The real drivers of tech layoffs
- YT :: https://www.youtube.com/watch?v=tt_SloZ18Mk
- Original title :: AI Tech Layoffs Explained by ex-Meta Principal Engineer
A former L8 engineer at Meta, Microsoft, and Atlassian uses layoff.fyi data to trace tech layoffs back to three overlapping causes: pandemic over-hiring, rising interest rates, and, most recently, AI-driven reallocation of capital and headcount, and argues the AI-driven wave will keep growing.
Pandemic over-hiring and its correction
COVID-19 lockdowns had two opposite effects on demand: travel-dependent companies (Uber, OYO, Groupon, Airbnb) shrank, while online-activity companies grew sharply as people stayed home. Many of the latter assumed the growth was permanent and hired aggressively to match it — Shopify's 2022 layoff announcement explicitly cited its bet that e-commerce would "permanently leap ahead by five or even ten years." When growth reverted to pre-COVID trend lines, the resulting overstaffing became the first big driver of layoffs.
Interest rates and the pressure for efficiency
Federal Reserve data shows interest rates near zero through 2020–2021, then rising sharply from mid-2022 onward. Near-zero rates make borrowing cheap, so companies can invest based on long-term outlook without urgency. Once rates rise, servicing debt costs real money every year, pushing companies to prioritize near-term profit and margin over long-term bets. Overlaying interest-rate data on the layoff timeline shows a close correlation: layoffs stayed low while rates were near zero and surged once rates began climbing, coinciding with widespread corporate talk of "efficiency."
A separate hardware downturn
2024 layoffs at companies like Intel, Dell, and Cisco trace to a third, distinct cause: a hardware demand collapse. The pandemic drove a one-time surge in device purchases (home-office PCs, remote-work equipment); once that replacement cycle finished, hardware demand fell off, squeezing hardware vendors independent of the interest-rate or AI stories.
AI as a rising, distinct category of layoffs
Starting in 2024, "AI" emerges as a distinct and growing category of layoff justification in the layoff.fyi data. Before then, AI was often invoked as a rhetorical efficiency claim rather than a real cause. The turn became concrete with Meta's AI push: a roughly $14B deal for part of Scale AI (bringing in CEO Alexander Wang to lead Meta's AI org), aggressive AI-researcher hiring from frontier labs (OpenAI accused Meta of offering nine-figure sign-on bonuses), and a plan to spend up to $65B on AI, much of it on data centers and compute.
To fund this at a time of high interest rates, Meta reallocated rather than simply adding headcount: reassigning roughly 7,000 people who could contribute to AI work, and laying off people whose skills didn't map to the new priority. The presenter frames this explicitly as AI becoming the dominant strategic priority — not AI directly performing people's jobs — as the mechanism behind these cuts.
Where AI is actually replacing engineering work
A separate, more literal case of AI displacing labor: a Reddit account from an Amazon employee described AWS Bedrock's routing and load-balancing layer being rewritten (not merely refactored) by a distinguished engineer and eight principal engineers using AI in about a month — work previously maintained by an organization of over a thousand engineers. This illustrates that small AI-augmented teams can now move faster than large ones.
That shift is reshaping org design. The traditional unit — an engineering manager coordinating a team of ICs, with managers stacking under directors as complexity grows — is being questioned in favor of a new unit: one or a few ICs paired with a large amount of AI agent capacity, which need no recruiting or performance reviews. As directors ask whether this new unit still needs as many engineering managers, middle management is taking a disproportionate hit in recent layoffs. The presenter expects AI-attributed layoffs to keep rising over the next few years.
What to do about it
The core advice: take working with AI seriously now, since career trajectories increasingly depend on it, drawing an analogy to prior technology transitions (factories replacing handcraft, tractors replacing manual farming, spreadsheets replacing manual accounting). Two viable paths are proposed: go deep into the AI stack itself (ML, training, inference, or chip hardware), or become the most AI-native person on your existing team by using AI to work faster and better. Early attempts at using AI will likely be frustrating and error-prone, consistent with the learning curve of any new technology, but the productivity gains are described as real and already realized by those who've pushed through it. For those already laid off, the advice is to treat it as an opportunity to reassess career direction and explore new interests — the presenter notes he voluntarily left his own L8 role about a month before this video and does not regret it.