In May of last year WPP and Nvidia announced that they were jointly developing a content engine that harnesses NVIDIA’s 3D content development platform Omniverse  and artificial intelligence. The effort is designed to enable creative teams at the holding company’s ad agencies to produce high-quality commercial content faster, more efficiently and at scale for clients.   

In an update on that effort this week WPP said the work has advanced to creating 3D landscapes and other products based on new generative AI models developed by Nvidia.  

The tools allow creative teams to build on WPP visual content already in existence, or search for 3D objects from a Shutterstock library of nearly 1.3 million 3D assets.  

The tools—known as Nvidia inference microservices or NIMs—help build a virtual environment object by object, move items around on command, and assemble objects into scenes. And do it quickly and at scale. (In a post on LinkedIn, technologist and consultant Karston Wade described the newly unveiled NIMs as “essentially AI-model-optimization-as-a-service.” 

Much of WPP’s 3D product development work is being done by production staffers (trained by Nvidia) through the company’s recently launched Production Studio that is housed within its “intelligent marketing operating system” WPP Open. 

Client Coca-Cola Co. will be one of the first WPP clients to begin scaling generative 3D ads across markets globally.  

WPP has also been working with Ford to build real-time digital twins of its vehicles to create car configurators that customers can use online and that the company hopes will help generate demand for new cars before they hit the production line. 

WPP provided the update at SIGGRAPH, the annual computer graphics conference in Denver earlier this week.