Turn your texts, PPTs, PDFs or URLs to video - in minutes.
In a jaw-dropping moment today at COMPUTEX 2024, NVIDIA founder and CEO Jensen Huang revealed an AI avatar of himself during his live keynote. Built by Synthesia, the avatar is powered by EXPRESS-1, our latest AI model, which has been trained to understand the intricate relationship between what we say and how we say it, and to reproduce human emotions and mannerisms with incredible detail.
The keynote featured several chapters with Huang’s digital human counterpart narrating over video demonstrations. Clad in his signature leather jacket, the avatar was nearly indistinguishable from the real Jensen Huang, mirroring his facial expressions, tone of voice, and even his characteristic enthusiasm for technology.
Synthesia’s efforts to research and develop AI-powered digital humans is well-aligned with the work of NVIDIA, which has long been at the forefront of AI and graphics technology.
Introduced in April 2024, our EXPRESS-1 model has relied on a large cluster of NVIDIA H100 Tensor Core GPUs for multi-node, multi-GPU training, representing a significant leap in the development of synthetic media that has been made possible by NVIDIA accelerated computing. EXPRESS-1 uses large, pre-trained models as a backbone to drive the performance for our Expressive Avatars, combined with diffusion to model complex multimodal distributions. EXPRESS-1 predicts every movement and facial expression in real time, aligning seamlessly with the timings, intonations, and emphasis of spoken language. This results in performances that are astonishingly naturalistic and human-like, setting a new standard in the industry.
Synthesia also relies on high-end NVIDIA RTX GPUs to capture and process 8K footage, which is used to train our models and for post-production work. Our avatar rendering cluster in AWS has hundreds of NVIDIA GPU instances that serve video rendering traffic for customers who create AI videos on our platform.
The introduction of Huang’s AI avatar is more than a technological milestone; it's a glimpse into a future where high-quality content creation is no longer limited by expensive cameras or complicated editing software. Founded in 2017, Synthesia has spent the last four years building the world’s largest video communications platform that allows businesses and individuals to create, share and distribute AI videos at scale.
More than 55,000 businesses use our platform to change how they communicate and share knowledge using video. Companies using Synthesia to leverage video in these areas see massive benefits in the form of improved employee and customer experience, resulting in accelerated operational efficiency.
Users simply type in text, select from over 150 diverse AI avatars, and choose from more than 130 languages to ultimately produce videos that speak to a variety of company audiences and use cases, including employee training and development, customer enablement, internal communications, and personalized sales or marketing.
If you want to learn what we’re building next, join us on June 24 for a keynote presentation from Synthesia CEO and co-founder Victor Riparbelli.
And for a chance to try out an Expressive Avatar today, head over to synthesia.io/avatars or contact our team to join NVIDIA and dozens of other Fortune 100 companies that have joined the Synthesia community.
