The R&D tax credit for devtools and infrastructure companies

Distributed systems, storage engines, query planners, and schedulers rest on computer science and proceed through measurement and iteration: design an approach, benchmark it, find where it breaks, redesign. The four-part test is met comfortably by this kind of work.

The uncertainty is unusually deep. Whether a consensus protocol holds under partition, whether a storage layout meets a latency target, whether a scheduler stays correct under load, none of these are knowable without building and testing.

Developing new architectures, performance and scalability engineering, novel algorithms, and reliability work under genuine uncertainty generally qualifies.

Running the service, routine operations, documentation, and packaging do not. A strong claim separates the experimental systems work from the operational work around it.

Infrastructure work leaves a rich trail: benchmark commits, reverted approaches, performance branches, and design changes. That trail is the evidence of resolving uncertainty.

R&D Binder reads it from GitHub directly and maps each piece to a business component and the four-part-test evidence behind it.

• What is a business component?
• What is a qualified research activity?
• The Section 41 four-part test
• R&D tax credit for AI companies

• 26 U.S.C. § 41, credit for increasing research activities - Cornell Law School, Legal Information Institute
• Treas. Reg. § 1.41-4, qualified research and the process of experimentation - Cornell Law School, Legal Information Institute
• IRS, Instructions for Form 6765 - Internal Revenue Service