R&D stands for research and development, the broad category of work companies and governments do to create new products, services, or processes and improve existing ones. Global R&D spending now exceeds $2 trillion annually, with the United States and China each investing roughly $1 trillion. It touches nearly every industry, from pharmaceuticals developing new drugs to tech companies building software to automakers designing electric vehicles.
The Three Types of R&D
The internationally recognized framework for defining R&D, called the Frascati Manual, breaks it into three categories:
- Basic research is work done to gain new knowledge without any specific application in mind. A university lab studying how a protein folds is doing basic research. There’s no product on the horizon; the goal is understanding.
- Applied research takes that knowledge and directs it toward a specific practical goal. If that same protein-folding insight gets used to identify a drug target, the work has shifted to applied research.
- Experimental development uses findings from research to actually build or improve a product, process, or service. This is where prototypes get made, software gets coded, and formulations get tested at scale.
Most corporate R&D falls into the experimental development category. Basic research is more commonly funded by governments and universities, though some large companies (notably in pharmaceuticals and semiconductors) maintain significant basic research programs.
Which Industries Spend the Most
In the U.S., five industries account for 79% of the $602.5 billion businesses spent on R&D in 2021. Software publishing and information technology led at 25%, followed by chemicals manufacturing (which includes pharmaceuticals) at 18%, and computer and electronic products (including semiconductors) at 17%. Professional and technical services contributed 11%, and transportation equipment manufacturing, covering cars and aerospace, made up 8%.
What matters more than raw spending is R&D intensity: how much a company spends on R&D relative to its revenue. Semiconductor manufacturing is one of the most R&D-intensive industries in the U.S., with companies spending about 20 cents of every sales dollar on research and development. Pharmaceutical companies typically fall in a similar range. A retail company, by contrast, might spend less than 1% of revenue on R&D.
How R&D Works in Practice
Most companies don’t just throw money at research and hope for results. They use structured processes to move ideas from concept to market. One widely adopted framework is the Stage-Gate model, which breaks product development into five stages, each followed by a decision point where leadership decides whether to keep investing or kill the project.
The stages move from an initial concept assessment (quick desk research to see if an idea has legs) through building a full business case with market research and financial projections, then into actual development where engineers build prototypes. After that comes testing and validation, including beta tests and field trials, and finally a full commercial launch. Each gate forces teams to prove the project is still worth pursuing before more money gets committed. This prevents companies from pouring resources into ideas that should have been abandoned early.
In pharmaceuticals, the process is even more rigorous and dramatically longer. A new drug typically moves through preclinical work, then three phases of clinical trials. Phase 1 takes several months and tests safety in small groups; about 70% of drugs advance. Phase 2 runs several months to two years and tests whether the drug actually works; only 33% make it through. Phase 3 involves 300 to 3,000 patients and takes one to four years; just 25 to 30% of drugs pass. From initial discovery to approval, the entire process often takes over a decade.
Measuring Technology Readiness
NASA developed a nine-level scale called Technology Readiness Levels (TRLs) that has since been adopted across defense, energy, and manufacturing industries. It provides a common language for describing how far along a technology is. At TRL 1, you’ve observed basic scientific principles but nothing more. By TRL 3, you have proof of concept in a lab. TRL 5 means a component has been tested in a realistic environment. TRL 9 means the technology is fully proven in real-world operation.
This framework helps organizations make funding decisions and communicate progress without ambiguity. When an engineer says a technology is at TRL 4, everyone understands it means a basic prototype works in the lab but hasn’t been tested under real conditions yet.
How R&D Costs Are Handled Financially
Under U.S. accounting rules, companies generally must treat R&D spending as an expense in the year it occurs rather than spreading it over time as an investment. This means R&D-heavy companies can look less profitable on paper than they actually are, since their spending on future products hits the income statement immediately. There’s one notable exception: software development costs can be capitalized (treated as an asset) once the software has reached “technological feasibility,” meaning all the planning, designing, coding, and testing needed to confirm it can be built to spec has been completed.
To encourage R&D investment, the U.S. offers a research tax credit under Section 41 of the tax code. To qualify, a project must pass all four parts of a test: the spending must be eligible as a research expense, the work must aim to discover information that is technological in nature, the result must be intended for a new or improved business product or process, and substantially all of the work must involve a process of experimentation. This credit applies across industries, not just to traditional lab work. A manufacturer improving a production process or a software company developing a new algorithm can both potentially qualify.
Global R&D Spending Trends
The geography of R&D is shifting fast. Projections for 2025 put China at roughly $1.05 trillion in R&D spending (adjusted for purchasing power), nearly matching the U.S. at $1.07 trillion. A full crossover is expected by 2026. Asia as a whole is projected to capture 42% of global R&D spending in 2025, up from 34% just five years earlier. This shift reflects massive government investment in China, South Korea, and other Asian economies, particularly in semiconductors, artificial intelligence, and clean energy.
For businesses, this means the competitive landscape for talent and innovation is increasingly global. Companies that once dominated R&D from a handful of labs in the U.S. or Europe now face rivals with comparable resources and growing technical expertise across Asia.