Why we bank hardware: the case for credit in the industrial era.

The past fifteen years of American venture and growth capital flowed overwhelmingly toward software. The returns were extraordinary. They will not be repeated in the same form. The next decade of American capital formation is shifting toward businesses that produce physical things — electric vehicles, grid hardware, automation systems, specialty chemicals, defense components, medical devices. Those businesses have fundamentally different capital structures than the last cycle’s winners, and they need fundamentally different lenders.

Steelbanc was built to be one of those lenders. This piece lays out, in plain terms, why we believe specialty commercial credit is the right capital for U.S. hardware companies — and why we are deploying our book toward EV platform manufacturers, robotics suppliers, and the capital-equipment OEMs that sit around them.

Hardware has collateral. Software has logos.

The first-principles case for hardware credit is the most boring one: there is something to lien against. A precision-machining business has CNC equipment a lender can finance against forced-liquidation value. An EV platform manufacturer has tooling, body lines, inventory, and intellectual property registrations. A specialty coatings producer has reactors, packaging lines, and a borrowing base of receivables and finished-goods inventory.

Software companies have customer logos, gross retention cohorts, and a contracted ARR base. Those are respectable assets. They are not what a senior secured term lender holds in a workout. Through-cycle credit performance is, more than anything, a function of what the lender can sell when things go wrong. Hardware lenders, on average, recover meaningfully more on meaningfully fewer events. We took our first credit loss in 2025 — final recovery, 91 cents on the dollar. That recovery rate is, in our view, the entire case for this asset class.

Hardware needs more capital than venture can supply.

Bringing a vehicle program to market, building a battery plant, or scaling a robotics fleet costs significantly more capital than even the largest venture rounds support. A $300 million Series D pays for product development, an initial pilot line, and roughly one year of operating burn. It does not, in any serious sense, build a manufacturing company.

What builds a manufacturing company, at scale, is a stack: equipment finance against the production line, a working capital revolver against receivables and inventory, a senior secured term loan against the plant and tooling, and a measured layer of growth equity on top. Money-center banks generally do not underwrite the first three layers for early-stage industrial businesses — the credit committees were not built for it. Private credit megafunds chase larger LBO transactions. That leaves a structural gap, and it is the gap Steelbanc operates in.

Platform architectures change the credit case.

Among the EV manufacturers we follow, the most credit- durable are the ones building on a proprietary platform architecture rather than as a single-vehicle program. Skateboard platforms — chassis-and-drivetrain modules onto which different vehicle bodies can be mounted — allow a manufacturer to amortize its core capital expenditure across multiple future programs. A single platform investment supports several vehicles; per- program capex falls; the credit story strengthens. Rivian followed a version of this approach with its R1 and EDV programs; REE Automotive is built on the premise; and Olympian Motors is bringing its Model O1 to market on a proprietary EV skateboard platform of its own.

From a credit perspective, the implication is the same in each of these cases: the platform itself is the durable collateral. It is a fungible asset whose value does not collapse if any one model program is delayed, redesigned, or repositioned, and it is exactly the kind of asset specialty equipment-finance teams are built to underwrite.

Robotics and automation are the supply chain underneath.

You cannot have a domestic EV industry without the automation suppliers that build the lines. The companies producing welding cells, body-in-white stations, automated guided vehicles, and end-of-line test stations are the picks-and-shovels of the industrial revival. The names span scales: Symbotic and Berkshire Grey at the warehouse layer; Boston Dynamics and a generation of younger humanoid robotics shops on the floor; ATS Corporation and Rockwell Automation in capital equipment. They share the same credit profile as the OEMs they serve — real collateral, real backlog, real customer concentration to manage — and we expect a growing share of our originations through 2027 to be directed at this layer of the supply chain, alongside the EV platform manufacturers and Tier 2 component shops we already bank.

What we look for.

Across the hardware book, our underwriting is consistent. We look for borrowers with $30 to $250 million in revenue, real collateral (equipment, real property, or a tested borrowing base of working capital), a customer portfolio that is not catastrophically concentrated, and management teams that have run a P&L through at least one downturn. For pre-revenue or early-revenue hardware companies, we will look harder at the underlying platform asset, the production line economics, and the equity capital already in the company before we extend committed credit.

We are not the only specialty lender in this market, and we will not be the only one for long. But we believe the credit demand from U.S. hardware manufacturers — the EV platform builders, the robotics shops, the equipment OEMs, the specialty industrials — will outpace the supply of disciplined balance-sheet capital for the rest of the decade. That is the opportunity we were built to take.