Dell and NVIDIA launch more cost-effective AI servers

During the Power Next Edge, Dell and NVIDIA announced an important novelty for professionals working with artificial intelligence. Companies are joining forces to launch AI servers that will power a variety of tasks such as speech recognition, cybersecurity, recommender systems and a growing number of innovative language-based services.

The new generation will have 15 server models power edge, and interested companies can now take the free trial. Check out what was announced.

15 New AI Servers with NVIDIA Ecosystem

A Dell announced that its next generation of servers power edge, focused on AI, will be released with support for AI applications focused only on developing artificial intelligence functionality.

The companies hope that the development will help professionals in diverse fields lead a more productive day-to-day through the advancements offered by the GPUs and DPUs of the NVIDIA, which were also developed and focused on the artificial intelligence market. The companies have been partners for 25 years and hope that this new launch marks a new stage.

As the amount of data in the world expands, most information technology capability will be at the service of machine intelligence. Building systems for AI first is a great opportunity for us to collaborate.

Michael Dell, CEO of Dell

Greater energy efficiency and savings

According to the press release, the new AI servers will be fitted with Bluefield Data Processing Units (DPUs), which can deliver "offloading, accelerating and isolating the operating system and networking stacks of the data center." In practice, it is possible to have access to a 25% reduction in the use of electricity, which will bring significant savings at the end of the month.

The expectation is to allow access through several channels: private installations, hybrid or completely online. Servers running the vSphere 8, NVIDIA's secure data center solution, are also able to deliver greater productivity for professionals working with AI. AI ecosystem applications include:

• Clara Parabricks: for genomics;
• MONGE: medical imaging;
• Morpheus: cybersecurity;
• Metropolis: video analysis in the retail sector;
• DeepStream: computer vision;
• Merlin: recommendation systems;
• And many others.

The big focus is on offering a server that manages to generate energy savings without the need to leave power aside, in addition to having AI applications that were developed to advance with different types of projects.

In addition to the Bluefield DPUs, H100 graphics cards, developed for the enterprise sector, will be included in the new AI servers. It has been promised that data processing will be 25 times faster compared to the previous generation, as well as Power Edge servers, with these GPUs, will be 300 times more power efficient for running inference on large language models – those exceeding 500 billion of parameters – when compared to previous generation non-accelerated servers.

Novelty can be tested for free

To allow those interested to already have access to the benefits of the new servers, a online page which must be filled out and shortly afterwards, if it is within the established standards, access to the preview will be released through the NVIDIA LaunchPad. The maximum power will be enjoyed without any payment required.

The launch forecast is for the fourth week of January 2023 and price information has not yet been reported. It is expected that companies that have high demand and need power to deliver their projects are the main target audience.

What did you think of the announcement of the two companies? Tell us in the comments!

See also other features

NVIDIA Retail AI Launches for Theft Prevention