, industrial efficiency, agriculture, and mobility, so this isn't a robotics only phenomenon. Robotics is one of the newest industries to learn a playbook that solar and storage developers have run for years.

Zach Martin

Now let's open the machinery. If the asset tier is where the cheap money lives, how do you actually get it? You can't just walk into a lender with a pile of robots and a spreadsheet. The lender needs the robots and their cash flows separated from the startup's other risks, and that's what a special purpose vehicle, or SPV, is for. It's a legal box. And what goes in the box, and how it's built, determines whether a lender will fund it. I want to walk through this slowly, because it's one of those ideas that sounds intimidating and turns out to be fairly intuitive once you see the pieces.

Zach Martin

Start with two entities. The first is the operating company, which people in finance call the OpCo. That's the startup: the engineers, the salespeople, the technicians, the brand. The OpCo builds or buys the robot and signs a customer to a multi year contract. The brief says the minimum is typically twenty four to forty eight months. Now the second entity: the SPV, a separate company created for one purpose, holding a pool of robots and the contracts attached to them. The OpCo sells the hardware and the future customer payments into the SPV. In legal terms that's a true sale, and lawyers care a great deal about the phrase. A true sale means the assets genuinely belong to the SPV, not to the startup. The result is an SPV that's described as bankruptcy remote, meaning that if the startup runs into trouble, the creditors of the startup can't simply reach into the SPV and take the robots.

Zach Martin

Why does that matter to a lender? Picture yourself deciding whether to lend eight million dollars. If you lend to the startup directly, you're exposed to everything that can go wrong with the startup: a failed product launch, a lawsuit, a bad hire, a competitor with deeper pockets. You'd demand a high interest rate. But if you lend to an SPV that only owns a pool of robots, each earning money from a creditworthy customer under a signed contract, your risk shrinks to something you can measure. How many robots are in the pool? Who are the customers? How reliably do they pay? That's a question a credit analyst can actually answer, and a question with an answer is a question with a price. Narrowing the risk is how you get a cheaper rate. It's a bit like the difference between lending to someone to start a business and lending against a rental property that already has a tenant.

Zach Martin

Who provides the money? According to the brief, debt funds, infrastructure investors, and private credit syndicates provide delayed draw debt facilities directly into the SPV. Named examples include S2G's Special Opportunities strategy and Generate Capital. S2G's report says it launched that strategy in 2022 to back business models that sit between traditional growth and infrastructure investing. Delayed draw is the part to understand. The lender commits to a total amount, say fifty million dollars, but the startup doesn't receive it all on day one. It draws money in pieces as it deploys robots. That's how you avoid paying interest on money you haven't used yet, and it's the same just in time idea Trinity talked about. The commitment is like a credit line sized to the plan, and each draw has to be supported by robots that actually exist and contracts that are actually signed.

Zach Martin

How much does the lender advance against each robot? That's governed by what's called the borrowing base, and it's the heart of the structure. The brief says advance rates typically run seventy to eighty five percent of the equipment cost or of the contracted receivables value. So for a hundred thousand dollar robot, the SPV might borrow seventy to eighty five thousand dollars against it. The remaining fifteen to thirty percent has to come from somewhere else, usually equity from the parent company. That slice is the lender's cushion. If something goes wrong, the first losses fall on equity, not on the lender. It's the same reason a mortgage lender asks for a down payment: when the borrower has skin in the game, the lender sleeps better, and the rate comes down.

Zach Martin

But not every robot qualifies for that advance. The lender sets eligibility criteria, and this is where the details bite. The customer has to meet a creditworthiness standard. The contract has to meet a minimum length. And there are concentration limits. The brief gives an example: no single end user should account for more than fifteen to twenty percent of the collateral pool. Think about why. If one customer represents half the pool and that customer goes bankrupt or cancels, half the cash flow disappears at once. So a lender who sees fifty robots all sitting in one customer's plant will advance less, or nothing, against the excess. This means a startup can't just chase the biggest customer and stack robots there. The financing structure pushes it toward a diversified book of customers, which is also a healthier business.

Zach Martin

Next comes the part I find most elegant: the cash flow waterfall. When the customer pays, the money doesn't go to the startup's operating account. According to the brief, payments flow into a segregated lockbox account. And then the money flows downhill in a strict order. First, it pays the interest and principal on the senior debt. Second, it funds mandatory reserves: a maintenance reserve and a debt service reserve, which are like savings accounts the structure keeps for rainy days. Third, it pays a servicing fee back to the OpCo for running the robots. And only after all of that, whatever is left, the residual, is distributed up to the parent company as equity cash flow. The order is the whole point. The lender gets paid before the founders do. That's what lets the lender accept a lower rate.

Zach Martin

Notice a subtle consequence of that waterfall. The startup now has two sources of cash from the same robot. One is the servicing fee, which compensates it for the work of keeping the robot running. The other is the residual, which arrives only after debt, reserves, and fees are covered. In the early years of a facility, the residual may be thin. The startup is effectively working for a modest fee while the lenders get repaid. Later, as debt is paid down, the residual grows. So the SPV converts a startup from a company that sells machines into something closer to a small infrastructure owner, with a patient, long dated stream of cash. That's a big identity shift. It changes how the company is run, how it's valued, and who wants to invest in it.

Zach Martin

If you want evidence that this isn't theory, look at a job posting. Sereact, a German AI robotics startup that builds intelligence for robots in warehouses and factories, posted a role for a Director of Asset Backed Financing, based in Stuttgart. I want to be careful here. A job listing isn't proof that a facility has closed. It's evidence of intent, and of what the company believes it needs to build. But what it says is remarkable. The listing says that as Sereact scales, financing its robotics deployments becomes a strategic capability in its own right. The person hired is expected to build the program from zero: incorporate the SPV, design the structure, raise the debt, and run the vehicle long term. That's a full stack job. Not just finance. A founder's mandate, inside the finance department.

Zach Martin

The details in that listing are almost a syllabus for this episode. The director would define asset class and eligibility, choose the jurisdiction, and settle tax and accounting treatment. They'd work with outside counsel on true sale opinions, security packages, and servicing arrangements. They'd build the lender universe, prepare the data room, and negotiate term sheets with banks, private credit funds, and alternative lenders, starting with a first warehouse or seed facility and progressing to larger or syndicated structures. They'd negotiate advance rates, eligibility criteria, concentration limits, covenants, triggers, and reserve mechanics. And they'd stress test the structure against downside scenarios. All the words I've been using for the last ten minutes are in the job description. The brief's structure isn't an academic idea. It's what a funded robotics company is actually hiring for.

Zach Martin

One line in that posting made me stop. It says the person would own data quality end to end, from the operating company's source systems to the data tape that lenders see each month. Let that sink in. A data tape is a spreadsheet, basically, listing every asset in the pool: which robot, which customer, how many hours it ran, what it was billed, whether it paid. The posting also lists the ongoing operational backbone: borrowing base calculations, lender reporting, compliance certificates, servicer reports, and audit trails. This is the unglamorous part of the story. A robotics company that wants cheap debt has to run its robots and its record keeping to a standard that a lender's auditors will accept. If the telemetry from the robots is messy, or the billing system disagrees with the field service system, the borrowing base shrinks, and the cash shrinks with it. In this world, clean data is a form of capital.

Zach Martin

There's also a line in the posting about the board. The director would brief the board and equity investors on the program, including capacity, cost of capital, and, quote, the trade off between dilutive and non dilutive funding. That's the central tension, said in a single sentence. Every dollar of debt is a dollar of equity the company doesn't have to sell. But every dollar of debt is also a promise to pay on a schedule, with rules about what the company can and can't do. The interview process even includes a take home assignment: size a facility for a hypothetical asset pool, propose a structure, and flag the key risks. The company wants to see the candidate think like a lender and a borrower at once. I find that reassuring, honestly. Somebody is supposed to be asking what could go wrong before the money goes out.

Zach Martin

Let me also be fair about the costs of this machinery, because it isn't free. Setting up an SPV means lawyers, tax advisors, auditors, corporate service providers, and bank accounts. It means ongoing reporting. It means covenants and triggers that can take cash away from the startup if the portfolio underperforms. A trigger is a tripwire, a condition like delinquencies rising above a set level, and when it trips, the waterfall can change so the lender gets paid faster and the residual stops flowing to the parent. That's by design, because it protects the lender. But it means a startup in a rough patch can lose access to the very cash it needs most. This is why the sources keep returning to discipline. These structures reward companies that are organized, and they punish companies that are not. The sources don't disclose specific advance rates or pricing for Sereact, so everything I've described is the general design, not that company's actual terms.

Zach Martin

So far, I've described a structure that a company builds for itself. But in practice, there are two broader routes that robotics companies have taken to keep the balance sheet from breaking. The first is to become a financing company that happens to deploy robots. The second is to let a financing company do the financing, while the robot maker stays a robot maker. Both are real. Both have a named example. And comparing them teaches you a lot about who ends up carrying which risk. Let's take the first one, and its best known example: Formic.

Zach Martin

Formic Technologies is based in Chicago. It was founded in September 2020 by Saman Farid and Misa Ilkhechi. Farid was formerly a founding partner of Baidu Ventures and had invested in or sat on the boards of more than thirty AI and robotics companies. According to the announcement coverage, Ilkhechi was a former Universal Robots salesperson. Notice what that background suggests. These founders understood both the investing side and the selling side of robotics, and they looked at the industry and concluded something unusual. In Farid's words, quoted in the press coverage, what manufacturers needed was not any specific new technology, but a better way to access the technology that would best meet their needs. That's a company built around the financing problem from day one, not one that bolted a subscription onto a product afterward.

Zach Martin

Here's the key design choice. Formic doesn't build its own robots. According to the announcement, it delivers complete robotic solutions using technologies from leading robot vendors, including Universal Robots, FANUC, KUKA, and ABB. Formic designs the solution, buys the equipment, installs it, programs it, and maintains it, at its own cost, and customers pay by the hour. This has three effects on the capital stack. First, no research and development spending to build proprietary arms, so equity isn't consumed by hardware engineering. Second, the robots Formic buys are standard products from major manufacturers, which makes them more familiar to lenders. Third, the company's whole job becomes deployment, scoping, uptime, and customer relationships. In the S2G report's language, Formic combines off the shelf robotic systems with proprietary scoping, deployment, maintenance, and monitoring capabilities.

Zach Martin

Now the money. In January 2022, Formic announced a 26.5 million dollar Series A led by Lux Capital, with participation from Initialized, Correlation Ventures, Lorimer Ventures, One Planet VC, and individual executives. The announcement said the financing valued the company at over a hundred million dollars, only fifteen months after founding. And in the same announcement, Formic said it had secured more than one hundred million dollars of debt capital to fund equipment purchases. Look at the proportions. Twenty six and a half million in equity, and over a hundred million in debt capacity. The debt line is nearly four times the equity. That is the capital stack from the brief, made visible in a press release. Equity pays for the company. Debt pays for the robots.

Zach Martin

I want to flag a wording difference, because it matters. Formic's own post says it has secured more than a hundred million dollars of debt capital to fund equipment purchases. The trade press coverage on Automation dot com says Formic secured access to more than a hundred million dollars of debt capital. Access is not the same as drawn. A facility of that size is a ceiling, not a deposit. The company borrows as it buys robots, which is consistent with the delayed draw logic we discussed. So the right reading is that Formic had lenders committed to fund up to that scale, as it deployed. Separately, one aggregator page lists participation by Mitsubishi HC Capital America in the debt round, and the brief says the same. But that aggregator also lists funding details that differ from the company's own announcement on points like round sizes and dates, so I'm treating the company's announcement as the anchor and not leaning on the aggregator for anything else.

Zach Martin

The hiring in that announcement is telling too. Formic added a vice president of deployment operations named Steve Olszewski, who the announcement says was formerly general manager of FinanceIt, CEO of Spruce Finance, and a senior vice president at Discover Bank. The company's finance lead was listed as treasurer and VP of Finance. So when Formic added to its leadership team, one of the marquee hires came from consumer and home improvement lending, not from a robotics lab. That's interpretation, but I think it's fair: a company that owns robots and collects hourly fees from hundreds of customers is partly a lender and servicer, and it needs people who know how to run a portfolio of financed assets. The Formic press release even lists in house equipment financing as one of three things that make its model possible, alongside systematized deployment and maintenance and the solutions themselves.

Zach Martin

What does the customer's side look like? Formic's early customers included Polar Manufacturing, a 105 year old maker of custom door and truck body hardware, and Georgia Nut, a 75 year old maker of confections and nuts. Both automated for the first time in their histories. Farid wrote that the two companies have a combined 180 years in business, and that Formic expected to save them a combined 150,000 dollars a year on just the first deployed robots. He also noted that more than a third of existing customers had asked for more robots. And he pointed to a stat that explains the model's logic: factory equipment and plants often sit idle for more than sixty percent of the total hours in a year. A robot that works more hours, on a contract that pays per hour, is aligned with exactly what the customer wants. Those are the company's own claims, to be fair, but they show how the sales pitch and the financing pitch lock together.

Zach Martin

How far has the model gone? The S2G report, which counts Formic as one of its portfolio companies, so keep that relationship in mind, says Formic has deployed more than 500 systems across U.S. manufacturing facilities, and that its active robot fleet has delivered over 400,000 hours of production, with more than 1.2 billion products packed and stacked. S2G also says that 85 percent of U.S. manufacturing facilities operate without robots, which is the market Formic is chasing, and notes that Formic focuses on standardized, high value use cases such as palletizing and case packing. Standardization is a financing feature, not just an operating one. If every deployment is a one off engineering project, no lender can predict how it will perform. If a thousand deployments look alike, a lender can model them. S2G also notes that Formic's contracts are structured on short to medium term horizons, but that it has achieved strong retention and repeat business.

Zach Martin

That last point deserves a pause, because S2G's own stated preferences include long term contracts. The report says S2G prioritizes businesses with long term contracts and creditworthy customers, and says it avoids companies built on short term subscription agreements. Yet it describes Formic's contracts as short to medium term. How do these fit together? The report doesn't spell it out, so I'll offer a reading and label it as mine: S2G seems to be underwriting Formic on the combination of standard hardware, repeat customers, and a service organization that can redeploy robots, not on contract length alone. S2G does say retained ownership lets providers refurbish and redeploy hardware. That's an important distinction between a robot sitting at a customer's site for five years and a robot that can move to a new customer when a contract ends. A fleet of standard robots can be moved. A fleet of custom ones often can't.

Zach Martin

Now the second route: the captive vendor alliance. Here the robot maker doesn't become a financier. Instead it partners with one. The example in the brief is Universal Robots, the collaborative robot maker, and DLL, a global vendor finance provider also known as De Lage Landen. The two launched what the brief calls UR Financial Services, offering cobot leasing through Universal Robots' global distributor network. The brief says the program includes operating and finance leases, seasonal payment adjustments, usage based financing, and end of term technology refresh options. DLL underwrites the customer's credit and finances the cobots. So the customer sees a lease. The distributor sees an easier sale. And the balance sheet exposure sits with a company whose entire business is holding equipment finance assets.

Zach Martin

Think about how different that is from the Formic model. In the Formic model, the service company owns the robots, carries the customer risk, and takes the operating responsibility, and in return it keeps the upside. In the Universal Robots and DLL model, the robot maker sells the robot, in effect, to the finance company, and the customer leases from the finance company. The robot maker gets paid close to the traditional way, when the sale happens, while the customer gets the monthly payment experience. Seasonal payment adjustments are a nice example of why this appeals to certain customers. A food or agricultural manufacturer with a busy season and a slow season can pay more when cash is flowing and less when it's not. A lease like that can make the purchase decision feel safe in a way a lump sum never does.

Zach Martin

The trade off is control and margin. When a financing partner is in the middle, the robot maker shares the economics with the lender, and it doesn't own the customer's ongoing payment stream. It also may not control the uptime promise, since a lease is typically a passive financing arrangement, not a performance guarantee. That's a significant difference we'll come back to in a moment. For a robot OEM that doesn't want to build balance sheet operations in house, which is a serious undertaking as the Sereact job posting shows, the alliance model is a sensible way to offer financing without becoming a bank. And it's not unique to robotics. The S2G report notes that customers increasingly prefer paying over time, from cell phones to cars, with options to upgrade or bundle services.

Zach Martin

One more piece of the market context. A market research report from Dataintelo values the robotic system leasing market at 12.8 billion dollars in 2025 and projects it to reach 38.6 billion by 2034. It's a market research projection, so it deserves the usual caution that such forecasts are estimates and the methodology isn't something I can verify here. A separate prediction, cited in a Formic article from November 2023, attributes to ABI Research the forecast that by 2026 RaaS installations would generate 34 billion dollars in revenue. Again, that's a forecast from a few years ago, and I'm not in a position to tell you it came true. The reason I mention these numbers is that they show where analysts expect the leasing and service models to head. Whether the exact figures hold, the direction across the sources is consistent, and the capital behind it, venture debt, equipment lenders, infrastructure funds, is showing up in line with it.

Zach Martin

Let me draw the contrast in plain terms before we go into risks. The Formic route is a bet on operational excellence. You need to be great at scoping, installing, maintaining, and billing, and you capture a recurring stream. The vendor alliance route is a bet on distribution and partnership. You stay focused on the robot, and a finance specialist takes the credit and funding job. A third path, which the Sereact posting describes, is a robotics developer building its own financing program, with an SPV, to support whatever commercial model it ends up choosing: Robotics as a Service, equipment leasing, or customer financing, those are the posting's examples. None of these is automatically the right one. Which is best depends on whether your edge is the machine, the service, or the software that makes the machine smart.

Zach Martin

Everything so far makes this sound tidy. Cheap money, matching assets, clean structures. So let me be the cautionary voice for a while, because the brief spends real time on what's different about lending to robots, and because the places where this model can crack are exactly the places investors and founders should be watching. Lenders in trucking or aircraft have decades of data. Robotics lenders are mostly working with much less history. And there are three specific problems: what the robot is worth if things fail, who's paying for it, and whether it actually works.

Zach Martin

First, the residual value problem. Imagine you lend against a Boeing 737 or a Caterpillar excavator. If the borrower defaults, you can take the machine and sell it, because a liquid secondary market exists. Other airlines, other contractors, brokers, auctions, all ready to bid. The collateral holds value in the lender's mind because it can be turned into cash. Now imagine a custom industrial robotic cell running proprietary firmware. The brief describes the historical secondary liquidation value of such a cell as near zero if the startup fails. A lender can't easily repossess it and lease it again. Who wants a bespoke machine built for someone else's pallets, controlled by software only the failed company's engineers can maintain? You'd own a very sophisticated, very expensive piece of scrap, or something close to it.

Zach Martin

S2G's report makes the same point from the investor's side, and I think it's the most practically useful passage for anyone considering this market. It says S2G prioritizes borrowers that use off the shelf equipment and flexible system architectures, because that makes it easier to pivot to best in class technologies and preserve serviceability. Then it describes the failure mode precisely. If an early stage original equipment manufacturer goes out of business and its collateral is tied to proprietary technology, it can be difficult to find replacement servicers, which directly impairs equipment and contract value. S2G also says it avoids companies built on niche or proprietary hardware and unproven technologies. Put that next to Formic's choice to buy robots from FANUC, KUKA, Universal Robots, and ABB, and the strategic logic of its design becomes obvious. Standard robots have technicians available. Orphaned robots do not.

Zach Martin

This produces an uncomfortable irony for the most ambitious robotics startups. The companies building genuinely new machines, humanoids, novel manipulators, hardware designed around their own physical AI, are exactly the ones whose hardware is hardest to finance on an asset backed basis. Their robots are proprietary by nature. That's what makes them valuable to equity investors. But proprietary is a dirty word to an asset lender. Here's my interpretation, and I'll flag it as mine: that tension may push frontier hardware companies toward funding with more equity for longer, or toward partnering with an integrator or service company that wraps their machines in a financeable structure. The sources don't say that directly. What they do say is that lenders underwrite durability of cash flows and serviceability, and that is a very different test from technical brilliance.

Zach Martin

Remember the S2G line that underwriting emphasizes lease and offtake cash flows as primary security rather than equipment value. Now it makes more sense. If the robot's resale value is shaky, the lender has to lean on the customer's payment stream. And that brings us to the second problem: counterparty risk, which the brief calls the SME paradox. Small and medium sized manufacturers are the biggest untapped market for automation. The brief cites industry surveys saying less than twelve percent are automated globally. I'll note that's a different statistic from the S2G figure about U.S. facilities, so I wouldn't merge them. But the point holds: the customers who most need the financing model, the ones who can't write a check for a robot, are the very customers most likely to fail to pay for one. A blue chip logistics giant is a safe credit and doesn't need the help. A thirty person machine shop is a riskier credit and needs it badly.

Zach Martin

S2G's report touches this too. It lists, among the customer profiles for Hardware as a Service, small and medium sized businesses that benefit from an all in one service offering. It also warns that industries with weak counterparty credit can present challenges for long term contract underwriting. So how do RaaS providers manage it? The brief says through performance monitoring, telematics, and minimum billing tiers. Telematics means the robot reports back: how many hours it ran, whether it's healthy, whether it's even powered on. That lets the provider see trouble early. Minimum billing tiers mean the customer owes a floor amount whatever the usage, which protects the cash flow if the customer's business slows. S2G lists common contract structures that match this, including contracts with set minimum payments, with overage payments for additional use at a predetermined rate. A usage only contract with no floor is very customer friendly and very lender unfriendly.

Zach Martin

Here's the quiet tension in Formic's pitch, and I want to be fair about it. Formic's customers pay by the hour, only when the robots are running, and only when they hit promised performance. That's about as aligned with the customer as it gets. It also means that the revenue depends on the customer actually using the robot and on the robot actually performing. I don't have Formic's contract terms, so I can't tell you whether they include minimum commitments, and I won't guess. What I can say is that for any provider the question lenders ask is how much of the cash flow is contracted versus how much is merely hoped for. Pure usage billing makes the second answer bigger. That's the sort of detail that sits in a data room and decides the advance rate.

Zach Martin

The third risk is the one I find most distinctive: operational risk, specifically uptime. A passive equipment lease on a forklift is simple. The customer pays whether or not the forklift runs well, and if it breaks, that's the customer's maintenance problem. RaaS is different. The brief says RaaS contracts stipulate operational uptime of ninety five to ninety nine point seven percent. If a robot suffers a mechanical failure and the vendor lacks field technicians to fix it, customer billing freezes. Think about what that does inside the SPV. The waterfall depends on customer payments arriving in the lockbox. If a pool of robots goes down, and billing stops, the cash that was supposed to pay the senior debt isn't there. So a repair truck that doesn't show up becomes a credit event. Operations and credit, which in most companies live on different floors of the building, become the same thing.

Zach Martin

Think about what that means for how a lender evaluates a robotics company. They're not only looking at financial statements. They're looking at the density of field technicians relative to the number of deployed robots. They're looking at spare parts inventory. They're looking at mean time to repair. A company that grows its fleet faster than its service organization is quietly weakening its own collateral. And here's the connection back to Formic: its Series A post said the new capital would help it build a world class service and maintenance organization, and that the company was tripling its team with a focus on sales, engineering, finance, and marketing. That's the company's own statement of intent, not evidence of results. But it shows an understanding that the service team is part of the credit story, not a cost to minimize.

Zach Martin

S2G's report adds one more piece. It says that bundled operations and maintenance can ensure hardware remains well maintained and upgraded, and that retained ownership incentivizes providers to refurbish, reuse, and recycle equipment across multiple lifecycles. That cuts both ways. If the provider is good at refurbishing and redeploying robots, the residual value of the fleet is higher than the near zero scrap value of an orphaned custom cell. A fleet of standard robots managed by a competent operator is closer to a rental car fleet than to a pile of one off prototypes. But the same sentence shows the dependence: the value lives in the operator's competence. Take away the operator and you've got the orphan problem again. That's why lenders write servicing arrangements into these structures, and why the Sereact posting lists servicing arrangements among the legal work. Someone has to be able to step in.

Zach Martin

Let me pull this together into what I'd tell a founder, and what I'd tell an investor, with the usual caveat that none of this is investment advice. For a founder: financing is part of the product. Notice that every structure we've discussed rewards the same behaviors. Standardize your deployments. Collect clean data. Diversify your customers. Sign contracts with real minimums. Keep your uptime high. Those are good operating habits anyway, and the capital markets are now paying you a lower cost of capital for having them. The brief's description of the stack is a progression: equity for intellectual property, venture debt for runway and milestones, asset finance for the robots themselves. Founders who keep each layer doing its own job, and don't use the expensive layer for the cheap job, keep more of their company.

Zach Martin

For an investor, or just a curious observer, the sources suggest a few questions worth asking whenever a robotics company announces a big financing. Is the number committed capacity or drawn debt? Formic's own wording and the trade press differed on that, and it matters. Is the hardware standard or proprietary, and could a stranger service it? How long are the contracts, and how much of the revenue is a guaranteed minimum? Who are the customers and how concentrated are they? What does the uptime commitment say, and does the company have the technicians to meet it? And finally, the Steve Brotman test from the Runway review: is the debt amplifying momentum or masking the absence of it? Those aren't gotcha questions. They're the questions the lenders are asking behind closed doors, and they're better to ask early.

Zach Martin

I'll leave you with the thought I started with. A hundred thousand dollar robot, earning a couple thousand dollars a month. On its own, that's a slow, cash hungry machine, the kind that can sink a young company by being too successful too fast. But wrap it in a signed contract, a bankruptcy remote vehicle, a lockbox, a waterfall, a service team, and clean data, and the same robot becomes something a lender can understand and price. David Spreng of Runway Growth Capital says that debt, used well, is no longer a sign of distress. It is a sign of discipline, and increasingly, of competitive advantage. I'd add a coach's footnote. Discipline isn't glamorous. It looks like a spreadsheet, a data tape, and a repair truck that shows up on time. But that's where the physical AI revolution is going to be won or lost: not in the demo, but in the boring, compounding habits that let a company deploy the next hundred robots without selling the company to do it. This has been ROBOT CAPITAL. Follow the companies. Follow the technology. Follow the capital.