Alex Domash finds AI added one million net US jobs

The mass job displacement narrative just broke. AI technologies generated nearly one million net new jobs across the United States, completely overwhelming corporate layoffs.

Economist Alex Domash calculated that while companies announced roughly 200,000 AI-related layoffs over recent years, net employment exploded elsewhere. The boom split into two distinct tiers: roughly 350,000 blue-collar infrastructure roles for electricians, utility crews, and HVAC technicians building data centers, alongside 750,000 white-collar software engineering and data science positions. Even heavily exposed fields like paralegals and financial analysts saw employment grow as cheap intelligence expanded market demand.

Data from Anthropic aligns with the broader labor resilience. Head of economics Peter McCrory analyzed 18 months of user data alongside Bureau of Labor Statistics figures, showing national unemployment holding at 4.2 percent despite quality-adjusted AI output surging over 2,000 percent annually. McCrory observed that workers use artificial intelligence to augment output and expand scope. However, investor Trace Cohen argued that corporate hiring freezes quietly mask long-term harm to junior workers.

The disruption is real, but it is taking the form of physical construction rather than simple labor destruction.

Building and running this infrastructure requires massive capital. On The AI Daily Brief, discussion focused on how enterprise buyers are enforcing strict token budgets to control inference costs. Microsoft cut operational costs by 84 percent by shifting default enterprise workloads to internal small models, while Stripe entered talks to acquire model router OpenRouter for $10 billion. Running specialized small models on existing hardware like NVIDIA H100s extends chip lifespans and de-risks data center capital expenditures.

The scale of this buildout could redefine federal debt dynamics. On What Bitcoin Did, fund manager Jeff Ross argued that Washington will treat data center infrastructure as a national security imperative. Ross projects that heavy capital spending on power grids and compute infrastructure will drive a manufacturing boom. Rapid productivity gains generate price deflation across healthcare and energy, allowing real economic growth to outrun government debt.

The predicted labor apocalypse remains theoretical. For now, the physical world is demanding millions of workers to build the engine.