Dell Technologies and NVIDIA are joining hands to provide computing infrastructure that supports the adoption of generative AI across a range of industries.

Dell Technologies announced that it held the '2024 Dell Technologies Vision Tech Industry Symposium' at the Grand InterContinental Seoul Parnas on Thursday the 23rd. The event discussed the current status and future prospects of the AI portfolio provided by Dell and Nvidia in areas such as finance, manufacturing, distribution, video, and public safety.

■ Dell Supports AI Deployment from Cloud to Edge




Dell Technologies recently announced that it is expanding its collaboration with NVIDIA, adding new servers, edge, workstations, solutions, and services to the 'Dell AI Factory with NVIDIA' to support rapid adoption of AI.

'Nvidia-based Dell AI Factory' is Dell's AI factoryThis offering integrates the portfolio with the NVIDIA AI Enterprise software platform. It is based on technologies such as NVIDIA Tensor Core GPUs.

In particular, the new product 'Dell PowerEdge XE9680L' unveiled by Dell is expected to provide high performance with a smaller 4U form factor and eight NVIDIA Blackwell architecture GPUs compared to existing models. The product provides 33% higher GPU density per node compared to the existing model. The Dell PowerEdge XE9680L server is scheduled to be released in the second half of 2024.

Dell emphasizes that it can support AI deployment across environments from the cloud to the edge. On the 21st, Dell unveiled the new Copilot+ PC product line that provides Microsoft (MS) AI experiences based on Qualcomm Snapdragon X Elite and Snapdragon X Plus processors. Dell said, "With this product, users will be able to experience an innovative PC environment based on local computing and processing across GPU, CPU, and NPU."

At the event, Dell CEO Yong Jeong introduced the Copilot+ PC product line, saying, "ChatGPT is an AI PC that runs on all applications on the PC. It will expand from personal PCs to B2B products in the future and demonstrate even more powerful performance."

He added, “Every data center must be designed to handle the speed and scale of AI, and Dell’s portfolio and partner ecosystem support AI deployments that are fully controllable across on-premises, edge, and cloud environments.” For example, Samsung SDS implemented generative AI on-premises by collaborating with Dell and receiving Dell PowerEdge XE9680.

■ Generative AI permeates every industry




At the event, experts from each field introduced application cases of solutions provided by both companies. Key speakers included H20 Managing Director Kwon Byeong-hee, Gauss Labs Vice President Choi Byeong-won, Leaders Systems Managing Director Cha Jeong-ho, E-Mart Manager Jeong Yu-seon, Twelve Labs Director Jeong Jin-woo, and NVIDIA Managing Director Kim Chang-min.

The Dell and NVIDIA collaboration will provide a full-stack offering, including computing, networking, and software, to a variety of industries around the world. “We plan to help customers build copilots, coding assistants, virtual customer service agents, industrial digital twins, and more,” NVIDIA CEO Jensen Huang said.

“We have achieved predictive quality management and digital twin-based process optimization using machine learning and AI vision,” said Kwon Byung-hee, managing director of H2O.ai, which advocates open source AI democratization. H2O.ai’s solutions enable global manufacturers to forecast supply chain inventory levels, demand, and predictive maintenance management.

For example, H2O.ai’s client Petronas used AI to collect and analyze 17,000 pieces of machine sensor data per minute to predict oil and gas equipment failures. And semiconductor manufacturer NXP achieved fully automated engineering with H2O.ai’s solution. LG Chem has identified factors affecting battery life and created a model to predict the remaining battery life in various devices.

'Twelve Labs' introduced a multimodal AI that understands images like humans. Twelve Labs possesses the image language generation model 'Pegasus-1' and the multimodal image understanding model 'Marengo 2.6'.

“The reason why computer vision has had difficulty spreading so far is because images are extremely complex multimodal data, and existing vision AI has limitations in analyzing images,” said Twelve Labs Director Jinwoo Jeong. “Twelve Labs has created a next-generation vision AI, an ultra-large multimodal image language model, so that AI can understand images, and searches based on text, similar images, and images are now possible.”

Twelve Labs has built an intelligent image system based on a platform based on Marengo and Pegasus. It extracts images into vectors in kb units, enabling various tasks, and as of March, it is recognized as the most promising image understanding model.

Image language models can be used in a variety of fields, including public safety, media content, sports, and national defense. For example, a body camera OEM provider stored 200PB of data in storage, which took a lot of time to analyze, but by utilizing image AI, the desired scene can be found quickly.