By Dan Jones, Head of Streets Innovation

If the R&D landscape has felt different this year, there is good reason. HMRC released statistics in September 2025 relating to R&D tax relief support claimed for the 2023 to 2024 tax year. The new picture is of a system that is still supporting substantial investment but is doing so through fewer, larger and better evidenced claims, while a tougher compliance stance is discouraging many first-time or marginal applicants. This is not an innovation collapse, it is a filter and strong claims are still going through but it’s clear that the market is shifting and claimants are levitating to high quality advisors, instead of cheap ones.

The headline numbers point in the same direction. Total support for 2023 to 2024 sits at around £7.6 billion, a modest decline on the prior year, while qualifying R&D expenditure is roughly £46.1 billion, which is only slightly down and indicates that real activity continues. What has shifted is the distribution, with total claims falling to about 46,950 – a drop of roughly 26 percent, driven by SMEs where claim numbers fell by around 31 percent and with RDEC claims slipping only about 5 percent. Value has migrated accordingly. RDEC delivered about £4.4 billion and now exceeds SME support at around £3.15 billion and the average claim size rose by roughly a third as the smallest submissions dropped away and the higher RDEC rate lifted outcomes.

Regional and sector patterns remain familiar, which matters for planning. London and the South East still dominate by count and value and Information and Communication, Manufacturing and Professional, Scientific and Technical continue to carry most of the activity. It is worth remembering that HMRC attributes by registered office rather than the location of labs or plants, so outreach should be shaped by where engineering and trials actually happen, not just where a company files its accounts.

The most noticeable change on the ground is the compliance environment. Every claim now requires an Additional Information Form and new claimants require Claim Notification Forms in advance of submission, enquiries are more frequent (albeit slowing) and more granular and there is closer scrutiny of subcontracting, connected parties and who truly bore the technical and financial risk. Many genuine claimants have felt that intensity, and it is clear that the combination of paperwork, uncertainty and the perceived tone of interactions has discouraged borderline and first-time filers. First-time claimants fell by about 41 percent in the most recent year reported, which is a striking signal that confidence has been shaken among smaller businesses that are not set up to evidence their work in a structured way. We’re aware there are many conversations ongoing about this, where first time claimants have been caught out by claim notification requirements, along with HMRC themselves getting things wrong!

That does not mean SMEs should step back from the regime, it means they need to approach it with a higher standard. R&D-intensive SMEs still have a viable route through the enhanced relief for intensive firms, with the intensity threshold set at 30 percent for periods beginning on or after 1 April 2024. The practical response is to model intensity during the year and to build the technical record as the work progresses, rather than attempting a reconstruction at year end, because contemporaneous records are precisely what today’s submissions demand.

For mid-market and larger groups, the centre of gravity is now the merged scheme, so the focus should be on disciplined project selection, clear narratives that establish the baseline and the advance, and tight cost traceability from first principles to final figures. Treat the claim as an audit-ready workstream, make sure contracts and scopes explain who did the R&D and why, and be explicit about why the uncertainties could not be readily resolved by a competent professional. If HMRC asks for clarification, respond quickly, keep technical leads in the conversation, and stick to facts, since that combination usually shortens the process and improves outcomes.

Step back from the individual levers and the message is consistent. The UK is still investing nearly fifty billion pounds a year in qualifying R&D, but the system now rewards documentation and discipline, with much less tolerance for generic descriptions or light-touch submissions. More value is flowing through robust RDEC-type profiles, and, for SMEs that genuinely live and breathe R&D, enhanced support remains available provided they plan and evidence properly. If you invest seriously and can show it, there is still meaningful value on the table. If you cannot, this year’s statistics tell you what happens next.

Here is a practical way to adapt. Front-load eligibility decisions and select the correct route early, test ERIS intensity during forecasting instead of at year end and build the claim pack throughout the project lifecycle so the story is detailed rather than retrospective. Strengthen subcontracting evidence by mapping the chain, clarifying scopes and identifying who bore risk and prove the state of the art with a crisp explanation of why the advance was not readily deducible. Above all, favour quality over quantity, prioritising projects with clear uncertainty and clean cost trails and retire marginal candidates that will soak time for little return. Leadership teams should back this by resourcing the record-keeping properly, aligning finance, project and technical leads on what the claim requires and when, and adding R&D tax checkpoints into stage gates so documentation does not become a year-end scramble.

Key figures to remember: about £7.6 billion of relief in 2023 to 2024, around 46,950 claims in total, down 26 percent, SME claims down about 31 percent, RDEC up to roughly £4.4 billion overtaking SME at about £3.15 billion, average claim size up by around one third, first-time claimants down about 41 percent and three sectors accounting for more than two thirds of activity while London and the South East remain dominant, even though the registered office is not the lab.

My view, plainly: compliance first, quality claims, no shortcuts. If your work advances science or technology and the evidence is in order, you will still do well. If not, these statistics are the writing on the wall.