The recent HM Treasury R&D Tax Reliefs Report has highlighted an interesting point on the future inclusion of cloud computing services as eligible costs within the R&D tax regime. The document suggests that while cloud computing costs will become eligible from April 2023, any costs relating to cloud servers and storage (defined by the Treasury as general overheads) will remain ineligible for R&D tax purposes.
The fact that the Treasury has identified one category of cloud services as eligible and another as ineligible for R&D tax purposes poses several questions and could make claiming for cloud costs an extremely complex process for many companies. Amazon Web Services (“AWS”), for example, is a cloud provider that many companies use to facilitate their R&D activities. AWS identify several product categories for their services, including “Compute”, “Storage”, “Database”, “Machine Learning” and “Blockchain” among others. Looking at the specific services within each of these categories provides an insight into what these services are used for and how they could contribute to R&D activities. The Compute category, for example, includes services such as:
- AWS Lambda (used to run serverless processes and code for applications and backend services),
- AWS Elastic Beanstalk (used to run and manage web applications), and
- Amazon Elastic Container Service (used to run and manage Docker Containers).
These are all examples of Software as a Service (“SaaS”) and Platform as a Service (“PaaS”). It is expected that where these “computing” services will be used directly for R&D, the relevant portion of these costs will be included within a R&D tax claim under the software cost category. The point to note, is that all these services are ultimately hosted on Amazon’s servers and any data will be stored on their databases. This includes services within both the Storage and Database product categories and so these costs will remain ineligible under the proposed rules. As these cloud computing services, and the hosting and storage of them, are so interconnected it will be extremely difficult for businesses to separate and accurately apportion the R&D activities and associated costs between them. A prime example of this is Elastic Beanstalk, which is hosted on either EC2 instances or S3 buckets (examples of storage services). AWS do not charge for the use of AWS Elastic Beanstalk; customers pay for AWS resources (such as EC2 instances or S3 buckets) created to store and run their applications.