AI Generated Video Content to Drive Demand for GPUs, says Nvidia CEO

The generative artificial intelligence (AI) boom continues to fuel demand for Nvidia's advanced graphics processors, particularly driven by the need for new AI models capable of creating video and engaging in human-like voice interactions. Nvidia CEO Jensen Huang highlighted that the next significant development in AI involves systems that can process 3D video and other complex information, which will require substantial computing power.

Nvidia's advanced Grace Hopper chips, such as the H200, have already been utilized in multimodal models like OpenAI's GPT-4, which can conduct realistic voice conversations and interact with text and images. The demand for these chips has been bolstered by Nvidia’s customers, including Google DeepMind and Meta Platforms, who are also developing AI platforms for image and video generation.

The company's data center unit reported more than five-fold growth in sales in the first quarter, leading to a forecast of quarterly revenue significantly exceeding estimates. Derren Nathan, head of equity analysis at Hargreaves Lansdown, noted that the increasing need for large language models to be multimodal—capable of understanding video, text, speech, and both 2D and 3D images—has driven this broad-based demand.

Additionally, the automotive industry's use of AI models for video has emerged as a substantial driver of Nvidia chip demand. Tesla, for instance, has expanded its AI training processor cluster to around 35,000 H100s to support its autonomous driving goals. Nvidia's finance chief, Colette Kress, mentioned that the automotive sector is expected to be the largest enterprise vertical in Nvidia's data center business this year, underscoring the industry's significant role in the company's growth.

Overall, the rapid development and deployment of advanced AI models across various sectors, including automotive, content production, and more, are poised to continue driving robust demand for Nvidia's high-performance graphics processors.