Hewlett Packard Enterprise (HPE) has secured a significant contract worth over $1 billion to supply AI-accelerating servers to Elon Musk's xAI, which is closely linked to the social media platform X.

The servers provided by HPE will enhance xAI's capabilities, particularly for AI inference and training, and are expected to power the Grok model.

Plans are underway to upgrade xAI's supercomputer with an additional 100,000 GPUs, with a long-term goal of expanding it to one million GPUs.

In December 2024, X expanded Grok's image generation capabilities to all users, moving beyond previous restrictions for X Premium subscribers.

xAI has also launched a standalone Grok app for iOS, which is expected to increase demand for computing resources as future models evolve.

Although the agreement with HPE was reportedly made at the end of 2024, the company has chosen not to comment publicly due to confidentiality.

Currently, xAI operates a supercomputer in Memphis, Tennessee, which consists of 100,000 GPUs sourced from Dell and Supermicro.

The news of HPE's contract was first reported by Bloomberg, highlighting the company's role in supporting Musk's ambitious AI projects.

X's decision to partner with HPE may be influenced by the need for alternative suppliers following recent accounting issues faced by Supermicro.

Grok, developed by xAI, is a generative model that competes with similar offerings like Claude, ChatGPT, and Gemini, and is primarily accessed through the X platform.

Musk has previously redirected 12,000 Nvidia H100 GPUs from Tesla to xAI, indicating a strategic move to optimize resource usage.

HPE is recognized for its expertise in accelerated computing, with its Cray division responsible for developing major supercomputers.