How Gavin Newsom’s AI Executive Order Raises the Age-Old Question of Who Owns the Means of Production

California has just done something that, while obvious, seems revolutionary: it has stopped treating AI as technological innovation and instead as economic policy, wealth distribution and power.

Theexecutive orderthatGavin Newsomhas just issued doesn’t stem from the childish enthusiasm of “let’s prepare everyone for the jobs of the future,” but from a much more uncomfortable premise: there will be disruption, many sectors will be hit hard and many people will lose their jobs. Unless the state calculates the impact, anticipates and regulates the process, it will be the same old story.

California isn’t talking about training, reskilling, or AI literacy; tasks already underway through partnerships with Google, Microsoft, IBM and Adobe to provide certifications, internships and generative AI programs for public high schools and universities. What is important is the shift in tone: AI is no longer just a tool to increase productivity, and is now emerging as a phenomenon that may require unemployment insurance, compensation standards, equity participation, worker ownership models, changes to the WARN Act, public sector monitoring panels, and even mechanisms to ensure that a portion of the value generated by AI companies is allocated to public interest uses.

That shift is fundamental. During the early years of generative AI, the dominant narrative was a mix of tech evangelism, determinism and cheap consulting: “AI won’t take your job; someone who knows how to use AI better than you will.” It was a perfect line for conferences, courses, and headlines, but terrible as a social diagnosis. California seems to be beginning to understand that the problem isn’t whether every worker learns to write prompts, but what happens when companies discover they can reorganize entire processes, reduce entire layers of skilled labor and further concentrate the benefits of automation among shareholders, executives, and owners of computing infrastructure.

Newsom’s executive order holds significant symbolic value because it comes from the region that has benefited most from that concentration. California rightly boasts of being the world’s largest innovation ecosystem and home to 33 of the top 50 private AI companies. But precisely for that reason, its move is particularly revealing: when the very place that is the source of so much disruption begins to prepare to protect itself from it, it’s time for the rest of the world to stop seeing AI as an abstract promise and instead as a top-tier industrial policy.

The Stanford AI Index 2026 reveals several facets to the AI phenomenon, among them skyrocketing private investment, growing organizational adoption, rapidly expanding technical capabilities, and uneven labor market impact, particularly among younger workers in vulnerable occupations. According to Stanford, employment among software developers aged 22 to 25 has fallen by nearly 20% since 2024, and a significant portion of companies expect to cut staff in the coming year. Stanford also notes that Americans in general are particularly skeptical: only 33% believe that AI will improve their work, reflected by low confidence in the government’s ability to regulate it.

In other words, this is not a revolution taking place by stealth , but a growing tension between those who design the technology and finance it and those who will have to adapt to its impact. The experts tend to see the potential, while the public sees risk, opacity and loss of control: a technology advancing faster than society’s ability to understand or regulate it is not progress; it is simply power without checks and balances.

Of course, Newsom’s move hasn’t satisfied everyone. The California Labor Federation criticizes it for not going far enough: the governor acknowledges the potential harm to workers of AI, but mass layoffs should not be treated as inevitable, and instead as a political decision. That objection is highly relevant, because it highlights the difference between governing the transition and merely documenting it. A dashboard on the impact of AI may be useful; but if it only serves to certify the decline once it has occurred, it will be simply an elegant way of arriving too late.

The most thought-provoking part of the order is, precisely, the part that breaks with the usual language of employability. Talk of universal basic capital, worker ownership or a share in the value generated by productivity amounts to recognizing that AI is not just about wages, but ownership. If productivity no longer stems solely from human labor, and instead from models trained on collective data, massive infrastructure and an extraordinary concentration of capital, then the question ceases to be what laid off workers should do and becomes who is making a profit from automation; one that most governments continue to avoid.

California will be a laboratory not because its solutions are readily exportable, but because the contradictions wrought by AI will emerge there first. For decades, Silicon Valley has been the place where the promises of technology were first seen; now it may become where its social costs are first seen. Countries that today view AI as a race to attract data centers, startups and talent should pay more attention to Newsom’s executive order than to any keynote speech from Big Tech. True maturity does not consist of applauding every new model, but of asking what institutions will be necessary when those models begin to reorganize entire labor markets.

There is also an important difference from the early interpretations of generative AI. At first, there was much talk of augmented creativity, individual productivity, personal assistants and the democratization of knowledge. All of that exists, and it would be absurd to deny it. But the current phase is different: companies are no longer experimenting with chatbots, but rather redesigning cost structures. AI has gone from a tool in the hands of the worker to a talking point in the hands of the board of directors. That semantic shift is crucial: when a technology enters Excel, it is no longer an abstract, but a question of margins, workforce, and return on capital.

That is why California matters. Not because its governor has found some magic formula, but because he has put a more mature framework on the table: innovation, yes, but with early warning signs. Adoption, yes, but with collective bargaining. Productivity, yes, but shared benefits. Training, yes, but not as an excuse to abandon those who are displaced. Collaboration with big tech, yes, but without confusing the public interest with the sales pitch of those selling the technology.

The question for Europe, for Spain, and for any economy that aspires to more than just consuming tools designed in other countries is clear: are we going to wait for disruption to manifest itself in the form of unemployment statistics, or are we going to design the mechanisms for distribution, protection, and governance first? Because the consequences of AI are not inevitable. What is inevitable, if we do nothing, is that the usual story will unfold: socialization of costs, privatization of profits and an optimistic narrative carefully packaged to make what is actually a massive transfer of power appear to be progress

California has just recognized that AI will not be easy to regulate. Doing so will require data, rights, negotiation, taxation and discussions over ownership; all subject to a question that many would prefer not to hear: if machines are going to produce more wealth than ever before, why should we accept that this wealth belongs to fewer people than ever before?

—

This post was previously published on Enrique Dans‘ blog.

***