SKF’s AI Garbo Splits Celebrity Fame From the Performance.

On September 30, 2026, SKF unveiled a campaign featuring an AI recreation of Greta Garbo to promote magnetic bearings. The industrial brand can now commission a new performance from a historical star, turning celebrity advertising into a combination of production technology, permissions and inherited reputation. Advertisers must reconsider what they are buying: a recognizable likeness, or a persuasive connection between that person and the offer.

This is a released campaign, not a demonstration of something that might eventually become possible. Its significance reaches beyond industrial marketing because the recognizable spokesperson and the newly created performance can now come from different places.

A famous face with a specific business job

SKF’s announcement connects Garbo to an early Swedish industrial film that featured the company before she became an international star. The new initiative pairs her recreated presence with magnetic bearings and a revived financing principle: customers fund investment through the energy savings the technology generates. SKF says the campaign was developed with her family and Harriet Brown & Company.

That combination matters. A famous person can attract attention to a difficult subject, but the commercial work happens when attention changes a buyer’s understanding of the proposition. Here, the underlying offer concerns adoption economics. The campaign needs to make a technical purchase and its financing intelligible, not merely make the manufacturer look culturally interesting.

The historical connection gives SKF a reason to use this particular star. Without such a connection, synthetic casting could become an expensive way to advertise the production method rather than the product.

Production access is not ownership of meaning

Reuters reported on October 2 that the production used several U.S. and Chinese AI tools. Garbo’s likeness was generated through prompts rather than copyrighted archival photographs or footage; her voice used authorized early-film audio, with an actor helping the delivery. That is a constructed performance assembled from different inputs.

The important shift is the separation of recognizability from a living performer’s participation. Creative teams gain another way to stage a campaign, while families, representatives and source-material owners can become consequential participants in its development. Easier image generation does not automatically simplify every permission or agreement around a recognizable person.

This also changes the brief. A traditional spokesperson contributes a present-day identity and can explain a decision to participate. A recreation supplies a performance selected by the advertiser. The audience may recognize the face while interpreting its appearance very differently from a voluntary contemporary endorsement.

Permission cannot settle the audience’s reaction

The Guardian’s Peter Bradshaw published a sharply negative review of the advertisement. That reaction is not evidence of commercial failure, just as publicity is not evidence of commercial success. It demonstrates that a technically ambitious campaign can still face disagreement over its creative value.

Authorization, realism and relevance answer different questions. Authorization concerns whether the production has the necessary support. Realism concerns whether the recreation convinces the eye and ear. Relevance concerns whether the association helps the intended audience understand or trust the business proposition.

A brand can satisfy one question and struggle with another. A convincing imitation may make viewers notice the construction more intensely. Equally, an openly synthetic performance might work if its purpose is clear. The strategic risk is treating recognition as proof that the intended meaning transferred.

The offer needs to survive the spectacle

For a manufacturer, the useful response is interest from people who can evaluate the technology, financing and adoption conditions. General fascination with a recreated film star could create a large audience with little purchasing relevance. Those are different kinds of attention and should lead to different judgments about the campaign.

OrionPilot’s published positioning connects business context, strategy, creative production and performance learning within a marketing operating system. For an engineering company considering this kind of campaign, that connection matters: the reason for casting a famous likeness should stay attached to the buyer problem and the evidence used to assess response. The creative decision becomes more useful when it can be examined alongside the commercial objective it was meant to serve.

The test is whether the audience remembers the proposition after the novelty fades. Requests for technical information or financing discussions would reveal something different from people sharing a clip because they dislike, admire or debate the recreation.

Reconsider the casting decision, not just the tool

SKF’s campaign broadens the available creative vocabulary. It does not establish that synthetic celebrity advertising is cheaper, more effective or universally acceptable. The cited announcements and reporting provide no comparative conversion result that would support those conclusions.

The decision now is whether a recreated personality carries meaning that the product actually needs. A heritage connection can make the choice coherent; a recognizable face alone leaves that work unfinished. Before expanding this format, advertisers should demand a reason for the casting as strong as the reason for the offer—and evidence that customers understood both.