R&D tax relief for agritech companies

Agritech work often qualifies for R&D tax relief, and it usually comes with the evidence a claim needs: replicated field trials, recorded growing protocols and the data trail regulatory approval demands. Work in crop science, precision farming, controlled environment growing and livestock technology regularly meets HMRC’s definition of qualifying R&D, and the relief is worth between 14.7p and 26.97p per £1 of qualifying spend depending on the scheme. The harder questions are about boundaries: where the science ends, who claims when the money arrives as a grant, and how to evidence work that runs on growing seasons rather than sprint cycles.

What agritech work qualifies as R&D?

Work that seeks an advance in science or technology by resolving an uncertainty a competent professional in the field could not readily settle. A claim answers four questions per project, in order. What the field could already do, from published knowledge and what a competent professional could deduce from it — the baseline. What the work set out to add — the advance. What was not known about whether that was achievable, or how — the uncertainty. And what was attempted to settle it, failures included — the resolution.

Agritech meets that test across both of its halves, the biology and the engineering. At concept level, qualifying work looks like this:

• Precision agriculture. Variable-rate application, yield prediction and decision models whose field accuracy cannot be deduced from published work, and which degrade in ways the training data did not anticipate.
• Sensing and imaging. Spectral, acoustic and in-soil sensing where the signal has to be separated from canopy, weather and soil variation before it means anything.
• Controlled environment agriculture. Interactions between light, nutrient delivery and energy use in indoor growing systems where no published data settles the design.
• Breeding, biologicals and crop protection. Trait selection methods, biological control agents and treatment regimes where existing agronomy cannot predict the outcome and the mechanism has to be established rather than looked up.
• Robotics and automation. Field robotics, selective harvesting and weeding in unstructured environments, where the difficulty is the system’s combined behaviour rather than any one component.
• Soil and water. Nutrient cycling, irrigation scheduling and run-off management where the field’s models do not predict behaviour for the soil, catchment or cropping system.
• Livestock technology. Monitoring, welfare and yield systems where biological variability defeats standard approaches.

One of the Guidelines’ own worked examples is an agrochemical one. Finding a new active ingredient for a weed-killer and developing a formula around it is an advance, and systematically testing the resulting formulations for performance, toxicity, solubility and damage to other plants is R&D even though the testing uses established methods. Assessing what the product needs to appeal to consumers is not.

The advance must be to the field’s knowledge or capability, not just your company’s. Adapting knowledge from another field counts, but only where the adaptation was not readily deducible, and much of agritech sits on that line: machine vision written for a factory line, put to work in a canopy. Applying an established sensor platform to a new crop is unlikely to qualify on its own; developing the platform because no existing one survives the conditions is a different matter. Our guide to what counts as qualifying R&D sets out the test in full.

When is a field trial R&D, and when is it agronomy?

When it is run to settle a question the field cannot already answer, and only until that question is settled. Testing that directly contributes to resolving the uncertainty is R&D. Testing after it has been resolved is not, however necessary and however expensive.

From the tramlines, a qualifying trial and ordinary agronomy look identical. A replicated trial designed to establish whether a treatment works, and why, sits inside the claim. A demonstration plot run to show growers a product that already performs sits outside it. Variety and product comparison trials run to a standard protocol, on standard plots, to produce the data a customer expects, are ordinary agronomy.

Regulatory trials follow the same rule rather than a separate one. HMRC’s compliance guidance is direct: obtaining certification for a product which already has proven functionality does not qualify, and where certification requires further advances to materially improve functionality, the work aimed at those advances does. Efficacy, residue and safety packages are not R&D because the regulator demands them; they are R&D to the extent they resolve something still open. The same reasoning applies to the release and marketing notification route for precision bred plants in England, which Defra operates under the Genetic Technology (Precision Breeding) Act 2023 and its 2025 regulations. Developing the trait is the R&D. Notifying the release and clearing that route is compliance.

R&D also has an end: when the knowledge is codified in a form a competent professional can use, or when a prototype or pilot plant with all the functional characteristics of the final product exists. In a seasonal business, that end point rarely falls at a year end.

Do pilot plots and prototypes that produce a saleable crop still qualify?

Yes, with one cost category carved out. Designing, building and testing a prototype rig, growing module or machine is R&D, and a later sale does not take that work back out of the claim. What drops out is the consumable materials that ended up inside the item sold. A pilot plot works the same way: seed, nutrient and crop protection consumed in the trial qualify, except for the proportion absorbed into produce transferred in the ordinary course of business. Output scrapped, ploughed in or sold only as waste is unaffected, and where only part is sold, only that part of the cost comes out. The harder case is the crop or unit that was always going to be sold, where the restriction reaches wider than materials. Can I claim R&D tax relief on a prototype that is later sold? works it through.

Does grant funding reduce an agritech claim, and who claims in a consortium?

Grant funding does not reduce the claim. Who claims is a separate question, and it is answered below. Under the old SME rules a grant did reduce relief, and for accounting periods beginning on or after 1 April 2024 those rules are abolished: a grant-funded project claims in full under the merged scheme or ERIS. Much of the advice still circulating online predates the change. See grant funding and R&D tax relief, and Innovate UK grants and R&D tax relief together.

Much of the public money reaching English agritech comes through Defra’s Farming Innovation Programme, delivered in partnership with Innovate UK, part of UK Research and Innovation. It funds by competition, and the strands differ by scale and by who can lead: Research Starter awards led by farmers, growers and foresters; feasibility studies and small and large R&D partnerships led by UK businesses; longer Farming Futures projects; and the ADOPT fund for on-farm adoption of existing solutions. Collaboration is a condition of funding. A grant does not certify that the work is R&D for tax purposes. ADOPT is the clearest case: where the work is adopting a proven solution, there is no advance to claim.

Who claims turns on contracts rather than on where the money came from. Where a customer contracts for activities, and it is reasonable to assume from the contract terms and surrounding circumstances that the customer intended or contemplated that R&D of that sort would be done, the customer claims and the contractor does not. Where the customer did not, the contractor claims on its own costs. A grant offer letter does not settle this; in a consortium the collaboration agreement and the partner-to-partner work packages do.

Agritech consortia raise three points the general rule does not. Universities and other institutions of higher education, charities, scientific research organisations and health service bodies cannot claim R&D relief at all, and a company doing R&D contracted out to it by one of them can claim in its own right. So can a contractor whose customer is not acting in the course of a trade, profession or vocation within the charge to tax — the question to ask where a levy body or public funder commissions work directly rather than awarding a grant. And work done at the company’s own risk before any contract exists stays its own R&D even if a contract follows and later reimburses it. Contracted-out R&D sets out the intended-or-contemplated test in full. Get it wrong in either direction and the same work is claimed twice or by nobody.

How do growing seasons affect a claim?

They compress your chances to experiment, which makes contemporaneous records more valuable, not less. A field trial may give you one data point per year, so hypotheses, protocols and results should be captured as the season runs rather than reconstructed at year end. Trials that fail still qualify: the relief rewards the attempt to resolve the uncertainty, not the outcome.

Apportionment is the harder half, because agronomists, technicians and engineers move between trial work and commercial production inside the same week. HMRC’s compliance guidance accepts an estimated proportion of known expenditure where the estimate is arrived at using evidence and reason and based on facts. It recommends recording the claim methodology and the method behind each apportioned figure. Trial plot records, spray diaries and equipment logs usually supply it.

Which costs go into an agritech claim?

Staff costs apportioned to development time, covering field and glasshouse technicians and trial agronomists as well as the lab and the data team. Consumables carry unusual weight: seed, growing media, nutrient, crop protection product, feed, and the water, fuel and power consumed by the R&D, apportioned on a basis a reader can follow — plot areas, treatment counts or run hours rather than a round percentage of the farm bill. Agency staff enter as externally provided workers at 65% of payments to unconnected providers, and only where those workers’ earnings bear UK PAYE and Class 1 National Insurance, or the overseas conditions below are met. Unconnected subcontracted work, contract research organisations and trial sites included, enters at 65% too; connected parties are restricted to the lower of the payment and the other party’s own relevant expenditure.

Software used for R&D has always qualified, and data licences and cloud computing joined it for accounting periods beginning on or after 1 April 2023. The category is larger in agritech than most people expect: licensed satellite imagery and weather series, agronomic datasets, the compute behind model training. Capital expenditure does not qualify, whatever the glasshouse or rig cost, though capital spending on the R&D itself can attract R&D allowances instead; rent and patent costs qualify for neither, though patent rights have their own allowances. The category rules sit in which costs qualify for R&D tax relief.

Who is the competent professional in an agritech claim?

Whoever holds the relevant expertise, which in agritech is usually more than one person: the plant breeder or crop scientist speaks to the biology, the agricultural or control engineer to the machine, the data scientist to the model. HMRC expects each to be knowledgeable about the relevant scientific and technological principles, aware of the current state of knowledge, and to have accumulated experience and a recognised track record. Having worked in a field, or taking an intelligent interest in it, does not by itself qualify someone. Their opinion has to explain, without jargon, what the advance is and why it advances the field’s knowledge rather than the company’s own. A bare assertion that the project was R&D is unlikely to be enough. Because that judgement carries the claim, we interview those people directly rather than working from a project list.

Which R&D scheme applies to agritech companies?

It depends on profitability and R&D intensity. Profitable agritech companies claim under the merged scheme: a 20% expenditure credit worth £20,000 gross on £100,000 of qualifying spend, which nets to £15,000 at the 25% corporation tax rate, £14,700 at the 26.5% marginal rate where augmented profits fall between £50,000 and £250,000, or £16,200 where the 19% rate applies. Loss-making companies under the merged scheme also receive £16,200 net in cash on the same spend.

Pre-revenue agritech companies should look hard at Enhanced R&D Intensive Support (ERIS). A loss-making SME whose relevant R&D expenditure is at least 30% of its total relevant expenditure receives up to 26.97p per £1: on £100,000 of qualifying spend, a £26,970 payable credit. That denominator is broadly the trading costs in its accounts for the period, not just the R&D ones, and many venture-backed agritech businesses pass comfortably in their development years. A later dip need not cost the relief, but the grace period runs for a single year, and only for a company that met the intensity condition in its most recent prior 12-month accounting period and obtained relief for it. The ERIS intensity calculator works the 30% ratio, and the claim value calculator estimates both schemes on your own numbers.

Can overseas field trials qualify?

Sometimes, under a deliberately narrow exception. The default rule is that contractor and externally provided worker costs qualify only where the R&D is undertaken in the UK. The exception applies where conditions necessary for the R&D are not present in the UK, are present where the work is done, and would be wholly unreasonable to replicate here. Those conditions include geographical, environmental or social ones, and legal or regulatory requirements as a result of which the R&D may not be undertaken in the UK. Trialling a crop system in a climate the UK does not have is the case the exception was written for. Cost and workforce availability are expressly excluded as justifications. The detail is in overseas R&D costs.

How we work with agritech companies

LimestoneGrey is a firm of Chartered Tax Advisers and Chartered Accountants, regulated by ICAEW, specialising in R&D tax relief, with a deliberate focus on R&D-intensive companies, agritech among them. Every claim is prepared by our specialist team and signed off by a chartered adviser, enquiry support is included as standard, and the fee is agreed before work starts. Preparing the claim includes telling you which trials do not qualify.

Two compliance points are worth flagging early. A company claiming for the first time, or that has not claimed in the three years ending with the notification deadline, must notify HMRC within six months of the end of the period of account, or the claim is invalid: see claim notification. And HMRC checked around one in six R&D claims (17%) in 2023-24, the most recent year it has published, which is why agritech claims should be built on trial records, not year-end recollection.

If you are weighing a first claim, an ERIS position or a grant interaction, get in touch for a considered view from a chartered adviser.

Sources

• Guidelines on the meaning of R&D for tax purposes — paragraphs 6 and 9 on the advance and on adaptation from another field where not readily deducible; 13 and 14 on uncertainty and on fine-tuning; 20 and 26 on overall knowledge or capability and on direct contribution; 27(c) and 28(c) on testing that directly contributes and on production and distribution; 29 and 30 on system uncertainty; 34 and 39 on where R&D ends and on prototypes; and example E, the weed-killer active ingredient, the formulation testing that is R&D and the consumer-appeal work that is not.
• GfC3: How to identify qualifying R&D activities (part 4) — testing after the uncertainties are resolved does not qualify, and obtaining regulatory certification for a product with already proven functionality is not R&D unless certification requires further advances that materially improve functionality.
• GfC3: Recommended approach to claims and record keeping (part 5) — estimates arrived at using evidence and reason and based on facts, and HMRC’s recommendation to record the claim methodology and the method used for each apportioned figure.
• Check what R&D costs you can claim — the cost categories, consumable items including fuel and power, and the 65% contractor rule. CTA 2009 s1125 — computer software at subsection (1)(a) from the outset, with data licences and cloud computing services inserted at (1)(aa) and (ab) by Finance (No. 2) Act 2023 Sch 1, with effect for accounting periods beginning on or after 1 April 2023; subsection (2), consumable materials including water, fuel and power.
• CTA 2009 s1132A — subsection (2), earnings are qualifying where PAYE and Class 1 National Insurance must be accounted for; subsection (3), where it does not apply, earnings attributable to R&D undertaken outside the UK to which section 1138A applies are still qualifying. CTA 2009 s1138A — subsection (2), conditions not present in the UK, present in the location used and wholly unreasonable to replicate here; (3)(a), the geographical, environmental, social and legal or regulatory limbs; (3)(b), cost and worker availability excluded.
• CIRD161000 — the intended-or-contemplated test, contractual chains, pre-contract scoping work remaining the contractor’s own R&D even where later reimbursed, and claims by a contractor for an irrelievable client, including a customer not acting in the course of a trade, profession or vocation within the charge to tax. CIRD163000 — ineligible companies: charities, institutions of higher education, scientific research organisations and health service bodies.
• CIRD81300 — the competent professional: knowledgeable about the principles, aware of the current state of knowledge, with accumulated experience and a recognised track record, and what their opinion must explain.
• CIRD82300 and CTA 2009 s1126A — consumables absorbed into items transferred in the ordinary course of business, the waste position, and apportionment where only part of the output is transferred.
• Farming Innovation Programme — Defra’s programme delivered with Innovate UK, part of UKRI, for farmers, growers and foresters in England; the Research Starter, feasibility, small and large R&D partnership, Farming Futures and ADOPT strands, who can lead each, and the requirement to collaborate.
• Precision breeding register: notices and decisions — Defra’s register of release and marketing notices under the Genetic Technology (Precision Breeding) Act 2023 and the Genetic Technology (Precision Breeding) Regulations 2025; currently only precision bred plants can be released or marketed in England.