Hewlett Packard Enterprise has reportedly secured a contract to provide Elon Musk’s X, the location higher often called Twitter, with greater than $1 billion in AI-accelerating servers.
The settlement, first reported by Bloomberg, cites nameless sources that declare HPE is the most recent to learn from the Tesla tycoon’s synthetic intelligence ambitions.
Musk’s model-building outfit xAI, which makes its generative Grok mannequin primarily out there by way of the billionaire’s X, has a mixture of Dell and Supermicro methods in its 100,000 GPU Colossus supercomputer in Memphis, Tennessee.
That system is because of be upgraded with a further 100,000 GPUs, with plans to finally increase the system to one million accelerators sooner or later.
These HPE-supplied methods, as reported, are in our view both going to finish up at X to run AI inference and coaching work, or at xAI doing the identical. X and xAI are fairly tangled; as we stated, xAI offers a Claude/ChatGPT/Gemini-like mannequin referred to as Grok that netizens work together with primarily from X. Musk has beforehand redirected 12,000 Nvidia H100 GPUs initially sure for Tesla to xAI, arguing the elements would have sat unused in any other case.
Final month, X opened Grok and xAI’s new image-gen capabilities to all X-ers past the earlier limitation to X Premium subscribers. This month, xAI additionally launched a standalone Grok app for iOS. Broader availability of those companies probably will necessitate extra compute assets, particularly if future Grok fashions proceed to develop in dimension.
HPE, for its half, is not any stranger to accelerated computing. Its Cray division is chargeable for constructing a few of the largest supercomputers on this planet, together with the El Capitan, Frontier, and Aurora supercomputers, which rank among the many three strongest methods on the publicly identified world top-500 record.
Extra just lately, HPE introduced a sequence of recent liquid-cooled GPU methods out there along with your selection of Nvidia, AMD, or Intel GPUs and AI accelerators. Which of the three X has opted for has but to be disclosed, nevertheless, we’ll observe that almost all deployments by Musk’s enterprises have utilized Nvidia accelerators.
It is also doable X’s choice to supply methods from HPE might have been pushed by a need to search out various suppliers within the wake of Supermicro’s accounting drama.
The Register reached out to HPE and X for remark, and had not heard again on the time of publication. ®
Hewlett Packard Enterprise has reportedly secured a contract to provide Elon Musk’s X, the location higher often called Twitter, with greater than $1 billion in AI-accelerating servers.
The settlement, first reported by Bloomberg, cites nameless sources that declare HPE is the most recent to learn from the Tesla tycoon’s synthetic intelligence ambitions.
Musk’s model-building outfit xAI, which makes its generative Grok mannequin primarily out there by way of the billionaire’s X, has a mixture of Dell and Supermicro methods in its 100,000 GPU Colossus supercomputer in Memphis, Tennessee.
That system is because of be upgraded with a further 100,000 GPUs, with plans to finally increase the system to one million accelerators sooner or later.
These HPE-supplied methods, as reported, are in our view both going to finish up at X to run AI inference and coaching work, or at xAI doing the identical. X and xAI are fairly tangled; as we stated, xAI offers a Claude/ChatGPT/Gemini-like mannequin referred to as Grok that netizens work together with primarily from X. Musk has beforehand redirected 12,000 Nvidia H100 GPUs initially sure for Tesla to xAI, arguing the elements would have sat unused in any other case.
Final month, X opened Grok and xAI’s new image-gen capabilities to all X-ers past the earlier limitation to X Premium subscribers. This month, xAI additionally launched a standalone Grok app for iOS. Broader availability of those companies probably will necessitate extra compute assets, particularly if future Grok fashions proceed to develop in dimension.
HPE, for its half, is not any stranger to accelerated computing. Its Cray division is chargeable for constructing a few of the largest supercomputers on this planet, together with the El Capitan, Frontier, and Aurora supercomputers, which rank among the many three strongest methods on the publicly identified world top-500 record.
Extra just lately, HPE introduced a sequence of recent liquid-cooled GPU methods out there along with your selection of Nvidia, AMD, or Intel GPUs and AI accelerators. Which of the three X has opted for has but to be disclosed, nevertheless, we’ll observe that almost all deployments by Musk’s enterprises have utilized Nvidia accelerators.
It is also doable X’s choice to supply methods from HPE might have been pushed by a need to search out various suppliers within the wake of Supermicro’s accounting drama.
The Register reached out to HPE and X for remark, and had not heard again on the time of publication. ®
