"AI" is going to make getting software engineering jobs harder over the next year... but not because "AI" will replace humans, but rather because of the interaction of accounting and the need to pacify stock market investors.

1. Megacorps have spent massively on AI datacenters. This is a capital expenditure which means the expense isn't booked immediately.
2. Instead, the expense gets turned into deprecation over the next few years, i.e. reported expenses are going to be much higher for the next few years.
3. Higher expenses shrink reported profits, so these companies will start doing layoffs.
4. So far there is no evidence of offsetting profits.

Notice this effect happens _even if all AI datacenter spending stops today_. It's baked in as depreciation.

So far we've had layoffs at Amazon, rumored layoffs at Meta, and layoffs at Oracle (much more dire straits financially than the others). Today Microsoft is doing voluntary worker buyouts: https://www.bloomberg.com/news/articles/2026-04-23/microsoft-offers-voluntary-retirement-to-about-7-of-us-workers

I've seen rumors of Google doing buyouts too. And when these layoffs hit software engineers, that's going to mean a lot more competition for any open jobs across the whole industry. All those laid off engineers will want jobs too.

And these companies have also hired a lot of people in the past, and will be hiring less going forward, so this also means fewer jobs even if the rest of the industry continues as normal.

So that means more applicants for fewer jobs.

Update: Meta announces the rumored layoffs https://www.bloomberg.com/news/articles/2026-04-23/meta-tells-staff-it-will-cut-10-of-jobs-in-push-for-efficiency?srnd=homepage-americas

@itamarst AI is a massive misallocation of capital
#CapitalStrike
@itamarst I'm working on my leaf-blowing skills.

@itamarst
Does it really work like that, though? I mean yes, for companies like Nvidia, Google, MS etc who directly invest in building datacenters it is capital expenditures.

But how about everybody else? Those who actually use AI offerings? AI is billed based on usage and/or monthly so aren't we instead talking about operating costs (or whatever the right book keeping term is)?

@shadowdancer I'm only talking about the financial impacts on megacorps who are investing in datacenters. The thing is, they hire _many_ software engineers, so layoffs and decreased hiring there are very impactful in terms of availability and competition for jobs across the whole industry.
@itamarst
Okay, fair enough. They are indeed large employers of software engineers and can cause disruption in the job market by mass layoffs.

@itamarst @prietschka @davidgerard @edzitron.com

I mean the writing has been on the wall for years now that this is how it was going to go. All the major megacorps lighting more money than God on fire every quarter forever? There was no way it could last, and no way it could end other than hilariously terrible implosion. @davidgerard has been writing about it in his newsletter for years now, @prietschka and tons of other industry veterans have been screaming about it to anyone who would listen for years now, @edzitron.com has been writing his massive newsletters about it for years now…

That it's taken this long for people at these companies to start feeling the punch is truly a testament to how much money they can burn before they feel anything. What comes next will NOT be pretty…

@itamarst

Hub's company has fantastic growth, record earnings and the market goes meh. I'm not sure why companies even bother trying to jack their stock market price anymore, it's just stupid really. I know, compensation in stock blah blah blah but if the market isn't even going to be rational what the fuck is the point?