Manufacturing and Supply Agreements for Hardware Startups: NRE, Tooling Ownership, MOQs, and the Clauses That Protect Your Product

The Contract Hardware Founders Sign Without Reading

When you move a hardware product from prototype to production, your attention goes where the numbers are. You scrutinize the bill of materials, push the contract manufacturer on unit price, and argue over minimum order quantity and payment terms. That instinct is reasonable — price, payment, and MOQ are the most heavily negotiated terms in almost any contract-manufacturer agreement, and they're where most founder energy concentrates. But the dollars per unit are not where the agreement decides your company's fate.

The contract you skim past the cover page does far more than set a price. It decides who owns the injection-mold tooling you paid to build. It defines what "conforming" product actually means — and that definition matters more than it sounds, because acceptance is what triggers your duty to pay. Until product is accepted, you can reject it; after acceptance, the leverage shifts. The same document sets your warranty terms, your remedy when an entire production run fails for the same defect, and whether you can move your product to a different factory without losing your tooling or your supply.

Here's the trap. Most founders sign the manufacturer's own boilerplate, and that boilerplate is drafted from the manufacturer's side of the table. It is built to narrow your remedies, not to protect your product. So while you're winning the fight over unit cost, you may be quietly losing the fights over tooling title, IP, and your ability to walk away. This guide works through the clauses that actually carry that risk — NRE and tooling ownership, MOQs, acceptance, warranty, and exit — so you know what you're signing before you sign it.

NRE and Tooling: Who Pays, and Who Owns What You Paid For

Non-recurring engineering (NRE) is the one-time, customer-specific cost of standing up your production line: the injection molds, fixtures, test jigs, and the programming and software that are built specifically for your program. By definition, these are tools your contract manufacturer (CM) cannot share across its other customers, because they exist only to make your product. That specificity is exactly why you pay for them up front, and it is also why the ownership question matters so much. A tool built for nothing but your product is a tool that can hold your product hostage.

Here is the trap that catches first-time hardware founders: paying the NRE or tooling invoice does not automatically transfer ownership of those tools to you. You can wire six figures for a mold and still have no legal right to take it with you when the relationship sours. A CM that owns the tooling can leverage that ownership to stall a transfer, raise prices, or simply refuse to release the molds when you want to dual-source or move production. As a rule of thumb, if tooling is fully owned by the manufacturer and reused across its book of business, it should not be billed to you as NRE in the first place — NRE is supposed to buy you something.

The contract language is where this gets decided, and the defaults usually run against you. In representative manufacturing and supply agreements filed publicly with the SEC, tooling ownership is commonly structured so that anything categorized as standard "Equipment and Tooling" defaults to the seller, while only "Custom Tooling" that is expressly itemized on a schedule belongs to the buyer. "Tooling" in these agreements is typically defined broadly — dies, test and assembly fixtures, gauges, jigs, patterns, and the supporting documentation. The practical consequence is blunt: a buyer who pays the NRE but never itemizes its tools on the schedule may walk away owning nothing at all.

The fix is concrete, and you should negotiate for all three of these together. Your target is straightforward: upon completion of payment, you own every tool and fixture required to produce your product, ideally tied to milestone payment terms so ownership tracks the money as it goes out the door.

• Title on payment. Buyer-funded tooling vests in you as soon as it is paid for — not at some undefined later date, and not at the CM's discretion.
• Right to take possession and transfer. A return-of-buyer-property clause gives you the right to physically retrieve your tools and move them to a new manufacturer on termination.
• Consignment or bailment language. State that the CM holds your tools as a bailee — your property in their custody — not as an owner.

MOQ, Pricing, and Payment Terms: The Three Cash-Flow Levers

Tooling ownership decides who controls your product. The next three terms decide whether you can afford to build it. Price, payment terms, and minimum order quantity (MOQ) are the most heavily negotiated terms in a contract manufacturing (CM) agreement, and they pull directly on your bank balance. Treat them as a single system, because a concession on one almost always shifts pressure onto another.

Start with payment structure, which is a cash-flow term before it is anything else. If your startup has no credit history, expect the CM to ask you to fund a larger deposit upfront and pay the balance on shipment. That is normal, but it means your money is committed weeks or months before you have finished goods to sell. Model that gap honestly, and negotiate milestone payments tied to production stages rather than one large deposit where you can.

MOQ is where founders get surprised, because the real minimum is rarely the number of finished units. Components arrive in reels and lots — often in increments of 1,000 or 5,000 — so your CM has to buy more than your forecast needs even when not all of it gets used. That gap becomes excess inventory, and the agreement has to say who owns the liability for it. The same problem grows when long-lead components are bought against your forecast, which can lock up large amounts of inventory before you have committed to the demand.

Watch for economic order quantities, too. A CM may order above your needs to capture better unit pricing and improve its own margin, then pass the cost through to you. Require your written approval for those purchases and lock the pricing basis in the agreement rather than accepting open-ended cost pass-through. For raw-material or currency-driven price changes, negotiate a cap on increases plus advance notice and audit rights, so a quoted price does not quietly drift upward between purchase orders.

Quality, Acceptance, and Warranty: Defining "Conforming" Before You Ship

A manufacturer's job is to build what conforms to your specification, but "conforms" only means something if your agreement defines it. Two mechanisms do that work, and founders routinely conflate them: acceptance governs whether you have to pay for a unit at all, while warranty governs what happens if a unit you already accepted fails later. Keep them separate in your drafting, because they trigger different remedies on different timelines. Getting the language right here is the difference between rejecting a bad lot and being stuck with it.

Acceptance: the trigger for your payment obligation

Acceptance is the legal event that obligates you to pay. Until a unit is accepted, you can reject it and refuse payment, so the acceptance clause is one of the few places where leverage sits with you rather than the factory. Define a fixed acceptance period as a set number of days after you receive the goods, and state that product not rejected within that window is deemed accepted. Most contracts land at 30 days, with a workable range of roughly 20 to 60 days depending on how long your incoming inspection actually takes.

Pair your acceptance period with a defined inspection method. The standard sampling framework, AQL under ISO 2859-1, lets you inspect a statistically meaningful sample rather than every unit, and naming it in the contract removes arguments about whether a lot passed.

Warranty: a promise that holds after acceptance

Warranty is a separate promise that accepted product will keep meeting its criteria for a defined period. Warranty remedies are generally limited to repair, replacement, or credit, so don't expect a warranty clause alone to cover field costs or recalls. Expect the manufacturer to push for exclusions, and accept the legitimate ones: defects traceable to your design, to test equipment or software you provided, or to a test plan you approved that failed to catch the defect. Note too that sample, first-article, and preproduction units are typically sold without any warranty at all.

When a warranty isn't enough: epidemic failure

Ordinary warranty covers units you know are defective. An epidemic or systemic failure clause (sometimes called an EFR clause) covers product suspected defective when too many units fail from a single root cause. The trigger is a defect rate above a set percentage within a fixed window — 90 days is common — plus a minimum number of failed units sharing that root cause. Once triggered, it unlocks remedies beyond repair-or-replace, such as an extended warranty or coverage of recall and field costs. For any product where a systemic defect could mean a recall, this clause is where you put the manufacturer on the hook for the cleanup.

Protecting Your IP: Your Design, Their Process, and Anti-Cloning Terms

Manufacturing means handing a stranger the complete blueprint for your product. Before you do, the agreement has to draw a clean line between what stays yours and what belongs to the factory. Your buyer IP covers the things that make your product yours: the product design, firmware, drawings, and any dies or molds. The contract manufacturer's process IP covers how they build things in general — their equipment, methods, and know-how. Spelling out which side owns what is the foundation everything else in this section rests on.

That line only matters if your designs stay confidential in the first place. A signed NDA or confidentiality agreement is the baseline mechanism for keeping your design data private and for defining exactly how the CM may and may not use what you share. Sign it before you send a single drawing, BOM, or firmware file — not after the first sample comes back. Once data is out without a confidentiality framework around it, you have lost the leverage to control it.

The structural risk you are guarding against is overbuild. When you outsource production, the CM can run your tooling beyond the order you actually placed and divert the surplus into gray-market or black-market channels as cloned or counterfeit product. Two contract terms blunt this: an explicit no-overbuild obligation that caps production at authorized quantities, and anti-cloning language that bars the factory from making your product for anyone but you. Critically, that anti-cloning protection has to survive termination. A rogue manufacturer's incentive to keep running your molds does not end when the contract does, so plan for how you stop gray-market production after the relationship is over.

One area where a hardware founder should bring in counsel rather than rely on a boilerplate clause is jointly-developed improvements. As a drafting principle, you keep rights in improvements to your own IP, the CM keeps its background process IP, and any IP the two of you genuinely develop together should be expressly assigned or licensed — never left to silence. Ambiguity here is where ownership fights start.

Supply Continuity: Capacity, Lead Times, Exclusivity, and the Exit

A signed manufacturing agreement does not guarantee that units keep shipping. Continuity runs on a forecast-commitment mechanic: you give your contract manufacturer (CM) a rolling forecast, the CM commits capacity and orders long-lead components against that forecast, and the agreement defines the liability windows and flexibility parameters that govern how much of that forecast you are actually on the hook for. The single most consequential line in that mechanic is where the boundary sits between a non-binding forecast and a firm purchase order. Everything on the firm side is your inventory exposure; everything on the forecast side is the CM's bet on your business.

That boundary matters because some components carry punishing lead times. It is common to see parts with 26 to 39 week lead times, meaning the CM is committing to buys more than half a year ahead of any firm PO you have placed. The further out the firm-order line is drawn, the more material you own before you have committed to building anything. Negotiate the liability window and the "drop-in" flexibility for changing quantities inside it as deliberately as you negotiate price, because this is where continuity and cash both live.

The structural defense against a single CM controlling your supply is dual sourcing — qualifying the same component or build at a second approved supplier. A second source preserves continuity if one line goes down, gives you real negotiating leverage, and keeps you flexible. The cost is duplicated tooling and a second qualification effort, which is why founders often defer it. Build the right to dual source into the agreement now even if you do not exercise it until volume justifies the expense.

Plan the exit before you need it. On termination, the long-lead components ordered against your forecast naturally transfer to a successor manufacturer — the new CM can simply buy them from the current one — so a clean transition clause plus a defined last-time-buy right turns a messy divorce into an orderly handoff. Last-time-buy and pushing back on exclusivity are practitioner conventions rather than legal requirements, but both follow directly from the same logic: do not let any single CM hold your product hostage.

Actionable Next Steps

You now have the five structural areas that determine whether a manufacturing and supply agreement protects your product or quietly hands leverage to your contract manufacturer. Before you sign, work through this checklist and confirm each point is on paper in the agreement itself, not in an email or a sales conversation.

1. Tooling title. Confirm you own the tooling you fund the moment payment clears, that every mold, fixture, and die is itemized on a tooling schedule, and that you have an express right to take possession of it and transfer it to another manufacturer on termination.
2. The IP split. Draw a clean line between what stays yours (design files, firmware, dies you paid for) and the CM's own process know-how. Require an NDA before you hand over any design data, add no-overbuild and anti-cloning covenants that survive termination, and put your design and BOM into escrow.
3. MOQ, price, and payment. Pin down the minimum order quantity, a defined excess-and-obsolete window, a locked price basis, and price adjustments that are capped and require advance notice rather than open-ended cost pass-throughs.
4. Quality and warranty. Set a defined acceptance period with AQL inspection, spell out the warranty scope, and include an epidemic-failure clause that shifts the cost of systemic defects back to the manufacturer.
5. Supply continuity and exit. Separate non-binding forecasts from firm orders, secure a dual-source right, and build in a transition obligation plus a last-time-buy window so you are never stranded.

You can self-assess the commercial levers and draft the tooling schedule yourself; those are business decisions you understand better than anyone. Bring counsel in on the language that carries real legal risk: tooling-title and termination-transfer terms, the IP split and anti-cloning covenants, and the warranty, epidemic-failure, and limitation-of-liability allocation, where UCC defaults and the manufacturer's standard boilerplate can work against you in ways that are hard to see until it is too late.