HPE adds an AI training module to its GreenLake as a service platform
At the HPE Discover annual conference, HPE launched an exciting new offering for its GreenLake service platform. It's called HPE GreenLake for Large Language Models (LLMs), and it brings the power of high-performance computing (HPC) to artificial intelligence (AI) projects.
Training and deploying large-scale AI projects has been challenging, mainly because access to supercomputers is limited and expensive. Typically, only AI companies prioritise using them for AI training. But HPE is changing that with GreenLake for LLMs.
Justin Hotard, EVP and GM for HPC and AI at HPE explained that supercomputers are costly, complex, and not readily available on demand. HPE GreenLake for LLMs aims to solve these problems. It allows organisations to train their own LLMs using HPE's Cray XD supercomputers and Nvidia H100 GPUs whenever needed, without additional infrastructure or specialised skills.
To make this possible, HPE has partnered with Aleph Alpha, a German startup known for its Luminous software. Luminous is a pre-trained LLM optimised for analyzing and processing large amounts of data. Organisations in healthcare, finance, and the legal profession have successfully used it.
HPE has prioritised data privacy and sovereignty in developing GreenLake for LLMs. The customer's data is used only for their language model training, and this ensures adequate protection for trade secrets and sensitive information.
While this HPE offering will primarily benefit enterprises, including larger public sector bodies, startups in the AI space can also leverage it to accelerate their offerings and services.
It's important to note that GreenLake for LLMs is hosted in a remote private cloud facility in Canada, and therefore, it’s a true private cloud offering. This helps guarantee privacy and quality services.
HPE is committed to environmental sustainability and aims to provide a 100% carbon-neutral offering. The initial deployment uses nearly 100% renewable power and leverages liquid cooling technology to recycle heat.
While orders are being accepted now, availability is limited, and customers will be placed on a waiting list. North America will have additional availability by the end of the year, followed by Europe early next year. Pricing details are yet to be disclosed.
Overall, HPE GreenLake for LLMs is an exciting advancement that brings AI training capabilities to a broader audience, overcoming access and affordability challenges.
