Anthropic tripled its headcount over the past year. The share of junior engineers on its technical teams moved in the opposite direction.
According to an internal workforce report released on October 16, 2026, junior employees fell from 21% to 14% of the company’s technical staff since early 2025.
The numbers behind the shift
The most striking figure concerns new hires rather than existing staff. Junior technical roles made up 27% of Anthropic’s technical hiring before the shift. That figure has dropped to 9%.
The decline was not gradual. The report identifies a sharp drop in junior hiring that began after December 2025, with job postings for entry-level roles falling noticeably from that point onward.
Technical work as a whole also shrank relative to the rest of the company. Technical roles accounted for 61% of the total workforce before the shift and 56% afterward.
Technical hiring kept growing, just more slowly than hiring for non-technical roles. Business functions expanded faster, and the company says non-technical employees are increasingly using coding tools themselves.
The agents are clocking more hours than the humans
By August 2026, employee-directed Claude agents were logging six times the active work hours of Anthropic’s human staff.
Agents first overtook human work hours in March 2026. Five months later, they were working the equivalent of six shifts for every one a person put in.
Code commits per engineer rose by approximately 5x compared with six months earlier.
Meeting time per employee rose 20% year-over-year, and 16% for technical staff specifically.
Managers are hiring less because of AI
The report also draws on an April 2026 survey of Anthropic managers. About one-third of machine learning managers said they had reduced hiring because of AI use.
Among software engineering managers, the figure was about one-fifth. Junior roles were the most affected in both groups.
The research was produced by Anthropic economists, including Nathan Wilmers. It combines HR records, agent performance data, and employee feedback.
Why this report stands out
Anthropic is effectively the earliest adopter of its own product. Its staff had access to advanced Claude agents before most enterprises did, which makes its workforce a rough preview of what heavy AI adoption might look like elsewhere.
The changes described here unfolded over roughly a year, from early 2025 to late 2026. Usually, a company that triples in size hires a wave of junior people to handle the expanding workload. Anthropic grew without doing that, and the report points to AI agents as a big part of the reason.
What this means for the tech labor market
The most immediate concern is the entry-level pipeline. If junior technical hiring falls from 27% to 9% at a fast-growing AI company, graduates and early-career engineers may face a tougher market if other firms follow a similar path.
The rise in meeting time is worth watching as well. Reviewing, coordinating, and quality-checking machine output is real work, and organizations that ignore that cost may find their efficiency gains smaller than the commit counts suggest.
The caveat is that Anthropic is one company, and an unusual one. Its employees build AI for a living and are likely more comfortable delegating to agents than the average workforce.
Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.