What Qualifies for the R&D Tax Credit

The federal R&D tax credit under IRC Section 41 is not limited to laboratories, patents, or breakthrough discoveries. It applies to a broad range of business activities across many industries — if they meet the IRS four-part test.

The four-part test

Every qualifying research activity must satisfy all four criteria defined in IRC Section 41(d):

1. Business component test

The research must relate to a new or improved business component — a product, process, technique, formula, invention, or piece of software that the taxpayer offers for sale, lease, or license, or uses in its trade or business.

This includes:

• Developing a new product from scratch
• Improving the performance, reliability, or quality of an existing product
• Designing a new manufacturing process
• Developing internal-use software that supports business operations

2. Technological uncertainty

At the outset of the research, there must be uncertainty about one or more of:

• Whether the desired result can be achieved
• The method for achieving it
• The appropriate design of the business component

The uncertainty does not need to be novel to the industry — it only needs to be uncertain to the taxpayer. If your team did not know in advance whether a particular approach would work, that counts.

What does not qualify: applying established techniques with a known outcome, routine quality testing, or cosmetic changes that involve no technical uncertainty.

3. Process of experimentation

The taxpayer must engage in a systematic process to evaluate one or more alternatives to resolve the uncertainty. This includes:

• Modeling or simulation
• Systematic trial and error
• Testing prototypes or formulations
• Evaluating design alternatives
• Refining approaches based on results

The process does not need to follow formal scientific method. It does need to be more than ad hoc guesswork — there should be a deliberate evaluation of alternatives.

4. Technological in nature

The process of experimentation must rely fundamentally on principles of:

• Engineering
• Computer science
• Biological science
• Physical science

Activities grounded in economics, business management, social sciences, arts, or humanities do not qualify, even if they involve uncertainty and experimentation.

Examples of qualifying activities

Software development

• Building new application features that require solving technical design challenges
• Developing algorithms to handle data at scale where performance outcomes are uncertain
• Integrating systems where compatibility and data flow are not predetermined
• Developing automated testing frameworks for complex software

Manufacturing

• Designing new tooling or fixtures to improve production efficiency
• Developing processes to work with new materials where behavior is uncertain
• Improving yield rates through process experimentation
• Automating production steps that involve engineering challenges

Engineering and architecture

• Developing structural solutions for unusual site conditions
• Designing mechanical or electrical systems to meet novel performance specifications
• Engineering custom fabrication processes
• Prototyping components to meet tolerances not achievable with standard methods

Food and beverage

• Formulating new products where shelf stability, texture, or taste outcomes are uncertain
• Developing processes to scale production while preserving product characteristics
• Adapting formulations to meet new regulatory or nutritional requirements
• Experimenting with alternative ingredients or preservation methods

Biotech and life sciences

• Developing new assays, compounds, or biological processes
• Designing experiments to test drug candidates or diagnostic methods
• Developing laboratory processes to improve reproducibility or throughput
• Engineering delivery mechanisms for therapeutic compounds

Agencies and professional services

• Developing proprietary software tools for client delivery
• Building data processing pipelines where integration challenges are uncertain
• Creating automated workflows that require solving technical design problems

Common misconceptions

"R&D means a lab coat." The credit covers any qualifying activity across any industry. A machine shop improving its CNC process qualifies the same way a biotech lab does.

"We just improved an existing product, so it does not count." Improvements qualify. The credit covers new or improved business components. If there was technological uncertainty in the improvement, it counts.

"We did not succeed, so it does not qualify." Failed experiments qualify. The credit rewards the process of experimentation, not the outcome. Activities that did not achieve the desired result still count if they met the four-part test.

"Our competitors already do this." The uncertainty must be uncertain to the taxpayer, not to the industry. If your team did not know how to achieve the result at the outset, it qualifies — regardless of whether others have solved the same problem.

What does not qualify

Not every technical activity qualifies. See Common Disqualifiers for the specific exclusions in the tax code, including funded research, post-commercial-production activities, adaptation, foreign research, and activities in the social sciences.

Next steps

• Check if your business qualifies with the free calculator
• Qualified Research Expenses — what expenses count toward the credit
• Documentation Requirements — what substantiation you need