R&D Tax Credits for Manufacturing Companies
Written and reviewed by the R&D Tax Accountants editorial team. Last reviewed 8 August 2026.
Manufacturing claims are usually about process rather than product. A company changes how something is made, to a tolerance, a yield or a throughput it could not previously achieve, and does not know at the outset whether the change will work.
Those claims are frequently strong and frequently unmade, because the activity looks like production improvement rather than research and because the costs are mixed in with ordinary manufacturing spend.
Process R&D in Manufacturing
The qualifying activity is the attempt to establish something that was not known. Whether a material could be formed to a tolerance without a defect mode appearing, whether a line could hold a yield at a higher rate, whether a substitution for a restricted material would perform in service, or whether two processes could be combined without one degrading the other.
Routine optimisation is not R&D. Tuning known parameters within known limits, or rolling out a proven method to a second line, is skilled production work. The test is whether a competent professional in the field could readily have said what would happen.
Failed process trials are among the best evidence a manufacturer has. A trial that produced an unexpected defect mode demonstrates that the uncertainty was real, and the costs of that trial are claimable even though the outcome was negative.
Trial Runs, First Articles and Scrap
The costing question in manufacturing is which parts of a mixed activity belong in the claim. Consumables genuinely consumed in development and trials qualify, which includes material scrapped in a trial run and the power and water consumed by it. Production and distribution of goods and services are excluded, so saleable output from the same run does not.
That means a trial that produced 200 parts, of which 60 were scrapped and 140 were sold, is not a single claimable cost. Separating it at the time is far easier than reconstructing it from stock records afterwards, and it is the single most useful change a manufacturer can make to support future claims.
First-article inspection and qualification testing sit on the qualifying side where they were part of establishing whether the process worked. Routine batch inspection of a settled process does not.
Scale-Up From Lab to Line
Scale-up is where a lot of genuine manufacturing R&D happens and where companies most often assume the R&D has already finished. A process proven at bench scale is not proven at line rate, and the uncertainty about whether it will transfer is a technological uncertainty in its own right.
Claims here often run across more than one accounting period, which brings the deadlines into play: 24 months from the last day of the period of account where that period is 18 months or less, and six months after the end of the period of account for a first or lapsed claim notification. A programme that runs for three years produces three claims, and the first expires while the work continues.
HMRC sets out the cost categories and the exclusions in its guidance on which R&D costs can be claimed. Manufacturers combining relief with grant-funded development should check the interaction, and the funding route is Innovate UK.