Bezos at VivaTech 2026: AI won’t take away jobs, it will create jobs deficits – and why not everyone agrees
Main chat
A chat for vibe coders: news, guides, live cases, marketplace, and finding executors.
On June 17, 2026, on the stage of the largest technology conference in Europe — VivaTech in Paris — Amazon founder Jeff Bezos said a phrase that went through the headlines of dozens of publications: artificial intelligence will not destroy jobs, but lead to their shortage.
The statement directly contradicts what unions, some economists and even some figures in the tech industry themselves are saying. Let’s examine Bezos’ argument in its entirety, what his new startup Prometheus has to do with it, and why the reaction to these words has been so acute.
What Bezos said
In the session, moderated by former NASA astronaut Mike Massimino and next-door Blue Origin CEO Dave Limp, Bezos responded directly to the widespread alarm:
I know that a lot of people — including a lot of smart people — are worried that AI will make people unnecessary. I totally disagree with that view. And I think AI is actually going to lead to labor shortages.
The logic that he built next was not about replacing labor, but about expanding the number of tasks that, in principle, can be set and solved. "AI will create a labor shortage because it will allow people to discover more problems," Bezos said. The idea is that the constraint today is not a lack of ambition, but rather the barriers that technology can remove. The more problems can be formulated and tested, the more engineering and any other work emerges around their solution.
Prometheus: What is Bezos’s startup here
The argument is not abstract - Bezos illustrated it with his own new project. In 2025, he co-founded and co-CEO Prometheus, a startup that works on what Bezos calls an “artificial general engineer.”.
The idea of Prometheus differs from the usual large language models fundamentally. Bezos stressed that unlike text-trained models, Prometheus is built on engineering data designed to design physical objects — with the goal of dramatically accelerating the cycle from idea to product. He gave a figurative analogy to explain the boundaries of conventional language models: “Reading a thousand books about gymnastics does not make you a gymnast.”.
The practical goal is to compress the engineering development cycle. A ten-year development program can be reduced to five years, then to two, then to one, Bezos described the potential of the project.
At the time of the presentation, Prometheus did not release any products and did not publish detailed technical documentation on how the system works. This is a declaration of direction, not a report of the finished result.
The whole argument: from acceleration of engineering to shortage of personnel
Bezos’s line of reasoning goes like this: An AI capable of quickly prototyping, testing, and testing engineering ideas removes the barrier between an idea and its first test. That means more ideas are tested, more of them prove viable, and more engineering work follows.
In other words, according to Bezos, it is not AI that does work instead of a person – AI reduces the cost of trying, and as a result, the number of simultaneously going projects and directions grows so much that there are not enough specialists for everything.
Who disagrees and why
Bezos’s position is directly at odds with what other prominent figures in the technological and political environment say. Rishi Sunak, a former UK prime minister who is now an adviser to Microsoft and Anthropic, has previously said that AI is already affecting young people’s employment prospects.
Britain’s Trades Union Congress (TUC) has expressed concern that AI technology could repeat a “de-industrialization catastrophe” – a situation in which shareholders get richer while jobs “degrade or disappear.” At the same time, the TUC recognizes the potential benefits: with proper development, AI can have transformative potential, and workers could benefit from productivity gains.
A separate layer of criticism concerns not the content of the thesis, but the reaction to the details of the speech. Some commentators in Indian and other publications drew attention to the resonant quotes about the water consumption of data centers attributed to Bezos in some publications — noting that these specific formulations can not be found either in the video recording of the panel itself or in the official broadcast of VivaTech. This is an important reminder: with the rapid spread of resonant quotes through dozens of publications, some of the wording can be distorted or supplemented during the retelling, and it is worth checking the direct video recording where possible.
What the statistics say about the fact
It is useful to compare the statement with specific figures, not just opinions.
According to data compiled by the analytical publication IndexBox, May 2026 brought 38,579 AI-related job cuts – the highest monthly figure ever recorded. This doesn’t override Bezos’ long-term argument about new industries and new challenges, but it does show that the short-term effect on specific people losing their jobs right now doesn’t feel like a “labor deficit.”.
It is this gap between long-term and short-term effects that is at the heart of the discussion that has gone beyond one panel in one conference. Critics do not necessarily deny that new jobs will ever emerge. The question is whether they will emerge quickly enough to catch those whose current jobs are already disappearing. This is one of the defining economic disputes of the current moment in the development of AI.
Scope of investment as a background for discussion
Bezos’ statement comes amid an unprecedented influx of capital into the artificial intelligence industry. According to the research company PitchBook, presented on the same day on another stage of VivaTech, in the first quarter of 2026, $ 255 billion was invested in AI – more than in the whole of 2025 as a whole. Eighty percent of all venture capital investments now come from AI.
At the same time, McKinsey gave a much more restrained figure at a separate session of the conference: 80% of large companies invest in AI, but only 6% see a real impact on profits – a figure that has barely increased from 4% previously. The gap between the volume of investments and measurable returns is a separate, no less significant part of the overall picture, which fits Bezos's speech.
The Moon as a Way to Save the Earth
Bezos’s speech was not limited to the topic of jobs – and the second story helps to understand the general logic within which he talks about AI.
Bezos outlined a long-term vision for space exploration, using the phrase “limited by supply, not demand” – that is, in his opinion, the desire of people to expand into space already exists, just lack the infrastructure for this. The moon is convenient as a first step because of its proximity and resources: materials lifted from the lunar surface require 28 times less energy per kilogram than materials launched from Earth.
“We’re going to the moon to stay, not just to visit,” Bezos said. The long-term idea extends further: moving heavy industry—manufacturing, data centers, energy—beyond Earth to bring the planet back to its pre-industrial state. “We can have both progress and conservation of nature; we don’t have to give up one for the other,” he said.
There is a common thread with the thesis about jobs: in both topics, Bezos builds an argument through “not a choice between two bad options, but a technological path that removes the choice itself” – empowerment instead of compromise.
Context: a difficult year for Blue Origin
In the same session, Bezos touched on the recent failure of his space company: in May 2026, an unmanned New Glenn rocket exploded during a ground test at Cape Canaveral in Florida. “It was a blow in the breath for the whole team. But we realized that we were very lucky, Bezos said, noting that there were no casualties and that several critical elements of the launch infrastructure — including fuel systems — survived, although replacing them would have taken significantly longer.
Blue Origin CEO Dave Limp, who was present on the same stage, said restoration work is already underway and the company expects to resume launches before the end of the year.
Bottom line: dispute without final answer
Bezos’ speech at VivaTech 2026 is not neutral analytics, but the position of the interested party: the founder of an AI startup, who benefits from convincing the audience of the positive effect of this kind of technology. This does not automatically make an argument wrong, but the context is worth keeping in mind when assessing its credibility.
The question of whether AI will accelerate the emergence of new industries faster than destroying existing jobs or not remains open and unsettled. The May 2026 cuts statistics and Bezos’ optimistic thesis about the shortage of personnel describe, in fact, two sides of the same process, observed from different distances in time. Which of them will be more accurate on a scale of several years is a question that so far is answered only by further developments, not by the rhetoric of conferences.
*This article is based on the BBC, Euronews, TechGenyz, IndexBox, ThePrint and official VivaTech digests. Actual on June 22, 2026. *
