HMRC published its annual R&D tax relief statistics on 29 September 2026 at 09:30. The release covers 2024-25, and it carries counts no earlier edition could: the first published figures for the merged scheme, and the first to include Enhanced R&D Intensive Support.
HMRC’s own summary says so: this is “the first publication to include statistics for the new merged R&D expenditure credit (Merged RDEC) and enhanced R&D intensive support (ERIS) schemes”.
Until 29 September 2026, everything said about how the two current schemes have landed in practice was projection from the rules. There are now counts, though not yet a separate one for ERIS, as the scheme section below explains. The whole series, every year back to 2000-01 and every region from 2021-22, sits on our R&D statistics page, rebuilt from HMRC’s tables on publication day.
What this release is, and what year it covers
HMRC publishes these statistics once a year in the autumn. Each edition covers a single financial year that closed well before publication: September 2025 covered 2023-24, and this one covers 2024-25.
The lag is structural. A claim reaches HMRC with the company’s corporation tax return, and the claim window runs two years from the end of the period of account. A year’s claims are therefore not all in until long after the year has ended. HMRC allows for that by uplifting its headline counts for returns it has not yet received, and labels the figures provisional. It also rounds every figure independently, to the nearest 5 claims or £5 million, which is why the parts of a table need not add to its total.
Late returns also change earlier years. This release revises 2023-24 to 48,610 claims, £7.80bn of support and £47.8bn of qualifying expenditure. Every change below is measured against those revised figures, not the ones published a year ago.
One feature of this year matters more than the lag. The merged scheme and ERIS apply to accounting periods beginning on or after 1 April 2024. A period that began before that date stays under the old SME and RDEC rules however late it ends. HMRC counts each claim in the financial year in which its accounting period ends, so 2024-25 holds both systems at once: old-scheme claims from periods that began earlier, current-scheme claims from periods that began on or after 1 April 2024. The scheme figures below describe a transition year. Reading them as a clean first year of the merged scheme would overstate what they cover.
The clean year comes next. An accounting period cannot run longer than 12 months, so every period ending in 2025-26 began on or after 1 April 2024. That makes 2025-26 the first year in which every claim falls under the merged scheme or ERIS.
The headline numbers
| Measure, 2024-25 | Figure | Change on 2023-24 (revised) |
|---|---|---|
| Claims made | 40,325 | Down 17% |
| Total support claimed | £8.20bn | Up 5% |
| Qualifying R&D expenditure | £51.0bn | Up 7% |
| Average support per claim | £203,300 | Up 27% |
| First-time claimants, 2023-24 (latest year HMRC charts) | 5,250 | Down 43% on 2022-23 |
Claim volumes fell by a further 17%, on top of a 25% fall in 2023-24 on HMRC’s revised figures. At 40,325, the count is 38% below the 64,925 claims made in 2022-23. HMRC estimates that claims by SMEs fell 19% across every scheme while claims by large companies rose 4%, so the companies that stopped claiming were again overwhelmingly smaller ones. HMRC believes a key driver is the additional information form every claim must now carry. Whether the companies that left should have been claiming in the first place is the question the count cannot answer.
Read the claim count against the expenditure figure before drawing anything from either. Qualifying R&D expenditure of £51.0bn rose by 7%. Where the two move by very different amounts, the gap is telling you about the number of companies claiming rather than about the research being done. HMRC makes a similar point: with the schemes and rates changing between 2022-23 and 2024-25, it suggests expenditure is possibly the better guide to recent trends. The revised 2023-24 figures had the same shape, claims down 25% and expenditure up 2%. This year it is 17% fewer claims on 7% more expenditure.
The first-time-claimant figure carries a caveat that is easy to miss. HMRC uplifts its headline claim counts for returns not yet received, but it does not uplift the first-time-applicant table, so each year’s figure starts low and rises as late returns arrive. Last year’s release put first-time applicants in 2023-24 at 3,765; this year’s puts the same year at 5,250. That is still 43% down on 2022-23, and HMRC records it as the fifth consecutive annual fall. HMRC has left its partial 2024-25 figure out of its own chart, and so have we. The table does show 5,850, above 2023-24, but 4,530 of them sit in the RDEC and merged RDEC column, against 1,845 a year earlier, in the first year of the merged scheme. HMRC does not comment on the figure; it is not uplifted and is not comparable with earlier years.
What do the first merged scheme and ERIS figures tell a director?
They give a first count, not yet a verdict: 6,840 claims came through the merged scheme in 2024-25, 91% of them from SMEs, and ERIS has no separate count: HMRC reports it combined with the old scheme’s R&D-intensive SME claims.
The merged scheme is the general one: a 20% expenditure credit on qualifying expenditure, open to companies of any size. ERIS sits beside it for loss-making SMEs whose relevant R&D expenditure is at least 30% of total relevant expenditure, giving an 86% additional deduction with a 14.5% payable credit on the surrenderable loss. Whether ERIS is open to a company is decided by its size, its R&D intensity and its loss position. Our guide to which scheme applies works through the test.
6,840 claims came through the merged scheme in 2024-25, worth £1,425m, and HMRC estimates that 91% of them came from SMEs. The money ran the other way: SME claims carried £420m of that £1,425m, and the 625 large-company claims carried £1,005m.
ERIS has no count of its own yet. HMRC reports it in one column with the old scheme’s R&D-intensive SME claims: 4,965 claims worth £855m between them. The release cannot say how many of those were ERIS claims, and we have not estimated it.
One comparison to refuse. The current-scheme counts are not continuous with the old SME and RDEC counts: the old division was by company size, while the new one also turns on research intensity and profitability. A company that appeared under the SME scheme last year can appear this year under the merged scheme, under ERIS, or still under the old rules. HMRC says the same: SME-level figures for 2024-25 are not comparable with earlier years, and merged-scheme claims by SMEs are not comparable with earlier RDEC claims by SMEs.
That is the context for the scheme-level movements in the release. Relief under the SME scheme and ERIS fell 29% to £2.3bn, while relief under RDEC and the merged scheme rose 29% to £5.9bn. HMRC puts most of that down to SMEs claiming under the merged scheme and to the new rates applying for a full year. The comparison HMRC does allow is for large companies: their claims under RDEC and the merged scheme together rose 4%.
HMRC R&D statistics · by scheme · 2024-25 · provisional
Who claims, and where the money goes
Large companies made 9% of claims and received 56% of the relief. RDEC claims by large companies alone were 8% of claims and 44% of relief.
Share of claims40,325 claims
Share of relief£8.20bn of relief
SMEs made 90.6% of claims and received 43.9% of the relief. Large companies made 9.4% of claims and received 56.1%.
- SME scheme not R&D intensive 48.6% of claims and 17.4% of relief
- R&D-intensive SMEs and ERIS one HMRC column 12.3% of claims and 10.4% of relief
- RDEC, SMEs old scheme 14.2% of claims and 11.0% of relief
- Merged RDEC, SMEs current scheme 15.4% of claims and 5.1% of relief
- RDEC, large companies old scheme 7.9% of claims and 43.9% of relief
- Merged RDEC, large companies current scheme 1.5% of claims and 12.2% of relief
Hatched: current schemes, for accounting periods beginning on or after 1 April 2024. Dotted: HMRC's one column for R&D-intensive SMEs (old scheme) and ERIS (current scheme).
2024-25 is a transition year: accounting periods that began before 1 April 2024 still claim under the old schemes, so old and current schemes appear in the same year. HMRC says SME-level figures for 2024-25 are not comparable with earlier years. The figures are provisional, uplifted by HMRC to include an estimate of claims not yet received.
The figures behind the chart
| Scheme | Claims | Share of claims | Relief | Share of relief | Average per claim | HMRC cells |
|---|---|---|---|---|---|---|
| SME schemeold scheme | 19,600 | 48.6% | £1,425m | 17.4% | £72,700 | RD1!C15, RD1!D15, RD1!E15; RD2!C15, RD2!D15 |
| of which deduction only | 11,105 | Relief: £695m deduction and £730m payable credit. Relief for claims combining the two sits within those columns; HMRC publishes no relief column for combined claims | ||||
| of which payable credit only | 3,500 | |||||
| of which deduction and payable credit | 4,995 | |||||
| R&D-intensive SMEs and ERISold and current schemes, one column | 4,965 | 12.3% | £855m | 10.4% | £172,200 | RD1!F15; RD2!E15 |
| RDEC, SMEsold scheme | 5,740 | 14.2% | £900m | 11.0% | £156,800 | RD1!I15; RD2!H15 |
| Merged RDEC, SMEscurrent scheme | 6,215 | 15.4% | £420m | 5.1% | £67,600 | RD1!K15; RD2!J15 |
| RDEC, large companiesold scheme | 3,180 | 7.9% | £3,600m | 43.9% | £1.13m | RD1!H15; RD2!G15 |
| Merged RDEC, large companiescurrent scheme | 625 | 1.5% | £1,005m | 12.2% | £1.61m | RD1!J15; RD2!I15 |
| All schemes (HMRC total) | 40,325 | 100% | £8,200m | 100% | £203,300 | RD1!M15; RD2!L15 |
HMRC rounds every figure independently to the nearest 5 claims or £5 million, so parts need not sum to totals. The six categories sum to 40,325 claims and £8,205m of relief, against HMRC's totals of 40,325 and £8,200m; shares are of the six categories shown. The three kinds of non-intensive SME claim are merged in the chart because HMRC's relief table has no matching column for combined claims, so relief cannot be split three ways without estimating.
Source: HMRC, Research and Development Tax Credits Statistics: September 2026, tables RD1 and RD2. Shares and averages are LimestoneGrey's calculations from HMRC's figures. Contains public sector information licensed under the Open Government Licence v3.0.
Size, sector and where the money sits
Support has been concentrated in a small number of very large claims for years, and this release is no exception: 560 claims worth more than £2 million each account for 1% of all claims and 49% of all support. HMRC says the 27% rise in average claim value was driven by these claims. Their number rose from 495 to 560, and the relief they carried from £3.32bn to £4.03bn. 2024-25 was also the first year in which the 20% RDEC rate applied to a full year of expenditure, and RDEC is the scheme the largest claimants use.
At the other end, claims worth under £15,000 number 10,075, 25% of the total, down 24% on the year. The bottom of the distribution is where the fall showed first: on HMRC’s revised figures, sub-£15,000 claims fell 34% in 2023-24 against a 25% fall overall.
A further fall at the bottom of the distribution, and a faster one than the 17% overall, means the smallest claims are still leaving. That fall mixes companies that should never have claimed with companies that could have. The count cannot say how many of each left.
HMRC R&D statistics · by claim size · 2021-22 to 2024-25
Relief is concentrated in the largest claims
In 2024-25, the 58% of claims worth up to £50,000 carried 6% of the relief. The 6% worth £500,000 or more carried 69%.
2024-25: provisional, uplifted by HMRC to include claims not yet received
In 2024-25, the 58% of claims worth up to £50,000 carried 6% of the relief. The 6% worth £500,000 or more carried 69%. Select a point for a size band.
The figures behind the chart
| Claim size | 2021-22 | 2022-23 | 2023-24 | 2024-25 | ||||
|---|---|---|---|---|---|---|---|---|
| Claims | Relief | Claims | Relief | Claims | Relief | Claims | Relief | |
| Up to £5k | 12,500 | £30m | 7,530 | £20m | 4,420 | £10m | 3,280 | under £10m |
| £5k to £10k | 10,425 | £75m | 6,730 | £50m | 4,680 | £35m | 3,575 | £25m |
| £10k to £15k | 8,200 | £100m | 5,820 | £70m | 4,125 | £50m | 3,220 | £40m |
| £15k to £20k | 6,500 | £115m | 5,165 | £90m | 3,885 | £70m | 2,955 | £50m |
| £20k to £25k | 5,230 | £115m | 3,985 | £90m | 3,020 | £70m | 2,435 | £55m |
| £25k to £30k | 4,195 | £115m | 3,220 | £90m | 2,435 | £65m | 1,990 | £55m |
| £30k to £40k | 6,530 | £225m | 5,275 | £185m | 4,010 | £140m | 3,350 | £115m |
| £40k to £50k | 4,875 | £220m | 3,955 | £175m | 3,070 | £135m | 2,570 | £115m |
| £50k to £100k | 12,700 | £895m | 10,450 | £740m | 8,010 | £565m | 6,820 | £480m |
| £100k to £250k | 8,855 | £1,350m | 7,860 | £1,205m | 6,450 | £990m | 5,670 | £870m |
| £250k to £500k | 2,590 | £890m | 2,760 | £955m | 2,325 | £810m | 2,125 | £735m |
| £500k to £1m | 1,110 | £765m | 1,220 | £835m | 1,165 | £795m | 1,190 | £820m |
| £1m to £2m | 455 | £625m | 520 | £720m | 530 | £740m | 585 | £805m |
| Over £2m | 370 | £2,145m | 435 | £2,535m | 495 | £3,320m | 560 | £4,030m |
| Sum of bands | 84,535 | £7,665m | 64,925 | £7,760m | 48,620 | £7,795m | 40,325 | £8,195m |
| HMRC total (RD1, RD2) | 84,535 | £7,680m | 64,925 | £7,760m | 48,610 | £7,795m | 40,325 | £8,200m |
| Status | Revised | Provisional, Revised | Provisional, Revised | Provisional, Uplifted | ||||
Shares are of the sum of the bands, so each curve closes at 100%; HMRC rounds every figure independently to the nearest 5 claims or £5 million, which is why the bands and HMRC's totals can differ slightly. For 2024-25 HMRC prints the relief for claims up to £5,000 as “<10” in a £ million column, meaning under £10m; the curve counts it as nil.
Checked against HMRC's commentary. HMRC: 58% of 2024-25 claims were worth up to £50,000 and took 6% of the relief; 6% were worth £500,000 or more and took 69%. From the published cells: 57.97% of claims and 5.55% of relief (5.67% at most if the under-£10m band is added at its £10m ceiling); 5.79% of claims and 69.01% of relief. For 2023-24 HMRC gives 5% and 62%; the cells give 4.50% and 62.28%. Every figure agrees with HMRC’s after rounding to whole percentages.
Source: HMRC, Research and Development Tax Credits Statistics: September 2026, table RD7. 2022-23 to 2024-25 are provisional, and 2024-25 is uplifted for claims not yet received. Shares, cumulative shares and averages are LimestoneGrey's calculations from HMRC's figures. Contains public sector information licensed under the Open Government Licence v3.0.
By sector, manufacturing, information and communication, and professional, scientific and technical between them accounted for 75% of claims and 71% of support. Manufacturing was narrowly the largest by claim count at 11,090. HMRC assigns sector from the company’s registered industry code, which may not match the sector of its R&D.
Wales
HMRC allocates each claim to a region using the postcode of the claimant company’s registered office, and warns that this “might not correspond to where the R&D activity takes place”. A group registered in London running its laboratory in Swansea counts as London. The table is a map of where claimant companies are registered.
With that stated: Welsh-registered companies made 1,205 claims for £130m of support, on qualifying expenditure of £810m. That is 3.0% of UK claims and 1.6% of UK support. The average Welsh claim was £107,900, against a UK average of £203,300.
Wales has held roughly 3% of UK claims in each of the four years HMRC’s current tables cover. Its share of the money has been lower: about 2% in 2021-22 and 2022-23, then 1.6% in 2023-24 on the revised figures, and 1.6% again this year. The two shares have moved apart because the average UK claim has risen faster than the average Welsh one. Since 2021-22 the UK average has more than doubled, from £90,800 to £203,300, while the Welsh average has risen from £64,000 to £107,900.
Within Wales, South East Wales accounted for 640 of the 1,205 claims and £80m of the £130m. Our sub-region league table ranks all 46 UK sub-regions on claims and on relief, each figure traced to the HMRC table and cell it came from. It runs on the 2024-25 figures from HMRC’s September 2026 tables.
What the release says about error and fraud, and what it does not
These statistics count claims and money. They are not a measure of how many claims were right, and the two are routinely conflated. HMRC publishes its error and fraud estimates separately, in its annual report and accounts and in its approach-to-R&D-reliefs papers.
The release’s own error and fraud section confirms the separation. It points to the updated estimate in HMRC’s 2025–26 annual report, and says the statistics “do not include any assessment of the levels of error and fraud” in the schemes.
The figures that stand are these. Error and fraud in the R&D reliefs was measured at 17.6% in 2021-22 and at 6.4% in 2023-24, on random-enquiry evidence (for 2023-24, alongside a review of large-company RDEC claims). HMRC’s July 2026 annual report puts an illustrative 5.3% on the two years since, which is an estimate rather than a measurement. Separately, HMRC checked around one in six R&D claims (17%) in 2023-24, the most recent year it has published. That checking rate comes from HMRC’s approach paper, not from the 29 September 2026 release, and nothing in that release changes it. What one in six means in practice is set out in our guide to HMRC enquiries.
Should any of this change whether you claim?
Not on its own. A statistics release describes a population; it says nothing about whether a particular company’s work qualifies.
The test has not moved: an advance in science or technology sought against a stated baseline of what was already achievable in the field, technological uncertainties a competent professional could not readily resolve, and a record of the work done to resolve them. Every figure above aggregates claims that met, or failed, that test. None is evidence about yours.
One practical reading does follow. If your company’s period began on or after 1 April 2024, it is claiming under one of the two current schemes, and its size, R&D intensity and loss position decide whether ERIS is open to it.
We will not forecast next year’s figures from these. This is a transition year, the error and fraud series runs on its own clock, and the Autumn Budget is not something to guess at. HMRC plans its next release for Autumn 2027.
To see where your own company sits: check which scheme applies to you, the rates for your period, or talk it through with a Chartered Tax Adviser.
Sources
- R&D tax credits statistics, September 2026 — the release read here, with its commentary, its main tables (RD1, RD2 and RD4 to RD8; the former RD3 R&D-intensive figures are now folded into the main tables) and its supplementary tables (RDS1 to RDS3).
- R&D tax credits statistics, September 2025 — the first-time-applicant figure for 2023-24 as first published.
- HMRC annual report and accounts 2025–26 — error and fraud measured at 6.4% for 2023-24, with an illustrative 5.3% for the two years since.
- HMRC’s approach to R&D tax reliefs 2023–24 — the one-in-six checking rate for 2023-24.
This article describes the rules as they stood at the review date above. The rules change: for the current position, start with our guides or talk to us.
