VivaTech 2026: What the Luxury Industry Learns by Looking at What Others Reject

by Pascal Iakovou
0 comments

Two hundred thousand visitors, four thousand five hundred exhibitors, a stage where Emmanuel Macron, Narendra Modi, Jeff Bezos, and Bernard Arnault took the stage one after another: VivaTech has reached a whole new level for its tenth edition. But the data that matters to a company’s executive team isn’t found in attendance figures—it’s found in the six disagreements observed side by side on the same panels.

The most striking of these disagreements pits two different approaches to trust against each other. Delphine Viguier-Hovasse, Director of Innovation and Strategic Planning at L’Oréal, reminded Nigel Vaz (Publicis Sapient) and Drew Panayiotou (Keurig Dr Pepper) of a rule established in 2021 by the group’s ethics committee: no representation of beauty is produced using computer-generated imagery in any country where L’Oréal operates. This rule coexists with extensive use of AI elsewhere in the process—aggregation of clinical trial data, the Beauty Genius assistant that identifies users without ever altering their faces, and a digital hair twin that simulates thirty years of hair coloring. The group is accelerating AI wherever it aids decision-making, and halting it immediately whenever it touches on representation.

Drew Panayiotou argued the opposite: Dr Pepper owes its success to its customers’ adoption of the product on social media, which was fed back into the innovation pipeline by AI agents. Nigel Vaz put the debate into perspective with a startling statistic: 10 percent of companies consider AI central to their operations, while 73 percent already use it on a daily basis. The gap isn’t due to the technology itself but to the willingness to organize around it.

A little later in the program, Shantanu Narayen, CEO of Adobe, made a remark to Maurice Lévy (Publicis Groupe) that is rarely heard from an executive whose company sells mass-production tools: there is “AI slop”—content produced in excess and of no interest—precisely because Adobe’s tools make production limitless. Narayen himself sets the limit: “I’ve never met a creative professional who would rather spend six hours on a piece they could complete in an hour.” ” The question he raises is no longer whether machines are replacing manual labor, but what creators do with the five hours they’ve freed up—and whether those hours are spent on anything other than mere execution. With “content credentials,” Adobe has built a metadata system that accompanies every creation like a label indicating who produced it, when, and under what rights; the system now brings together thousands of partners, from camera manufacturers to distribution platforms.

The most subtle shift involves the pre-selection process. Cameron Fink, 20, co-founder of the American startup Aru, described on stage a mechanism that has no known equivalent in the luxury sector: out of 200 simulated concepts, only six make it to a human jury—having already been filtered by a model trained not on self-reported surveys but on traces of actual behavior: purchase histories, foot traffic, and music listening habits. The argument boils down to one figure: three out of every ten consumer products launched on store shelves are still available two years later—a failure rate that highlights the limitations of traditional surveys. For an art director, the challenge isn’t the abundance of data but the question of who decides—before a human even enters the picture—what deserves their attention.

On the hardware side, four companies—Essilor Luxottica, Meta, InBrain Neuroelectronics, and Garmin Health—have independently set the same goal: to become invisible to the wearer. The glasses Stellest measure the time worn without showing anything else; the Ray-Ban Meta use the tool with a movement they’ve already mastered. InBrain is developing implantable graphene interfaces for the central nervous system, in collaboration with Merck, the Mayo Clinic, and Microsoft, with a clause rarely formalized elsewhere: the data generated contractually belongs to the patient, and the device can be removed at any time. This is a governance standard that the luxury sector—which is also exploring connected devices—has not yet matched in writing anywhere.

“AGI is tired, but its usefulness is hardwired.”

Tony Fadell — co-creator of the Nest, VivaTech 2026

A Pew Research survey cited in the article shows that fewer than two in ten Americans expect AI to have a net benefit for society, compared with about four in ten who fear the opposite effect—and yet half of those surveyed already use a chatbot. The divide isn’t between supporters and skeptics, but within the user base itself.

None of these scenes represent a choice between acceptance and caution. Each one documents an actor who has chosen their own boundary and stood by it, whether publicly or in silence. In the United States, the agreement ratified in 2026 between the SAG-AFTRA union and the studios already regulates the use of synthetic performers; nothing equivalent exists for luxury boutiques, campaigns, or activations. The framework will emerge—whether through legislation or public pressure—before the fashion houses have to draft it themselves. VivaTech will be held again at Porte de Versailles from June 16 to 19, 2027; by then, the question raised this year will not have resolved itself.

Details

At InBrain Neuroelectronics, the data generated by the implant contractually belongs to the patient and the patient alone. If the patient wishes, the device can be removed at any time—and the data is erased along with it.

Cette publication est également disponible en : Français (French)

Related Articles