More than 280,000 technology positions are currently open. The strongest demand is moving toward the physical infrastructure required to make AI work.
Technology companies are still laying people off, but they are also hiring at a scale that complicates the idea that artificial intelligence is simply eliminating tech employment. TrueUp, which tracks more than 9,000 technology companies, is currently monitoring more than 280,000 open tech positions, up sharply from the beginning of 2026. At the same time, approximately 190,000 tech workers have been laid off this year. That is a significant number, but the industry is on pace to finish well below the roughly 430,000 layoffs recorded during the 2023 peak. The employment story inside technology is becoming less about whether jobs exist and more about which jobs companies now need.
AI Needs a Physical Economy
The clearest increase is appearing in hardware engineering. Companies building AI models also need GPUs, servers, data centers, robotics systems, electrical infrastructure, cooling systems, manufacturing capacity and the engineers who can design and maintain all of it. TrueUp founder Amit Taylor pointed to GPU development, AI robotics, U.S. manufacturing startups and companies such as SpaceX as drivers of the hiring demand.
That creates a useful correction to the way the AI labor story is often framed. Artificial intelligence may reduce demand for some software, support and administrative roles while simultaneously increasing demand for workers closer to the physical systems that make AI possible. An AI model may live in software, but the infrastructure underneath it is industrial. Someone has to build the chips, wire the data centers, maintain the networks, design the cooling systems and manufacture the equipment.
The Labor Market Is Rotating
That does not mean displaced software workers can simply walk into the new jobs. A front-end developer, customer-support specialist or product manager cannot necessarily move directly into semiconductor engineering or data-center construction. The skills, credentials and geography are different.
That is where the transition becomes a labor-market access story. Companies can truthfully say they are hiring while workers in adjacent occupations still experience the market as closed. The aggregate number of openings does not tell you whether the person who lost a job has a plausible route into the jobs being created.
We have seen the same pattern outside technology. Companies restructure one function while expanding another, which produces headlines that sound contradictory: layoffs and hiring at the same company, sometimes in the same quarter. The contradiction disappears once the organization’s priorities are visible. Employers are not necessarily becoming smaller. They are reallocating payroll.
AI Is Becoming an Industrial Policy Story
The shift toward hardware also intersects with U.S. efforts to bring semiconductor manufacturing and other strategic production back home. Demand for advanced chips, power-generation capacity and data-center construction means AI investment increasingly depends on industries that cannot be scaled by adding software engineers alone.
That changes who benefits from the boom. Regions with manufacturing capacity, energy infrastructure, engineering talent and available land may capture more of the next phase than traditional software hubs. Universities and workforce programs will face pressure to train people for jobs that sit between technology, manufacturing and engineering.
The next useful measure of AI’s employment impact will therefore not be a single national job-loss number. It will be whether workers displaced from one part of the technology economy can reach the occupations growing in another.
AI is not removing labor from technology uniformly. It is redrawing the map of where that labor sits.