Hewlett Packard Enterprise Co (NYSE: HPE) shares are trading higher on Tuesday. The company announced it will be one of the first to offer Advanced Micro Devices Inc’s (NASDAQ: AMD) “Helios” AI rack-scale architecture with an expandable Ethernet network.
What To Know: HPE has unveiled a major step forward in large-scale AI computing. The company is rolling out the first AMD "Helios" rack-scale architecture, designed specifically for cloud service providers (CSPs) who need to train and run massive AI models.
The system has built-in scale-up Ethernet networking, powered by HPE's Juniper Networking gear and Broadcom's Tomahawk 6 chip, all based on the open Ultra Accelerator Link over Ethernet standard.
HPE's new rack has 72 AMD Instinct MI455X GPUs, delivering 260 terabytes per second of bandwidth and up to 2.9 AI exaflops of FP4 performance. With 31 terabytes of next-gen high-bandwidth memory and 1.4 petabytes per second of memory throughput, it's designed to power today's largest and most demanding AI models.
HPE president and CEO Antonio Neri stated, “For more than a decade, HPE and AMD have pushed the boundaries of supercomputing, delivering multiple exascale-class systems and championing open standards that accelerate innovation. With the introduction of the new AMD ‘Helios’ and our purpose-built HPE scale-up networking solution, we are providing our cloud service provider customers with faster deployments, greater flexibility, and reduced risk in how they scale AI computing in their businesses.”
This is the first scale-up switch that has been made to function with AI workloads over standard Ethernet, created with the collaboration of Broadcom. The integration of HPE’s AI-native automation tools and Broadcom’s silicon means network operations are made more accessible, deployment times are significantly reduced and costs are lower.
**Hewlett Packard Enterprise Co (NYSE: HPE)** shares are trading higher on Tuesday. The company announced it will be one of the first to offer **Advanced Micro Devices Inc’s** (NASDAQ: AMD) “Helios” AI rack-scale architecture with an expandable Ethernet network.
**What To Know:** HPE has unveiled a major step forward in large-scale AI computing. The company is rolling out the first AMD "Helios" rack-scale architecture, designed specifically for cloud service providers (CSPs) who need to train and run massive AI models.
The system has built-in scale-up Ethernet networking, powered by HPE's Juniper Networking gear and Broadcom's Tomahawk 6 chip, all based on the open Ultra Accelerator Link over Ethernet standard.
HPE's new rack has 72 AMD Instinct MI455X GPUs, delivering 260 terabytes per second of bandwidth and up to 2.9 AI exaflops of FP4 performance. With 31 terabytes of next-gen high-bandwidth memory and 1.4 petabytes per second of memory throughput, it's designed to power today's largest and most demanding AI models.
HPE president and CEO **Antonio Neri** stated, “For more than a decade, HPE and AMD have pushed the boundaries of supercomputing, delivering multiple exascale-class systems and championing open standards that accelerate innovation. With the introduction of the new AMD ‘Helios’ and our purpose-built HPE scale-up networking solution, we are providing our cloud service provider customers with faster deployments, greater flexibility, and reduced risk in how they scale AI computing in their businesses.”
This is the first scale-up switch that has been made to function with AI workloads over standard Ethernet, created with the collaboration of **Broadcom**. The integration of HPE’s AI-native automation tools and Broadcom’s silicon means network operations are made more accessible, deployment times are significantly reduced and costs are lower.
<p><strong>Hewlett Packard Enterprise Co <a class="ticker-link" data-ticker="HPE" data-exchange="NYSE" href="https://www.benzinga.com/quote/HPE" target="_blank" rel="noopener">(NYSE:<a class="ticker" href="https://www.benzinga.com/quote/HPE">HPE</a>)</a> </strong>shares are trading higher on Tuesday. The company announced it will be one of the first to offer <strong>Advanced Micro Devices Inc’s</strong> <a class="ticker-link" data-ticker="AMD" data-exchange="NASDAQ" href="https://www.benzinga.com/quote/AMD" target="_blank" rel="noopener">(NASDAQ:<a class="ticker" href="https://www.benzinga.com/quote/AMD">AMD</a>)</a> “Helios” AI rack-scale architecture with an <a href="https://www.benzinga.com/node/49165485">expandable Ethernet network</a>.</p>
<ul class="wp-block-list">
<li><em>HPE is trading in a tight range. <a href="https://www.benzinga.com/quote/HPE" target="_blank" rel="noreferrer noopener">Check the full analysis here.</a></em></li>
</ul>
<p><strong>What To Know:</strong> HPE has unveiled a major step forward in large-scale AI computing. The company is rolling out the first AMD "Helios" rack-scale architecture, designed specifically for cloud service providers (CSPs) who need to train and run massive AI models. </p>
<p>The system has built-in scale-up Ethernet networking, powered by HPE's Juniper Networking gear and Broadcom's Tomahawk 6 chip, all based on the open Ultra Accelerator Link over Ethernet standard.</p>
<p>HPE's new rack has 72 AMD Instinct MI455X GPUs, delivering 260 terabytes per second of bandwidth and up to 2.9 AI exaflops of FP4 performance. With 31 terabytes of next-gen high-bandwidth memory and 1.4 petabytes per second of memory throughput, it's designed to power today's largest and most demanding AI models.</p>
<p>HPE president and CEO <strong>Antonio Neri</strong> stated, “For more than a decade, HPE and AMD have pushed the boundaries of supercomputing, delivering multiple exascale-class systems and championing open standards that accelerate innovation. With the introduction of the new AMD ‘Helios’ and our purpose-built HPE scale-up networking solution, we are providing our cloud service provider customers with faster deployments, greater flexibility, and reduced risk in how they scale AI computing in their businesses.”</p>
<p>This is the first scale-up switch that has been made to function with AI workloads over standard Ethernet, created with the collaboration of <strong>Broadcom</strong>. The integration of HPE’s AI-native automation tools and Broadcom’s silicon means network operations are made more accessible, deployment times are significantly reduced and costs are lower.</p>
<p><strong>HPE Price Action:</strong> <a href="https://www.benzinga.com/quote/HPE">Hewlett Packard shares</a> were up 0.23% at $22.01 at the time of publication on Tuesday, according to <a href="https://pro.benzinga.com/dashboard">Benzinga Pro</a>.</p>
<p><strong>Read Next:</strong></p>
<ul class="wp-block-list">
<li><strong><a href="https://www.benzinga.com/m-a/25/12/49154366/marvell-chases-amd-backed-celestial-ai-in-multi-billion-dollar-deal">Marvell Chases AMD-Backed Celestial AI In Multi-Billion Dollar Deal</a></strong></li>
</ul>
<p><em>Image: Miha Creative/Shutterstock.com</em></p>