---
title: "Anthropic staff expect AI to do 75% of their work by 2030, and headcount to grow 66%"
description: "An internal poll of 1,334 employees finds them far more bullish than US workers on job losses, growth and how fast AI improves"
author: "Nate Ledger"
published: 2026-10-16T22:48:53Z
modified: 2026-10-17T07:35:46Z
url: https://rews.cc/a/anthropic-staff-expect-ai-to-do-75-of-their-work-by-2030-and-d29710
language: en
tags: ["ai", "jobs", "economy", "anthropic", "automation", "tech", "business"]
publisher: "Rews (https://rews.cc)"
---

# Anthropic staff expect AI to do 75% of their work by 2030, and headcount to grow 66%

*An internal poll of 1,334 employees finds them far more bullish than US workers on job losses, growth and how fast AI improves*

By Nate Ledger · October 16, 2026 · https://rews.cc/a/anthropic-staff-expect-ai-to-do-75-of-their-work-by-2030-and-d29710

## In brief

- Anthropic published “The AI expectations gap,” comparing an internal poll of 1,334 staff with surveys of US adults
- Over half of staff expect AI to do at least 80% of their work by 2030, versus 5% of US workers
- Median staff answers imply 17.1% US GDP growth in 2030, versus 5.4% for the median US adult
- 95% of staff expect Anthropic’s headcount to grow within 12 months, with a median expectation of 66%
- Chief economist Peter McCrory said Anthropic runs about 18 months ahead of the economy on AI adoption

Anthropic asked its own employees what they think AI will do to work and the economy. It asked a large sample of ordinary Americans the same questions, then published the comparison on Friday under the title “The AI expectations gap.” As [Stephen Council reported for Business Insider](https://www.businessinsider.com/anthropic-survey-reveals-ai-potential-impact-on-unemployment-2026-10), the gap is wide. Most of the 1,334 Anthropic staffers polled said AI will be able to handle at least 80% of their jobs by 2030. About 5% of US workers said the same about theirs.

The [report itself](https://www.anthropic.com/research/ai-expectations-gap) fills in the details. The public side was a Morning Consult survey of 10,980 US adults, fielded August 11 to 23 and weighted to be nationally representative. The staff side was an internal poll run August 9 to 12, and it was not weighted. On average, staff think AI could do 75% of their current work by 2030, against 28% for US workers. Ninety-one percent of staff expect AI to make a Nobel-level discovery by 2030, compared with 39% of US adults. Among people who use AI at work, 59% of Anthropic staff say a task takes a quarter of the time or less with it. For US adults the figure is 26%.

The researchers’ summary is careful about where the disagreement actually sits:

> US adults and Anthropic staff largely agree about what AI can do today. However, they disagree about how quickly it will improve, how widely it will be used, and how large the productivity gains measured by time saved will be.

So the two groups agree on the level and disagree on the slope. That is the normal shape of an argument about technology, and also the normal shape of an argument about a stock.

![Two businesspeople follow diverging cream paths against a teal background, one climbing a steep curve carrying a coral microchip, the other walking with a briefcase.](https://rews.cc/img/f10f5d8745ac9d33747607094cb6b6017bb1352f.webp)

## The macro part

The researchers turned survey answers into implied forecasts for GDP growth in 2030. For the median US adult the number was 5.4%, which is already a lot. For the median Anthropic employee it was 17.1%. US GDP grew 2.2% last year, so the staff are forecasting a growth rate nearly eight times that. The report spells out what it would mean: under staff expectations, “GDP is doubling in size roughly every four and a half years.” At 2.2% a year, doubling takes about 32 years.

Then there is unemployment, which is now 4.2%. According to Business Insider, most Anthropic employees expect it to top 12% in 2030, while most US workers think it will sit around or below 8%. The report frames this in a more interesting way. US workers who expect AI to do less than half their work forecast unemployment of 7% to 8%. Those who expect it to do 80% or more forecast 12.2%, which the report notes is “above its peak during the Great Financial Crisis.” Each extra 10 points of a worker’s job that AI takes over goes with about 0.4 points of extra forecast unemployment, and the report says staff follow the same pattern.

I think that is the most interesting finding in the paper. The public and the AI lab share a mental model of the economy: if AI can do my job, it can probably do lots of jobs, so unemployment goes up. They are not arguing about economics. They are arguing about one input to the model, which is how good the software will get. And one side builds the software.

## The hiring part

The median staffer thinks AI does 51% of the work of building a frontier model today, 80% by the end of 2027 and 95% by the end of 2030. Those figures, the report says, “are not a measurement by Anthropic or a company forecast.”

![Two cream-colored industrial robotic arms with orange grippers stand against a blue background, with a person beside the smaller arm on a table.](https://rews.cc/img/aa910b0def0a048e8345ce80fee42a4f8d50667a.webp)

Here is my favourite number. According to the report, “95% of Anthropic staff expect headcount to grow over the next 12 months, the median staff member expects it to grow by 66%,” and 27% expect it at least to double. The company adds that these “are the personal expectations of staff members in an internal poll and do not represent Anthropic’s plans.”

So the median employee expects AI to do 95% of the work of building AI by 2030, and also expects the company that builds AI to have two-thirds more people by next summer. These views fit together better than they first appear to. If you believe your product is about to become the most valuable thing in the economy, you hire as fast as you can now, because you are selling the thing that does the replacing. The jobs disappear somewhere else first.

Peter McCrory, Anthropic’s chief economist, made the case on X for why outsiders should care: “We’re publishing this research because Anthropic runs roughly a year and a half ahead of the economy on AI adoption, so what happens at Anthropic could be a leading indicator for the broader world.”

Maybe. The report and Business Insider offer some reasons the staff might simply know more. They have bigger budgets for AI tools and often get to use models that haven’t been released. They also do a kind of work that is unusually exposed to AI. The average US worker is more likely to work with their hands or in a service job, and respondents in manual trades, production and transport expected AI to handle the least of their work by 2030. Tech workers and people in business and finance expected more.

![Two workers, one holding a wrench, stand beside an oversized cream microchip with an orange side panel against a teal background.](https://rews.cc/img/4193f6076474b7a635fc30192f3d646a33404ad9.webp)

There is also a selection effect. People who take jobs at an AI lab tend to believe in AI, and they work for a CEO who has said this sort of thing for years. In March 2025, Dario Amodei predicted AI would write 90% of code within six months and “essentially all of the code” within the year. That May he told Axios that AI could wipe out half of entry-level white-collar jobs within five years and push unemployment to 10% or 20%. In a later essay he predicted an “unprecedented” short-term labour shock alongside dramatic growth, and wrote: “AI isn’t a substitute for specific human jobs but rather a general labor substitute for humans.” Asked to forecast, his employees gave much the same answer he has been giving.

None of this is evidence one way or the other about who is right. So far the evidence on AI and jobs is mixed. A Yale team recently found that AI had mostly changed jobs and had “no connection” to changes in unemployment. Anthropic, the company whose [own prospectus warned](https://rews.cc/a/anthropic-prospectus-shows-42-billion-net-loss-ahead-of-poss-6f7244) that AI could pose an existential risk, is now [asking investors for a great deal of money](https://rews.cc/a/anthropic-would-like-100-billion-for-the-ai-its-ceo-says-sho-695b9d) on the strength of the more aggressive set of answers.

The staff forecast, then, runs like this: the US economy doubles every four and a half years, unemployment roughly triples, and a year from now Anthropic has 66% more employees. Of those three, only the last is something Anthropic can make happen by itself.