In October 2024, I wrote that the tech industry was entering an era of silent firing. Jobs were not being eliminated overnight, but subtly reshaped in ways that encouraged attrition, as companies quietly prepared for large-scale automation. At the time, this was largely a warning. With age, it looks more like a pattern.

Amazon’s January 2026 announcement of 16,000 layoffs brings corporate staff reductions to roughly 10% of its workforce. Publicly, leadership has been careful to separate these cuts from artificial intelligence. As CEO Andy Jassy put it after earlier reductions, “the announcement that we made…was not really financially driven, and it’s not even really AI-driven, not right now at least.” Yet in parallel, Jassy has been explicit about the role AI is already playing across the company, stating that “in virtually every corner of the company, we’re using generative AI to make customers’ lives better and easier,” and that new AI-driven agents are “coming, and coming fast.” The tension between these two statements highlights a growing accountability gap: AI is framed as transformative when speaking to investors, but incidental when explaining workforce reductions.

A NEW MODEL IS NEEDED AFTER AI INVESTMENTS

Let’s look at the case of Meta (Facebook). They will either have to trim other expenses significantly or start charging users. Meta’s current global average annual revenue per user sits at roughly $13–14. On the higher earning side are the U.S. and Canada, at around $68 per user. Seeing as Meta’s fastest user growth markets are in Asia-Pacific and developing countries, that $13-14 figure is more likely to go down rather than up.

Consider Meta’s commitment to spend $600 billion on infrastructure by 2028. Divide this by 3 billion Facebook users, and they would need to squeeze an additional $200 per year from each user just to break even on this investment alone. Said differently, they would have to ~15x their annual revenue per user to match this spend. This does not factor in their aggressive acquisitions in the AI space, which adds to payroll and layoff costs post-acquisition. The math is simple: Meta will need to find new revenue sources from a user base already inundated with ads. Or more likely, the most obvious target is to cut expenses, with headcount made redundant by their own AI investments.

We are seeing evidence of this accelerating towards a head: Meta is planning to track employees’ clicks and keystrokes to train AI. This seems like the exact step you would take before fully automating jobs and firing workers.

S&P RISES WHEN JOB POSTINGS FALL

What is notable is how early this shift appeared in the data. Since the launch of ChatGPT in October 2022, job postings have decreased by one-third, while the S&P has risen by 75%. This presents an eerie reversal of historical norms, where periods of market expansion have traditionally been driven by hiring, not job cuts. We still have not yet reached widespread AI deployment, yet hiring behavior has already adjusted!

This suggests we are still in the first phase. Companies are cutting staff while simultaneously insisting that AI is not the cause. That framing may hold for now, but it becomes harder to maintain as automation capabilities expand and productivity gains become more visible.

Read the full article here.