This analysis was first published in SvD Näringsliv, in Swedish, on February 16th, 2024. This piece was translated from Swedish by Claude. Some phrasing may differ from a human translation.

Tech giants are laying off thousands of employees even as they invest massively in AI. The irony is that the very tools they are building may be what makes those cuts permanent.

In January 2023, Google announced it would lay off approximately 12,000 employees — around six percent of its global workforce. The cuts came despite the fact that the company had grown its headcount by roughly 55,000 people since the start of the pandemic.

The pattern was not unique to Google. Microsoft, Meta, Amazon, and others made similar announcements in the same period. The narrative was straightforward: the pandemic had driven an unusual surge in demand for digital services, the companies had hired aggressively to meet it, and now the correction had arrived. The boom was over.

But something else is happening alongside this correction — something with longer-term consequences. These same companies are simultaneously making enormous investments in artificial intelligence. The capital expenditure required for AI — chips, data centres, infrastructure — is staggering. And it is running in parallel with a period in which the core growth of these businesses has slowed to single digits.

The term “dogfooding” comes from the tech industry. It means using your own product internally before releasing it to the world. The idea is to catch problems early, but also to build genuine conviction in what you are making. If you would not use it yourself, why would anyone else?

The tech giants are now dogfooding AI in a very direct sense: they are deploying it to handle tasks that engineers and analysts previously performed. Code review. Documentation. Internal tooling. Routine data analysis. These are precisely the kinds of tasks that junior and mid-level engineers spend much of their time on.

An average software engineer in Silicon Valley costs around 1.5 million kronor per year in total compensation. At that price, the incentive to replace even a fraction of that work with AI tools is significant — especially when the same companies are under pressure to demonstrate that their massive AI investments will eventually translate into efficiency gains.

This is the irony at the heart of the current moment. The companies that built the AI are among the first to face the consequences of deploying it. The layoffs of 2023 were framed as a pandemic correction. But the structural pressure that follows — from AI automating the kinds of cognitive work that tech companies pay most handsomely for — may mean that the giants never return to their previous headcounts.

We may, in other words, be at peak tech worker. The largest technology companies in the world may never be larger, in terms of employees, than they are right now.

That is a remarkable thought. These companies have long been among the most sought-after employers in the world. They have shaped the labour market, driven wage inflation across the knowledge economy, and served as benchmarks for how high-skilled work could be compensated. If they begin to shrink — not through failure, but through the success of their own technology — the ripple effects will extend far beyond Silicon Valley.

It would be the ultimate irony if those who built this technology were also among the first to lose their jobs to it.