Capitalized Software and R&D
Most R&D spending is expensed immediately, but software development crosses a bright line — once a project is judged technologically feasible, its costs move from the income statement onto the balance sheet.
Prerequisites: Capitalizing vs Expensing a Cost, Reading an Income Statement
A company spends $10 million building a new product this year. Should that $10 million reduce this year's profit dollar-for-dollar, or should it sit on the balance sheet as an asset and get expensed gradually over the years the product will actually generate revenue? US GAAP's default answer for research and development is blunt: expense it immediately, because the payoff is too uncertain to call it an asset yet. Software development is the major exception, and knowing where the line falls changes reported profit substantially for any tech-heavy company.
Under US GAAP, R&D is expensed as incurred — full stop, no capitalizing "promising" research. But once a software project passes technological feasibility (for products to be sold) or reaches the application development stage (for software built for internal use), further costs to build it can be capitalized as an asset and amortized over its useful life instead of hitting profit all at once.
Where the line falls
For software sold to customers, technological feasibility is typically reached when a detailed program design or working model exists — proof the product can actually be built as planned, not just an idea. Costs before that point (research, planning, early prototyping) are R&D expense. Costs after that point (coding, testing to prepare for release) are capitalized.
For software a company builds for its own internal use — an ERP system, an internal trading platform — the trigger is different: costs during the preliminary planning stage are expensed, but once management commits to funding the project and it's probable it will be completed and used, costs in the application development stage (actual coding, testing, installation) are capitalized.
Worked example
A company spends $4 million on early-stage research proving a new algorithm works, then $6 million building and testing the shippable product after feasibility is established. Useful life of the resulting software is estimated at 3 years, amortized straight-line.
- Year 1 expense: the $4 million research cost hits the income statement immediately as R&D expense.
- Capitalized asset: the $6 million build cost goes on the balance sheet as a software asset, not expensed at all in year 1.
- Amortization begins once the product is released: , i.e. $2,000,000 per year for three years.
Total year-1 income statement hit: $4 million R&D expense plus, once released, a partial year of amortization — far less than the full $10 million spent, with the rest smoothed into future years as the product actually earns revenue.
What this means in practice
Capitalizing software costs boosts near-term reported profit and shows up as a growth in intangible assets rather than an expense, which is why analysts often add back amortization of capitalized software when comparing companies with different capitalization policies, or check the cash flow statement (where the cash outflow appears in full regardless of accounting treatment) as the real spending number.
A company that aggressively deems more of its development "past feasibility" than peers do can report higher margins purely through accounting judgment, not superior efficiency. Compare capitalized software as a percentage of total R&D-plus-capitalized-development spend across companies before trusting a margin comparison.
Related concepts
Practice in interviews
Further reading
- FASB ASC 985-20, Costs of Software to Be Sold, Leased or Marketed
- FASB ASC 350-40, Internal-Use Software