AMD has expanded its AI infrastructure portfolio with the launch of Helios, an open, rackscale AI infrastructure designed for frontier AI and sovereign computing. Helios is built around AMD’s next-generation Instinct GPUs, EPYC Venice processors, Pensando networking and the ROCm software stack.
“Helios is AMD’s first complete AI rack system with GPUs, CPUs, and networking built together, instead of selling separate chips. It is well suited for training large AI models, memory heavy models, long context processing and high volume inference, and AMD’s biggest shot yet at challenging Nvidia’s dominance,” said Pareekh Jain, CEO at EIIRTrend & Pareekh Consulting.
AMD has also secured an early hyperscale deployment for Helios with Microsoft agreeing to deploy it to power its frontier model AI inference, its AI customers, and support Azure AI services.
The architecture behind Helios
The launch of Helios marks AMD’s latest attempt to strengthen its position in a market where Nvidia continues to dominate AI infrastructure. Unlike previous AMD AI offerings centred on individual accelerators, Helios is designed as a complete rack-scale system integrating compute, networking and software.
According to Jain, Helios goes up against Nvidia’s Vera Rubin rack. “Nvidia is faster on raw inference speed and has a faster internal connection between chips whereas AMD wins on memory size and offers better value for the price and power used. It’s standout feature is memory, where each rack packs about 50% more total memory than Nvidia’s competing system, which helps run very large AI models. It also uses open, industry-standard connections instead of Nvidia’s private technology, giving buyers more flexibility,” he said.
The AMD Helios rackscale design includes 72 AMD Instinct MI455X GPUs with AMD EPYC Venice CPUs and AMD Pensando Vulcano networking using UALink, optimized for compute, data movement, and system efficiency. The platform also supports both OCP and MX data types, delivering up to 2.9 EFLOPS of FP4 and 1.4 EFLOPS of FP8 compute for AI training and inference.
It also integrates 31TB of HBM4 memory with 19.6TB/s of memory bandwidth, while a liquid-cooling design uses quick-disconnect connections to efficiently dissipate heat. It is designed on open standards including OCP Open Rack Wide (ORW), Ultra Accelerator Link (UALink), and Ultra Ethernet Consortium (UEC) and can be scaled efficiently across datacenters while optimizing power, cooling, and serviceability for modern AI infrastructure, said the company.
On the security front, Helios incorporates a hardware root of trust and continuous attestation at every layer. It supports hardware-enforced isolation, encrypted memory and interconnects to help protect AI models, data and workloads in multi-tenant environments.
The software challenge
While the launch of Helios might help AMD close the hardware gap with Nvidia’s rack-scale systems, it will be the software compatibility that will be the real driver of enterprise adoption.
For this, AMD is expanding its ROCm AI software platform too, which supports frameworks including PyTorch, TensorFlow, and JAX, for enabling high-throughput inference and efficient distributed training while preserving familiar developer workflows.
Jain stated While hardware parity or superiority in memory bandwidth is achievable, software maturity remains the key differentiator for Nvidia. The Nvidia’s CUDA software has a 15-20 year head start, and almost every AI tool, tutorial, and codebase defaults to it.
He added software has been AMD’s weak spot. AMD has improved ROCm a lot but it still lags behind on the newest, most specialized optimizations, and setup is more complicated. For everyday AI work, ROCm is usable but for cutting-edge performance, CUDA still leads.
Evaluating the trade-offs
For CIOs evaluating AI infrastructure, Helios launch brings in another option to a market that has largely revolved around Nvidia’s dominance. But when considering Helios, CIOs will have to evaluate factors such as performance, software readiness, deployment models, procurement timelines and total cost of ownership before committing to a platform.
While AMD has not publicly announced a specific price tag for the Helios, Jain believes it to be noticeably cheaper to buy and run with lower chip prices and lower power use per GPU.
“It gives companies a real second option besides Nvidia, easing supply shortages and giving leverage in negotiations. The catch is software, where teams need to check whether their AI tools run well on AMD’s stack, since some advanced tools are still CUDA only,” Jain said.
For CIOs planning to deploy both, Jain warns the two systems can’t be plugged together into one combined machine as they use different, incompatible connection technology. But companies can and do run both side by side in the same data center, just as separate systems handling different jobs.