1. News
  2. Data, cloud and pure maths: What the R&D tax relief changes mean for tech companies

Data, cloud and pure maths: What the R&D tax relief changes mean for tech companies

Tom Heslin technical specialist
(Last updated on )
Read time:
Data Cloud

The government has made a clear statement of intent to support the technology sector with its upcoming reforms to R&D tax relief. As a result of the soon-to-be widened scope and reduction in ambiguity about what can be included in R&D tax relief claims, the way ahead is clearer for tech innovators.

For many years, two stumbling blocks lay in the path of businesses using artificial intelligence (AI), quantum computing and pure mathematics when looking to embrace the incentive.

  • That the often-significant costs of using cloud computing services do not qualify for R&D tax relief.
  • The omission of advances in pure mathematics from the definition of qualifying R&D activities.

These issues were highlighted by ForrestBrown and the wider industry, recognising that cloud computing and innovation have become ever more entwined. After all, why should an AI innovator be out of pocket for the mathematics or computing power needed to develop new solutions?  

Changes to R&D tax relief 2023: cloud, data and maths

After acknowledging the feedback received as part of a wider consultation on the future of the R&D tax relief incentive, changes announced in the Spring Statement 2022 will see both obstacles addressed in policy from April 2023. 

As a result, businesses using AI, quantum computing and pure mathematics, can apply for R&D tax relief with confidence and can consider making greater use of cloud computing to develop new products, processes and solutions.

With a wider scope of relief comes greater risk of scrutiny from HMRC. Therefore, as with all claims, an approach of due care must still be taken.

What these changes to R&D tax incentives mean in practice

This widening of the incentive will lead to new opportunities for tech firms once it comes into effect from April 2023. The inclusion of pure mathematical advances and licence payments for datasets used in R&D will unlock value in some cases. However, it is likely that the addition of cloud computing costs will have the broadest impact on R&D tax relief claims for companies working at the frontiers of these sectors.

Cloud computing costs & R&D tax relief

As most businesses in the tech sector will know, cloud computing isn’t a marginal cost. The sheer power that goes into creating, maintaining, running and storing machine learning (ML) and AI solutions, training datasets, models and simulations can be a huge overhead that is essentially unavoidable in spaces such as quantum computing.

For this reason, while a few smaller companies may revaluate cost effectiveness if they still house their own servers, for much of the sector this change will not introduce them to cloud computing. The impact, however, will be that firms can start to reclaim some of these costs back, allowing them to redirect funds to other areas of R&D or allow them to put additional budget behind computer power to more quickly and effectively develop new solutions.

Pure Mathematics & R&D tax relief

The inclusion of activities seeking advances in pure maths will allow businesses to claim with more confidence. In a commercial context, it is most likely that work in pure maths would already be linked to software development, with translating those advances into software being the primary aim of such projects. Often, the boundary between the two is not clear, which, without this update, potentially left R&D claims open to searching questions from HMRC.  With the disambiguation of the definition, however, tech firms can be confident in knowing that the government wants them to have access to funding and that they have clear policy on their side. 

Dataset costs & R&D tax relief

The addition of costs relating to licence payments for datasets is a positive move from HMRC and one that should benefit tech companies, in particular those operating in AI and machine learning. Being able to include the purchase of data to be used in training datasets, for example, could make it more affordable to build and improve AI using existing data. However, we’re still awaiting details of how these changes will be interpreted by HMRC. This is particularly key as some of the proposed exclusions appear to be excessively restrictive to the extent they limit the genuine application of the benefit. We know that HMRC wants to avoid overly broad cost categories due to the risk of abuse, but it is important that companies are able to claim relief for these costs with certainty. That’s why we’re pushing for them to provide definitive guidance as soon as possible.  

Our Technical Director Jenny Tragner has attended a number of round tables as part of the consultation on these measures. These discussions have looked at the use of datasets across multiple projects for example, the timing of licence agreements within R&D project boundaries, and what restrictions might be reasonable to limit sharing of the data with other parties.

Eventually, we expect a “practical use” position to be reasonable, where a dataset that is clearly an essential part of one or more R&D projects will attract relief. Our current understanding of dataset costs and R&D tax claims is as follows, but may evolve as the legislation and accompanying guidance is released:

  • If you import data and use “as is” in your live system with no processing, then it’s not used for R&D. 
  • However, if you do some transformation/enrichment of the data, where the source dataset was essential to an R&D process, then importing the dataset along with enrichment fields can be seen as an essential mapping process to use in the output of the R&D project, meaning this type of expenditure should attract relief. 
  • If the source dataset is not used after the R&D – essentially because it is consumed by the R&D process, such as being used for training a machine learning model – then associated expenditure should be eligible.

Although these changes should give tech and software companies more confidence, the introduction of any changes to the incentive is a good opportunity to ensure that your R&D tax claims are being carefully prepared. HMRC’s guidance and approach to these newly introduced activities and costs will evolve as they review more R&D claims, so make sure that you have a balanced and pragmatic approach to risk in your claim methodology.  

Be proactive: How to approach R&D claims in 2023

While the government has shown that it is committed to providing support to innovative tech firms, we know that HMRC is focused on tackling errors and abuse in R&D tax relief claims. It is reviewing more claims than ever before to ensure that the incentive isn’t being misused.

It is likely that HMRC will pay particular attention to these new rules, as the scope for error or abuse is higher where there isn’t established guidance and mutual understanding of interpretation. When tackling a first R&D claim for data, cloud or pure maths, a good first step would be to think about record keeping. For example, consider splitting financial records for cloud computing costs into amounts relating to R&D and normal day-to-day requirements. This could make it much easier to calculate your claim and provide you with robust evidence should HMRC review it.

ForrestBrown – a trusted adviser for tech companies

If a business is looking to claim significant sums of money or is thinking about making a first R&D claim, it might make sense to partner with a trusted adviser. When looking for the right partner, don’t assume that all advisers are the same. Making sure that the adviser’s credentials are properly considered will leave you with peace of mind that any claim they produce is watertight. Chartered tax advisers are technically qualified, committed to CPD and subject to ethical oversight from a professional body. You can be reassured that they will operate with due care, and if anything goes wrong, that body can step in to help you.

Advisers with genuine sector experience will know the ins and outs of the changes and will be able to advise on the best course of action for your business on a case-by-case basis. Keep in mind that a one size fits all approach is less likely to stand up to scrutiny from HMRC, so having an adviser on hand who understands how to apply the new rules to your case will set you up for success.

At ForrestBrown, we welcome these changes, but recognise that there is still significant untapped opportunity to use R&D tax relief to supercharge UK business investment in R&D. Nevertheless, these upcoming changes to R&D tax incentives prove that the government is willing to step up its support for the UK tech sector as it continues to thrive. With the proper advice and approach, the new, clearer road to R&D tax relief is an exciting prospect for those already innovating and those hoping to.