If your Inbox is anything like mine, you are being absolutely deluged with notifications of articles. Yesterday one such article I did find interesting was in Quartz and was written by Chris Morris. The article describes the monthly layoffs in January 2026:
Employers laid off 108,435 people last month, the highest January number since 2009, according to a new report from Challenger, Gray & Christmas. At the same time, hiring intentions haven’t been lower since then.
The number reflects a 118% increase over January 2025 and more than doubles the total layoffs of last December. New hires came in at just 5,306, the lowest number since Challenger began tracking that data point in 2009.
What struck me about the article was how little actual insight it offered. I’m going to attempt to correct that.
I’m not sure whether it will make you feel better or worse but just about half of the announced layoffs were in just two companies, Amazon and United Parcel Services (UPS), and UPS’s layoffs can be attributed to the loss of its Amazon contracts.
That naturally leads to the follow-up question: what’s going on with Amazon? The short answer is AI just not in the way most people think of it.
The longer answer is that Amazon is in a fight for survival of its current margin structure and strategic dominance. Contrary to what you may believe, Amazon isn’t just a big retailer. For the first ten years of its corporate history it was an IPO in search of a business model. In 2006 it finally happened on one and it wasn’t as an online retail bookseller (its origins). That was when Amazon Web Services (AWS) was born. Since then it’s been clear: Amazon is a web services provider that dabbles a bit in online retail sales. Sometimes it makes money in online retail sales, sometimes it loses money but it always makes money providing web services.
AI summaries in web searches are already cutting “click through rates” by more than 50%. Fewer click throughs changes the economics of web sites. That in turn will inevitably result in fewer web sites. A similar thing is happening to online APIs. There are fewer discrete service requests, fewer API calls, and lower marginal revenue per workload.
AWS’s main business is providing online infrastructure for web sites and online API service providers. Couple that with an enormous expansion in capex for training and inference and AWS clearly has a major problem.
What about LLM AI service providers? Don’t they use online infrastructure?
Ignore the names you might have heard of in LLM AI, ChatGPT (OpenAI) and Claude (Anthropic). There are only four actual prospective leading suppliers of LLM AI services:
Google
Microsoft
IBM (via enterprise integration rather than consumer scale)
Amazon
By “leading supplier” I mean vertically integrated, capital-scale LLM providers who can survive margin compression, not just model builders. Everyone else is either a tenant, a feature, or a future acquisition target.
Amazon’s recent actions (internalization of logistics, contract shedding, headcount discipline, and ruthless focus on the margin) are best explained as preemptive capital reallocation to defend its position in an online infrastructure market that is increasingly AI-driven. It doesn’t need AWS to grow; it needs AWS not to shrink while its AI capex explodes.
The other three candidates are much less exposed than Amazon, each for different reasons.
These layoffs are not cyclical weakness. They are the visible surface of a capital war.
