Comment by w4

I think that is broadly right in theory, yes, though the IRS isn’t saying that currently. They haven’t provided any guidance at all last I heard, in spite of being asked for it by industry groups many months ago.

Setting aside the apparently absurd state of things as they stand today, I agree in principle that coding that doesn’t result in a new asset shouldn’t be treated as R&D and thus should be deductible even under this regime. However, the issue that we’ve been grappling with is that even if we end up there, it’s still not totally clear where the line is with regards to software, as software is markedly different from other asset classes (e.g. machinery or equipment). Code written to support an existing revenue generating piece of software may still be R&D. For example, is adding a new feature to software R&D that creates a new asset (the new feature and new software version), or is it maintenance of an existing asset (the existing software)? It’s a mess. Under the previous regime this wasn’t an issue; you had an option to treat software development either way, so no one was worried about these nuances because it was a choice, meaning there was no way to get it wrong. Not so anymore, and the IRS is not helping to understand it right now based on what I’ve been told.

Then add in the administrative burden presented by this change: Software devs are likely to get very familiar with 0.1 time tracking that lawyers and other professionals suffer with if this change sticks, so that businesses can figure out what is deductible and what must be amortized.

Another issue is that if you dispose of or retire the asset there doesn’t seem to be any provision to accelerate or write off the depreciation immediately. This starts to look like a particularly pernicious challenge in the case of startups, for whom cash flow is an issue and who may very well fail before the 5 year amortization period is up. It’s also a major potential issue in the case of re-writes.

The apparent inability to write off failed R&D is also a devastating change in my opinion, and its impacts extend well beyond software. It greatly increases the risk of R&D at a time when we need it most to confront challenges like climate change, to remain competitive with China, to capitalize on opportunities in AI, and so forth.

So while I agree with you in principle, even if that’s where we end up - and it makes sense that it’s a reasonable outcome if deductibility is not restored - it’s still a real mess compared to where we were previously, and will have the effect of stifling innovation and economic growth.

EDIT: Cleaned up a bit. A lot of folks do contradict my point: They’re saying that you can still deduct software development R&D citing old info, or that you can deduct it if you don’t take the credit. That’s just wrong.