The Franco-American partnership expands the use of data, from generated content to atomic computing, to develop skincare and sunscreen formulas

Collaboration between L'Oréal e NVIDIA enters a phase closer to materials science than to marketing. After the announcement in June 2025, focused on the use of generative artificial intelligence, 3D rendering and tools to personalize the shopping experience, the French group announced on March 17, 2026 the extension of the agreement to computational chemistry applied to cosmetics. The transition is significant because it shifts AI from a creative and commercial production function to formulation research, that is, at the point where ingredients, textures and product performances are designed.

According to the material released by the company, L'Oréal will integrate NVIDIA ALCHEMI in your own environment Research and development to build an AI engine dedicated to beauty and skincare. The stated goal is to predict how molecules and ingredients might behave and interact at the atomic scale, before entering physical testing. This approach doesn't eliminate the experimental bench, but it aims to reduce the number of hypotheses to be tested, anticipating the most promising combinations with predictive models.

From product rendering to molecular simulations

The first chapter of the agreement, presented to Viva Technology in Paris on 11 June 2025, mainly concerned the use of NVIDIA AI Enterprise to develop and deploy AI applications in various beauty fields. Projects cited included CREAITECH, generative system for content and 3D rendering of products, and Noli, a multi-brand marketplace founded and supported by L'Oréal, built as an “AI Beauty Matchmaker”. The communication also indicated the use of over 1 million skin data and the analysis of thousands of formulations to profile needs and recommendations.

That model focused on the relationship between brands, consumers, and commercial channels: more scalable campaigns, reusable 3D assets, more granular recommendations across e-commerce, social media, and influencer marketing. The new phase, however, operates further upstream. The question is no longer simply how to represent, sell, or recommend a product, but how to select molecules, active ingredients, and combinations before they undergo stability, safety, efficacy, and sensorial testing.

The industry difference is significant. In marketing, AI improves the speed and variety of content production; in research, it impacts the hypothesis generation cycle. For a group with 40 international brands, 2025 sales equal to 44,05 billion euros, more than 95.000 employees, 22 research centers and over 4.000 scientists, even a partial improvement in formulation selection can have significant organizational effects.

ALCHEMISTRY and the decisive role of atomistic calculus

NVIDIA ALCHEMI, according to the official documentation of the American group, is a set of microservices NIM, software tools, and GPU-oriented components designed to accelerate chemical and materials discovery. The architecture addresses use cases such as batteries, catalysts, OLEDs, and beauty products, with atomistic simulations enabled by machine learning algorithm and learned interatomic potentials. NVIDIA describes the goal as the ability to achieve simulations close to the accuracy of quantum chemistry, but more feasible at industrial scale.

In the L'Oréal case, the press release talks about a discovery process 100 times faster compared to traditional methods. The data should be read in the context indicated by the source: it concerns the ability to simulate ingredients, performance, and textures in a virtual environment, simultaneously testing thousands of variables and reducing the journey from lab idea to finished product. This doesn't mean that every scientific and regulatory phase is automatically compressed to the same extent, but it does signal a change in methodology in the exploratory phase.

The two initial axes are photoprotection e skin tone managementThese are areas in which formulations depend not only on the presence of an active ingredient, but on the combined behavior of molecules, substrates, texture, stability, interaction with the skin, and perceived use. This is why simulation can be an interesting option: it allows for filtering options before material testing, with an approach closer to advanced materials research than to traditional cosmetics.

“We connect atomic-scale discovery to real consumer benefit,”

said Barbara Lavernos, Deputy Chief Executive Officer in charge of Research, Innovation and Technology at L'Oréal.

Lavernos's statement should be seen within a broader context: the cosmetics industry is increasingly driven by hybrid skills, where chemistry, modeling, data, AI, and experimental validation must coexist. Formulation remains a complex profession, but its scope is expanding to include computational infrastructures that until a few years ago were more common in the pharmaceutical, materials, energy, and semiconductor industries.

Data, creativity and research in the same architecture

The continuity between the 2025 and 2026 press releases lies in the attempt to unify multiple levels of the value chain: ideation, prototyping, content, recommendations, and formulation. In this reading, CREAITECH, Noli and the integration of ALCHEMI are not isolated episodes, but parts of a strategy in which Artificial Intelligence and Big Data they enter both into market-oriented activities and into the construction of the product portfolio.

In 2025 Asmita DubeyL'Oréal Groupe's Chief Digital and Marketing Officer linked the agreement with NVIDIA to changing consumer expectations, citing generative AI and agentic AI. In the same announcement, Azita MartinMartin, Vice President and General Manager Retail and CPG at NVIDIA, spoke about rapid innovation, scalability, personalized marketing, and more efficient advertising. In 2026, Martin herself brought ALCHEMI back to R&D flows, emphasizing the simulation of ingredient performance at the atomic scale.

“By integrating NVIDIA ALCHEMI into R&D workflows, L'Oréal can simulate ingredient performance at the atomic scale,”

said Azita Martin by NVIDIA.

The industrial focus is convergence. The same accelerated computing infrastructure that allows for the generation of 3D assets, the organization of creative pipelines, and the deployment of AI services can also be used for scientific simulations. It's not the same technical problem, but the same direction: transforming data, models, and computing power into an operational advantage. For the industry, this trajectory could widen the gap between companies with powerful computational resources and operators dependent on slower experimental cycles.

The implications for the young beauty-tech sector

The L'Oréal-NVIDIA case shows how the Digital transformation The beauty industry is no longer just about online sales, CRM, social commerce, or virtual configurators. The extension to computational chemistry introduces a deeper dimension: the ability to build predictive systems to select ingredients, reduce unpromising trials, and improve the consistency between scientific objective and final result.

This evolution has at least three implications. The first concerns skills: research teams must communicate with data scientists, software engineers, GPU specialists, and those capable of interpreting molecular models. The second concerns governance: if preliminary decisions are guided by models, traceability, data quality, simulation robustness, and error control become crucial. The third concerns the market: faster products to formulate are not automatically better, but they can increase competitive pressure on launch times, customization, and segmentation.

One necessary caution remains. In the beauty industry, physical validation, safety, stability, and sensory experience cannot be replaced by a model. Predictive AI can narrow the scope of options, but it cannot eliminate experimental work or regulatory responsibilities. Therefore, the true value of the agreement will be measured by the ability to integrate simulation and empirical evidence, avoiding treating calculation as an absolute shortcut.

The presentation of the new formulation science to the NVIDIA GTC AI Conference in San Jose, scheduled for March 16-19, 2026, confirms the desire to position cosmetics among the application sectors of advanced industrial AI. This isn't just a marketing gimmick: it positions beauty alongside materials, retail, manufacturing, and healthcare as a field where models, data, and accelerated computing can transform research.

For L'Oréal, the agreement with NVIDIA offers a way to connect global scale, data assets, scientific expertise, and computational capacity. For the industry, it signals a possible normalization of AI-assisted discovery Even in sectors historically based on iterative testing, formulation expertise, and brand awareness. The most concrete prospect is not fully automated beauty, but rather more selective research, in which hypotheses reach the lab already filtered by models capable of interpreting molecular interactions, expected performance, and application constraints.

Here are three insights that might interest you:

Gucci and Alpine: When Luxury Becomes a Platform in Formula 1
3D Printed Skin Will Replace Animal Testing
Sprout: When Medicinal Mushrooms Become Tailored Wellness