AMD is apparently no longer satisfied with being perceived as a perennial underdog, and is reportedly choosing to make a bold statement by pricing its Helios rackscale solution at a hefty premium to NVIDIA's Rubin. What's more, AMD has reportedly managed to rope in Microsoft as the product's first confirmed client, demonstrating the utility that Helios offers even at its supposedly elevated price point.
Futurum believes AMD will price Helios at $5 million to $5.5 million per rack, versus an expected retail price of $3.5 million to $4 million for the second-gen NVIDIA Rubin rack
As we explained in a dedicated post recently, Helios is AMD's first full-stack rack-level solution for AI workloads, consisting of:
- AMD Instinct MI455X GPUs
- 6th Gen AMD EPYC CPUs
- AMD Pensando AI Network Interface Cards (NICs)
- AMD Pensando DPU
- AMD Infinity Fabric
- AMD ROCm Software Stack
For the benefit of those who might not be aware, AMD's MI400-series GPUs leverage its latest CDNA 5 architecture, and sport increased HBM4 capacity and bandwidth, while bringing support for expanded AI formats with a higher throughput.
The AMD Instinct MI455X GPU that powers the Helios rack enables 40 PFLOPs of FP4 and 20 PFLOPs FP8 compute. For comparison, a NVIDIA Rubin GPU offers 50 PFLOPs of FP4 and just 17.5 PFLOPs of FP8 compute. Also, the MI455X GPU leverages 432 GB of HBM4 at a bandwidth of 19.6 TB/s vs. NVIDIA Rubin's 288 GB of HBM4 at 22 TB/s.
Next, the sixth-gen AMD EPYC CPU, codenamed Venice, is based on the new Zen 6 core architecture, and fabricated on TSMC's 2nm node. These CPUs offer up to 256 cores and 512 threads with eight massive compute dies and two even larger I/O dies.
The AMD Pensando "Vulcano" 800 AI NIC is an 800 Gbps high-performance network switch, with an 800 Gbps Ethernet network throughput. It is currently the only NIC to offer up to 2.4 Tbps of scale-out bandwidth per GPU, backed by full hardware and software programmability.
Finally, the AMD Pensando DPU bridges the AI servers to enterprise networks, improving network, security, and storage offloads to boost AI server efficiency, where each Salina DPU features 16 Arm N1 cores that are used for offloading networking, security, and storage for front-end server-to-client connectivity.
According to Futurum, AMD will price the Helios at $5 million to $5.5 million per rack, which would constitute a premium of around 40 percent relative to the expected price of $3.5 million to $4 million for the second-gen NVIDIA Rubin rack.
Even so, we would urge our readers to take these estimates with a pinch of salt. After all, BOM estimates are not always accurate, and ignore competitive pricing dynamics.
Finally, Microsoft has given its definitive consent to deploying the Helios rack for Azure AI services. In addition, AMD is also roping in OpenAI, META, Oracle, HPE, TCS, Celestica, Nutanix, and the US DOE as customers.
About the author: Writing is my one incontrovertible passion. Over the past six years, he has authored over 2,200 distinct articles on financial and tech-related topics, spanning nearly 1 million words. And he has been a member of Wcctech mobile team since 2025. As an alumnus of the University of Toronto, Rotman Commerce Program, I bring nuance, in-depth knowledge, and a unique perspective to every topic that I cover. When I'm not writing, I'm traveling the world, exploring hidden confectionaries and restaurants as an aspiring food connoisseur.
Follow Wccftech on Google to get more of our news coverage in your feeds.
