Dell has teamed up with Nvidia to launched a new range of products catering for the demand for on-premises generative AI. The new range includes servers equipped with Nvidia H100 GPUs for training large language models.
The company hopes its on-premises solutions will offer a viable alternative to the cloud-based AI services available through hyperscale providers like Amazon’s AWS, Google Cloud and Microsoft Azure. The hardware could help businesses train or fine-tune in-house models rather than using those from OpenAI or Anthropic.
It builds on the partnership with Nvidia that Dell announced in May, known as Project Helix. The new expansion adds additional software products and expands the hardware range offered as part of the solution. This includes desktops, core data centres and edge locations.
A wide range of technology companies are embracing generative AI and foundation models. Microsoft has a close partnership with ChatGPT creator OpenAI and is embedding its technology across its product range. Google has invested heavily in large language models and Salesforce is pivoting to become an AI-first company. IBM has rolled out its WatsonX enterprise AI platform and ITSPs are investing in training and new AI platforms.
A report by Bloomberg Intelligence released in June suggests the inclusion and demand for generative AI could add $280bn to the global software market. The report suggests the generative AI market will expand from a 1% impact on the industry to 10% by 2032. Its biggest impact will be in infrastructure-as-a-service products for training models.
Dell says generative AI is driving ‘fundamental change’
Jeff Clarke, co-chief operating officer of Dell Technologies, said generative AI represents “an inflection point that is driving fundamental change in the pace of innovation while improving the customer experience and enabling new ways to work”. Clarke told a call with investors: “Customers, big and small, are using their own data and business context to train, fine-tune and inference on Dell infrastructure solutions to incorporate advanced AI into their core business processes effectively and efficiently.”
One of the main announcements from Dell was the “validated design for AI” product. An inference blueprint engineered with Nvidia, it is designed to speed up the deployment of modular, secure and scalable platforms for generative AI in enterprises. Traditional inference approaches have struggled at scale. Dell says its approach allows for faster, higher-quality predictions and decisions in real time and on secure, private data.
Content from our partners
These validated designs are pre-tested and proven configurations for Dell infrastructure that can be used to train AI models. This includes configurations for the Dell PowerEdge XE9680 or PowerEdge R760xa, with a choice of Nvidia GPUs.
Dell’s services division is also adding AI capabilities as it bids to keep pace with competitors such as Capgemini. This will include strategies for high-value use cases and a full-stack implementation for enterprise AI that includes software, validated designs and hardware solutions.
The company is also boosting AI power inside its workstations. Bringing the Nvidia RTX 6000 to its devices will allow developers to fine-tune and build generative AI models locally before deploying them to dedicated AI hardware at scale, it says.
