9621 Agnes Crossing, Lake Suzanneview, New Mexico Island 84604-9295.
At Discover, HPE announced its private cloud AI offering, developed jointly with Nvidia. The private cloud must bring together the best that both companies have to offer. Nvidia
At Discover, HPE announced its private cloud AI offering, developed jointly with Nvidia. The private cloud must bring together the best that both companies have to offer.
Nvidia is also the guest of honor at HPE Discover. HPE CEO Antonio Neri and Jensen Huang completely unpack the Sphere in Las Vegas. Every provider that wants to be competitive in the AI industry is forced to turn to Nvidia and HPE is also happy to put you in touch with the GPU specialist. During the keynote, HPE and Nvidia jointly announce their joint offering: Private Cloud AI.
After renewing their wedding vows at the Discover event in late 2023, HPE and Nvidia are now welcoming their first love child with Private Cloud AI. “Generative AI has enormous potential for business transformation. But the complexity of fragmented AI technology presents too many risks and obstacles that could endanger a company’s most valuable asset: its own data,” says Neri.
Private Cloud AI is the result you get when you put HPE and Nvidia’s portfolio in a blender. HPE builds Nvidia’s GPUs into its own HPC servers, which come in a variety of sizes and flavors. The Cray XD670 supports up to eight H200s, while a ProLiant DL384 or DL380a can accommodate eight GH200s.
Hardware is just one aspect where HPE and Nvidia meet. Private Cloud AI also uses Nvidia’s software for optimal GPU performance. Nvidia AI Enterprise lays the foundation for the AI and data software stack. This includes the NIM platform with microservices for inferencing. HPE then spices up the Nvidia platform with its own GreenLake portfolio. GreenLake for File Storage is certified to support intensive workloads on Nvidia hardware.
HPE ObsRamp will be given the responsibility of keeping an eye on things. HPE connects the observability platform with Nvidia hardware and software to give IT administrators insights to help them detect anomalies in AI infrastructure and workloads. To make these insights manageable, ObsRamp will have its own co-pilot. For example, HPE and Nvidia are trying to bring the entire AI stack into Private Cloud AI and give users full control over data, management and security of AI models.
The Private Cloud AI will be delivered from September with full support for Nvidia’s Hopper generation. But the future has also been considered. The portfolio will be converted to Nvidia’s Blackwell, Ruby and Vera architecture in the coming years.
HPE believes that Private Cloud AI is the perfect alternative for companies that want to get started with (Gen)AI but don’t want an expensive cloud subscription and/or prefer to keep their data in a controlled environment. “Large-scale operations of GenAI in the public cloud will eat up your IT budget in no time. With Private Cloud AI, we want to provide an enterprise-grade solution rather than a reference architecture, so that the burden of implementation is not on the customer,” said Neil MacDonald, head of HPE’s HPC business.
Running GenAI at scale in the public cloud will eat up your IT budget in no time.
Neil MacDonald, VP HPC & AI HPE
Source: IT Daily
As an experienced journalist and author, Mary has been reporting on the latest news and trends for over 5 years. With a passion for uncovering the stories behind the headlines, Mary has earned a reputation as a trusted voice in the world of journalism. Her writing style is insightful, engaging and thought-provoking, as she takes a deep dive into the most pressing issues of our time.
