Published on March 18, 2024, 9:15 pm

At the NVIDIA GTC event, Hewlett Packard Enterprise (HPE) unveiled updates to its AI-native portfolio, emphasizing the operationalization of generative AI (GenAI), deep learning, and machine learning (ML) applications. Among the highlights are the availability of two HPE and NVIDIA co-engineered full-stack GenAI solutions, a preview of HPE Machine Learning Inference Software, an enterprise retrieval-augmented generation (RAG) reference architecture, and support for future products based on the new NVIDIA Blackwell platform.

Antonio Neri, President and CEO at HPE, highlighted the importance of hybrid solutions for GenAI’s promise and full AI lifecycle management. He emphasized that collaborative innovation between HPE and NVIDIA is crucial in delivering effective AI software and hardware solutions to accelerate GenAI development from concept to production.

Jensen Huang, Founder and CEO at NVIDIA, emphasized how Generative AI can leverage data from various sources to drive industry breakthroughs. The collaboration between HPE and NVIDIA aims to empower enterprises by harnessing data to create and deploy new AI applications for business transformation.

The supercomputing-powered GenAI training solution by HPE offers organizations a preconfigured full-stack solution for developing large AI models. This turnkey solution powered by NVIDIA supports up to 168 NVIDIA GH200 Grace Hopper Superchips. It enables customers to accelerate GenAI projects such as LLMs, recommender systems, and vector databases while streamlining the model development process.

Additionally, HPE’s enterprise-class GenAI tuning and inference solution provide a flexible pay-per-use model through HPE GreenLake. Co-engineered with NVIDIA, this solution reduces ramp-up time and costs by offering tailored compute, storage, software, networking components necessary for producing GenAI applications within a hybrid cloud framework.

Furthermore, enterprises can turn proofs-of-concept into production applications with assistance from HPE Machine Learning Inference Software. This offering integrates with NVIDIA NIM to deploy ML models on a scale securely. Moreover, a reference architecture for enterprise RAG based on NVIDIA NeMo Retriever microservice architecture is available for businesses seeking customized chatbots or generators leveraging private data.

Looking ahead, HPE plans to develop next-gen solutions based on the new NVIDIA Blackwell platform featuring second-generation Transformer Engine enhancements to accelerate GenAI workloads further. Details about forthcoming products incorporating technologies like GB200 Grace Blackwell Superchip will be announced in due course.

In conclusion, the advancements unveiled at GTC underscore the continued collaboration between HPE and NVIDIA in driving innovation across AI landscape—enabling enterprises to harness Generative AI effectively towards achieving transformative outcomes.


Comments are closed.