AI transformation has picked up tremendous velocity and enablement across the globe in the past year, thanks to generative AI (GenAI) and its potential for businesses: developing new features and products faster, boosting user productivity, enabling competitive advantages and more.
But bringing and expanding generative AI across an organization continues to be challenging, and businesses are looking for faster, comprehensive “easy-button” solutions.
In 2024 More AI Choices Drive Hesitation
Rewind to one year ago, when the Dell PowerEdge XE9680 server with eight NVIDIA H100 Tensor Core GPUs became available at NVIDIA GTC 2023—delivering a platform to accelerate generative AI applications. Dell quickly followed with complete solution blueprints, enabling organizations to jumpstart their AI development.
Today’s challenges reflect the progress the industry has made but still leave organizations short of realizing solutions that can move their business forward: high-quality results, data silos, trust, governance and security of their data, guidance and expertise to navigate through the myriad choices of models, architectures and solutions in market. Dell Technologies is bringing new approaches to enable customers and overcome these challenges, with our announcement today of the Dell AI Factory with NVIDIA.
Modernize Operations with End-to-End AI Solutions
Implementing an AI strategy and deploying it across the organization with a consistent framework and approach is difficult, even for the most advanced organizations. The Dell AI Factory brings all the necessary foundations and services into a turnkey approach to support the entire model lifecycle for AI and generative AI operations.
Dell and NVIDIA collaborated to deliver a fully integrated and seamless solution, designed to meet the varied needs within enterprises for AI and supports the needs for distributed, democratized AI operations across multiple locations, including from data center to edge and cloud.
The Dell AI Factory with NVIDIA is the industry’s first end-to-end solution integrating Dell’s compute, storage, client device, software and services capabilities with NVIDIA’s advanced AI infrastructure and software suite, all underpinned by a high-speed NVIDIA networking fabric.
Delivered as a fully integrated solution, it takes advantage of rack-level design, with rigorous testing and validation to deliver a seamless solution for transforming data into valuable insights and outcomes, specifically designed for enterprise data security and governance standards. The Dell AI Factory with NVIDIA is available now, offering enterprises a unique opportunity to harness the power of AI to drive innovation and achieve smarter, higher-value outcomes.
Deliver Better Results with New Capabilities
The growth in LLMs and use cases led early adopters to use pre-trained models and siloed approaches to develop uses for generative AI. Unfortunately, developers, IT and data scientists quickly learned that general-purpose LLMs hallucinate, that company data must always be secured behind the firewall and not be forwarded to the cloud and that fine-tuning and training takes time and computational power.
Enter a faster approach than fine-tuning on pre-trained models or building their own models (a much more challenging approach): retrieval-augmented generation (RAG). Not a new approach, RAG lets organizations integrate their data while reducing the need to rely on open-source models and expertise.
The Dell AI Factory with NVIDIA supports a wide array of AI use cases and applications. It includes end-to-end solutions, software, methodology and services to support the entire GenAI lifecycle from inferencing and RAG to model creation and model tuning.
The Dell AI Factory with NVIDIA broadens the GenAI portfolio with new solutions, allowing organizations to accelerate AI adoption in a trusted approach and integrate more of their data for higher accuracy results:
-
- New RAG-based capability enables customers to improve the quality of their LLM responses, reducing hallucinations. RAG reduces the computation overhead and expertise normally needed to fine-tune models. Customers can turbocharge their RAG deployments using NVIDIA NIM and NVIDIA NeMo Retriever microservices—coming to the NVIDIA AI Enterprise software suite—on their Dell infrastructure.
- Custom training model approaches enable data science teams to address complex problems with smaller language models built completely from an enterprise-specific data domain. Training smaller models also reduces computation overhead and hallucinations from open-source LLMs.
AI-optimized Innovation and Connectivity
As organizations expand their requirements and identify use cases that increase productivity and deliver value to their end customers, they look to adopt AI infrastructure strategies that serve today’s needs, as well as their future.
Today’s servers are assisted by powerful GPUs to enable faster time-to-results. And connecting data to compute requires high speed connectivity, optimized network fabrics and unfettered access to data, wherever it lives in the organization. Tomorrow’s demands will push technology even further to support multiple AI projects for multiple departments in an organization. Higher performance is required to support increasing inferencing demands and more RAG and model training for new use cases.
Dell Technologies is expanding our collaboration with NVIDIA to accelerate customers’ future requirements:
-
- The Dell PowerEdge XE9680 server is being enhanced to support NVIDIA H200 Tensor Core GPUs, as well as the newest NVIDIA B100 and B200 Tensor Core GPUs, delivering a clear performance path to sustain demanding organizations and multiple use cases.
- Dell is also announcing support for the NVIDIA GB200 Grace Blackwell Superchip as a standalone platform for specific deployments. More information will be available in the 2H of this year.
- Dell PowerEdge servers are also supporting the NVIDIA BlueField–3 SuperNIC, which accelerates scale-out connectivity for AI infrastructures, and when connected with NVIDIA Spectrum-4 switches as part of the complete NVIDIA Spectrum-X Ethernet for AI networking platform, enables a new level of optimized AI network communication performance. Furthermore, this high-performance AI fabric also powers the NVIDIA OVX architecture to accelerate generative AI workloads and enable digital twin simulations with PowerEdge servers.
- Dell PowerScale storage is critical for storing data sets, LLM models and development frameworks. The PowerScale F710 is now validated with NVIDIA DGX SuperPOD with DGX H100 systems. Customers can now build their AI solutions with PowerScale, the first Ethernet-based storage validated with DGX SuperPOD, ideal for data centers already designed with Ethernet. Dell PowerScale is a proven scale-out architecture that’s capable of keeping GPUs fully utilized while maintaining high availability, exceptional efficiency and protecting AI models from data poisoning, model inversion and cyberattacks.
- Dell’s NVIDIA-accelerated infrastructure solutions are combined with the NVIDIA AI Enterprise software suite, which now includes support for NVIDIA NIM inference microservices to supercharge production AI deployments with optimized inference on dozens of NVIDIA and community models.
Data Visibility and Expertise
Organizations often have silos of data, reducing the efficiency of finding and preparing the data for generative AI. And lack of expertise in generative AI can further cause businesses to fall behind, as they try to get their projects started. Finding the right data to deliver high quality results is key—as well as having the right partner to accelerate customer deployments.
Dell Technologies is accelerating time to business value today with:
-
- Full availability of the Dell Data Lake House, enabling customers to bring AI to their data, enabling data scientists to identify the right data sets for each GenAI project.
- Dell Professional Services continues to drive faster GenAI outcomes with new Services offers which:
- Accelerate the implementation of RAG systems.
- Implement model training infrastructure.
- Expand data management services for Generative AI, including data security.
Power Your AI Journey with the Leading AI Portfolio
With the new capabilities announced today in the Dell AI Factory with NVIDIA, organizations can drive faster ROI, as well as the outcome of faster time-to-results from their projects with an AI strategy that can address a multitude of their needs and use cases, all powered by a comprehensive end-to-end, turnkey AI solution. Dell Technologies expanding portfolio accelerates customers objectives with a validated infrastructure strategy, platforms and partners, which deliver innovation and proven approaches to accelerate insights, from their own data.
To learn more, visit Dell.com/AI.
