A common question is whether cloud scaling and performance development can qualify as research and development for the federal R&D tax credit under Section 41. The short answer is that cloud scaling and performance development may constitute qualified research when it involves a technical uncertainty and a process of experimentation. Routine cloud configuration and deployment generally is not qualified research, and internal-use software rules may apply. This page explains the framework in general terms. It is educational and is not individualized advice. For the foundational framework, see our page on qualified research.
When Cloud Scaling Development May Warrant Review
Cloud scaling and performance development may warrant review when the work involves a genuine technical uncertainty and a process of experimentation. Under the four-part test, the work must be for a permitted purpose, be technological in nature, be intended to eliminate uncertainty, and be conducted through a process of experimentation.
Common scenarios that may warrant review include:
- New scale — evaluating alternative approaches to resolve uncertainty about whether a system can scale to a new level.
- Distributed system — testing alternative distributed-system approaches to resolve uncertainty about whether a system can achieve the required performance across multiple nodes.
- Auto-scaling — evaluating alternative auto-scaling approaches to resolve uncertainty about whether a system can scale automatically to the required performance.
- Multi-region — testing alternative approaches to resolve uncertainty about whether a system can achieve the required performance across multiple regions.
Routine Configuration vs. Scaling Development
A central distinction is between routine cloud configuration and scaling development:
- Routine configuration — configuring standard cloud services using established methods for a known application. There is no technical uncertainty. This is configuration, not research.
- Scaling development — developing new scaling approaches where there is a technical uncertainty about whether the system can achieve the required performance, and evaluating alternatives to resolve that uncertainty. This may warrant review.
Hypothetical Example
Consider a company that is developing a new distributed system to handle a new class of high-volume workloads and is uncertain whether any available approach can achieve the required throughput and latency across multiple regions. The company evaluates alternative architectures, tests each, and systematically varies the approach to resolve the uncertainty. This systematic evaluation of alternatives may warrant review as qualified research.
By contrast, if the same company configures standard auto-scaling on a cloud platform for a known application, that is routine configuration, not research.
This example is illustrative only and does not state that the activity definitely qualifies.
Documentation That May Help
Records that can help support cloud scaling development claims include design records identifying the uncertainty and alternative approaches, performance test results, and records of how results informed architecture decisions. For more, see our page on R&D tax credit documentation.
Key Takeaway
Cloud scaling and performance development may constitute qualified research when it involves a technical uncertainty and a process of experimentation. Routine cloud configuration and deployment generally is not qualified research, and internal-use software rules may apply. Because the distinction is fact-specific, professional review is appropriate before claiming the credit.