Generally the US requires valuable assets to be depreciated and amortized over their useful life. This is arguably a fair way to tax businesses with fewer downsides than many alternatives.
Consider a different situation, a business pays employees to build a residential home for $275k total, the land is worth zero in this simple example. Currently they can deduct $10k a year for 27.5 years to depreciate the home, even though they paid $275k up front. Allowing the business to deduct the entire $275k at once, only recovering the difference when the depreciated asset is sold is basically a tax free loan at the expense of all other taxpayers.
To be fair there are many situations where the government wants to incentivize spending in certain areas. Certain types of businesses can avoid depreciation and deduct full expenses, like for farm equipment and heavy duty vehicles, previously most R&D. Or where accelerated or bonus depreciation is used because most of the income is in the first few years. Like a taxi follows a 5 year double depreciation schedule, in the first year a $25k taxi would depreciate $10k, then $6k the next year, there are many examples that are on a shorter schedule.
Keeping a 5 year straight line depreciation on R&D benefits large established businesses and burdens startups, this is primarily a political decision and not economic. Another issue is that not all R&D spending results in a valuable asset with a usable life
I think you are confusing the use case. If a business pays employees to build a residential home for $275k and the land cost $0, they don’t deduct anything. Assume they sell the home for $350k. They then have a cost of sale of $275k so they make a profit of $35k0 - 275k = $75k. They then pay tax on the $75 in the year the home was sold.
Section 174 relates to Research and Development expenses. The idea originally was they firms were spending a lot of money to develop a product that might have long useful life, but the expenses were all hitting the operating expense in the current year. This is great for cash tax purposes, but bad for metrics like operating income and net income which impact many firms valuations. Arguably it also misrepresents the firm’s business model. A fundamental idea is that costs should be reflected consistently with the period that revenue is earned. With R&D, the revenue is expected to be earned over many years but the costs are incurred upfront.
The better example is not a residential home, but an apartment building. The builder spends $300M to build the apartments then leases them for the next 30 years. In that example, the $300M is depreciated over 30 years so the costs track with the expenses.
Yes this is the same type of example. A builder that spends $300m to build apartments doesn't deduct $300m in the first year. The same way a real estate investor that builds a single family home to rent out doesn't deduct $275k the first year
I believe that this is and similar example are missing a very important point in the narrative: in case of developing land to build a building, you will still have the possibility to deduct 100% of the salaries of the construction workers.
Those construction workers will build the building (= the product) on top of the land that you acquired (another asset) which means you are assets become the land itself and the building. Land and building will have their own asset value (purchase price or evaluation) that will be used for over X years.
As far as I understand as an European watching this from abroad, they are trying to evaluate the value of the asset (= code, the product) of a startup by using development costs as a proxy.
Employee salary is part of the capitalized cost and depreciated.
Land value is generally not a deductible expense at all, and as such it is not depreciated. Some related costs like property tax, insurance, mortgage interest are deductible when paid though