Three board spins and a failed thermal design are not waste. They are the evidence.
Iterative design and physical test is what the four-part test asks for. What decides a hardware claim is whether the work was U.S.-performed and who funded it.
Photo by Magnus Engø on Unsplash
A new or improved device, board, mechanism, or system is a business component. The work qualifies when your team faces genuine technical uncertainty - about whether a design is even capable, which method will work, or what the appropriate design is - and resolves it through a systematic process of experimentation: modeling, prototyping, and physical testing.
Every qualifying activity must pass the IRC §41 four-part test - permitted purpose, technological in nature (engineering and the physical sciences), elimination of uncertainty, and a process of experimentation. Hardware development maps to this naturally, because iterative design and test is a process of evaluating alternatives.
Crucially, qualification is about the process, not the result. A prototype that failed, a design you abandoned, and the materials you consumed proving out an approach can all be part of a defensible claim - provided the record shows what you were uncertain about and what you tried. That record is what substantiation actually means.
General and illustrative; eligibility depends on your facts.
Not activity categories - situations. If any of these read like a week you actually lived, there is very likely a qualified business component underneath it.
Three heatsink and airflow concepts were simulated, then built and instrumented on the bench. Two were discarded on the data.
Why it can qualify
The information you had did not establish a capable thermal design. Simulation plus physical test is a documented process of evaluating alternatives.
Stack-up, impedance targets, and routing were revised across three board spins, with measurements between each one.
Why it can qualify
The right design was not readable off the datasheets. Each spin is an experiment against a technical requirement.
Interrupt priorities, two scheduler approaches, and a DMA path were profiled on real hardware before the timing budget closed.
Why it can qualify
Firmware developed against hardware constraints relies on computer science and engineering principles and carries genuine method uncertainty.
Three power-management strategies were measured against an actual duty-cycle model. One had to be abandoned after the sleep-current numbers came back.
Why it can qualify
Improving performance is a permitted purpose, and the abandoned strategy is still qualified experimentation - the test is the process, not the result.
Candidate resins and wall sections were tested to failure until one design met all three constraints without a redesign of the internals.
Why it can qualify
Competing requirements with no known solution is the textbook elimination-of-uncertainty pattern.
Shielding, filtering, and layout changes were iterated - each measured - until the design passed with margin.
Why it can qualify
The redesign work to reach the requirement is the qualified part. The certification run itself sits much closer to routine testing.
The fixture and the process were re-engineered, with instrumented runs, before the build was released to production.
Why it can qualify
A process is a business component in its own right (§41(d)(2)(B)) - even when the product design is already frozen.
Illustrative situations, not client work. Whether any of them qualifies for you depends on your facts, your contracts, and your evidence.
Hardware programs generate the cleanest §41 record of any industry, and most of it is thrown away. Board revisions, bring-up logs, thermal and EMC runs, the qualification report that failed before one passed, the bill-of-materials lines consumed proving a build - each is a dated record of an alternative evaluated.
We read those artifacts next to the agreements that funded the program, because on a hardware claim the funding terms usually decide more than the engineering does.
Representative activities we see meet the four-part test across hardware, electronics, and mechanical engineering.
What spending counts toward the credit - tailored to how engineering teams actually spend.
| Expense category | What goes into the base |
|---|---|
| Engineering wages§41(b)(2)(A)-(B) | W-2 wages for engineers and technical staff performing, supervising, or directly supporting qualified development and testing. |
| Prototype supplies§41(b)(2)(C) | Materials and components consumed in building and testing prototypes - not depreciable equipment. |
| Contract engineering (65%)§41(b)(3) | 65% of amounts paid to U.S. third parties for qualified research and engineering performed on your behalf. |
| Cloud & compute§41(b)(2)(A)(iii) | Amounts paid to rent compute for simulation, modeling, and analysis used in qualified research. |
A directional shape for a hardware program, not a benchmark. Supplies matter far more here than in a software claim - which is also where the sloppiest numbers tend to be.
Prototype materials and components consumed in development and testing are eligible under §41(b)(2)(C). Depreciable equipment, tooling that becomes a capital asset, and general overhead are not.
The awkward case is the prototype that survives. A unit consumed in destructive test is a supply; a unit that becomes a demo rig, a fixture, or an asset on the balance sheet has to be treated as what it became.
Two of the six below are the same question asked twice, because it is the question that most often destroys an otherwise excellent hardware claim on examination.
§41(d)(4)(H)
A grant is not automatically fatal - and “we have an SBIR” is not an answer in either direction. Research another party pays for is funded, and excluded, unless you both bear the financial risk of failure and retain substantial rights in the results. Small businesses commonly do retain rights in SBIR data, so the analysis usually turns on risk: was payment contingent on the technical outcome, or due regardless?
§41(d)(4)(H)
The same test, one level down the supply chain. A cost-reimbursement subcontract that pays your engineering hours whether or not the unit performs, with the prime taking the design rights, is the hard case. Fixed-price development where you absorb the overrun and keep the IP reads very differently.
§41(d)(4)(D)
Incoming inspection, functional test on the line, and QC of units already in commercial production are excluded outright. The design-stage testing that told you what to change is a different activity with a different purpose - but it has to be documented as such.
§41(d)(3)(B)
Color, finish, and enclosure aesthetics are style and taste factors, excluded from a permitted purpose. If the new shape created a real thermal, structural, or manufacturability problem, the engineering that solved it is what you claim - not the styling that caused it.
§41(b)(2)(C)
Oscilloscopes, chambers, printers, and tooling that becomes a capital asset are depreciable property, which the supply definition excludes. The components and materials consumed in prototypes and destructive test are exactly what it includes.
§41(d)(4)(F)
Offshore design houses, contract firmware teams, and overseas subsidiaries are excluded regardless of how the contract is written or who supervises the work. The location of the research decides it.
Manufacturers face the same funding question in a different costume - customer-paid NRE and tooling. If you build as well as design, read the manufacturing page alongside this one.
§41(h) lets a qualified small business elect to apply up to $500,000 of its research credit per year against payroll taxes instead of income tax - the employer share of Social Security tax first, and the Medicare share above that. For a hardware startup that is burning engineering payroll years before revenue, that converts a credit it could not otherwise use into quarterly cash.
The definition catches people out. It turns on gross receipts under $5 million for the credit year and on not having had gross receipts before the five-year window ending in that year - so early prototype revenue or a small pilot sale years ago can close the door. The election is made on a timely-filed return, claimed on Form 8974 with your quarterly employment tax return, and cannot be made for more than five tax years.
Summary only - the statutory definition and the aggregation rules decide it. We test them explicitly.
A hypothetical scenario to show how the pieces fit together. It is not a quote, projection, or promise of results.
A company on these facts would also test the §41(h) payroll-tax election - the difference between a credit carried forward and cash in the next few quarters.
Illustrative only. Figures are hypothetical and rounded; contract engineering is included at 65%, no funded-research exclusion is assumed, and the federal credit commonly works out to roughly 6-10% of QRE depending on method, filing history, and the §280C election. Your result depends entirely on your facts. This is not a quote or a promise of any result.
IRC §174A restores immediate, full expensing of domestic research & experimental costs - including domestic engineering and software development - for tax years beginning after December 31, 2024. Captured alongside the §41 credit, you get the deduction and the credit.
Tell us about your products and development work, and we’ll map your qualifying engineering to the four-part test - screening the funding terms, and capturing §174A - reviewed and finalized by R&D experts and backed by Audit Protection. Contact us for pricing tailored to your study.
We typically reply within one business day.
