French startup Omi has formed a technology partnership with Nvidia to accelerate AI-generated product visuals for consumer brands. The company, which specializes in 3D digital twins and generative AI, said the collaboration will help it generate images 10 times faster and serve twice as many simultaneous users. Omi also introduced ProductDrop, a tool that lets brands create highly accurate product visuals at scale.
Inside the Nvidia Partnership
Omi will use Nvidia RTX PRO 6000 Blackwell Server Edition GPUs to boost compute capacity and will receive guidance from Nvidia technical experts and engineers. The startup said this access will allow it to generate images 10 times faster and serve twice as many simultaneous users. The collaboration goes beyond chip supply by giving Omi direct support to optimize its AI workloads.
ProductDrop and Technical Pipeline
ProductDrop enables brands to generate ultra-faithful product images at scale by combining AI-generated scenes with realistic 3D Digital Twins. The feature is built on Nvidia OptiX and CUDA, along with open source image understanding and generation models including the FLUX family and Qwen-Image. Omi's pipeline also integrates Microsoft TRELLIS for 3D asset generation, and Verda optimized that inference serving to run 10 times faster, cutting overall latency by 4 times.
Digital Twins and Product Accuracy
Omi describes its Digital Twins as a product source of truth that prevents AI-generated scenes from altering packaging, colors, or label text. Generative AI can produce realistic environments but often distorts the product itself, creating a major challenge for brands that manage hundreds or thousands of SKUs. By combining AI creativity with approved 3D models, Omi allows teams to build entirely new visual worlds while keeping each product faithful to what consumers actually purchase.
Commercial Growth and Competitive Pressure
Founded in 2020 by brothers Hugo and Paul Borensztein, Omi now employs about 55 people across Paris and New York and serves more than 1,000 brands. The company has reached US$4 million in annual recurring revenue in the United States in less than 18 months and generates nearly 70% of its revenue outside France. Hugo Borensztein acknowledged competition from Adobe and traditional photography, but Omi has raised nearly 20 million euros and does not plan to raise new funds immediately.
Expansion Beyond Beauty
Historically focused on beauty and cosmetics, Omi is now targeting everyday consumer products and working with groups such as Procter & Gamble and Unilever on certain projects. The startup's platform hosts more than 50,000 digital twins and generates several million visuals each year for e-commerce, social media, and advertising channels. This diversification supports its ambition to become the product source of truth for consumer brands across broader categories.
What Comes Next
Omi will share more details about the new feature and the technology behind it at Nvidia GTC Berlin on October 21 at booth #7059. The announcement follows Nvidia's recent investments in other French tech companies, including Doctolib and Pigment. The event will give Omi an opportunity to demonstrate how accelerated computing can support product accuracy at global scale.
By aligning AI-generated creativity with 3D product accuracy, Omi is positioning itself as a product source of truth for consumer brands. The Nvidia collaboration gives the startup more computing power and technical depth as it competes with larger platforms and seeks new enterprise customers. Its ability to scale trustworthy visuals across catalogs, markets, and formats will likely determine how far this French startup can extend its international momentum.