AMD Is Building an AI ‘Escape Hatch’ From Nvidia. Here’s What It Means for AMD Stock.
AMD could be the AI industry’s best escape route from Nvidia.
Nvidia (NVDA) still stands strong as the dominant force in AI computing. However, Advanced Micro Devices (AMD) is building something that could provide another option to the industry's biggest AI companies. The idea of an "escape hatch" from Nvidia doesn't necessarily imply AMD needs to replace Nvidia. But if it can become a credible second source of AI computing for companies building enormous AI infrastructure, the opportunity for AMD stock could be much larger than simply selling more chips.
AMD stock has climbed 125.5% year-to-date (YTD), outperforming both the broader market and the tech-heavy Nasdaq Composite ($NASX). Yet, Wall Street expects AMD stock to more than double from here. Let's examine the ways in which AMD plans to do it.
1. AMD Is Becoming the Second Choice AI Companies Actually Want
For years, AMD's challenge in AI was to prove that its Instinct accelerators could compete with Nvidia's GPUs. While that is still true, the AI market has grown significantly larger and more complicated. AMD is evolving into a key infrastructure player for some of the world's biggest AI companies, including OpenAI, Meta Platforms (META), Anthropic, and Microsoft (MSFT). In the second quarter, its data center segment revenue more than doubled year-over-year (YOY) to $6.7 billion, with the segment now accounting for 58% of total revenue. AMD doesn't need to replace Nvidia everywhere to create a huge business. It must establish itself as a reliable and scalable option so that major AI companies want it in the mix. OpenAI and Meta have already committed to multigigawatt-scale deployments, while Anthropic plans to deploy up to two gigawatts of AMD's MI450-series GPUs through the Helios platform. The first gigawatt is expected to be deployed in the first half of 2027. Microsoft also intends to install Helios at scale via Azure. Specifically, AMD is building relationships that could eventually strengthen into very large deployments.
AMD is no longer approaching the market as a mere GPU supplier. And that matters because AI infrastructure is becoming much more complicated. Its Helios platform combines EPYC CPUs, Instinct GPUs, Pensando networking, and ROCm software into a rack-scale AI system that allows everything to work together. According to management, Helios can give up to 15% greater throughput at the same rack power and 30% more tokens per dollar than competitors. Nvidia's dominance stems from its efforts over the years to create an ecosystem around its GPUs. Now, AMD is developing its own entire infrastructure platform, giving investors another alternative.
Topics in this story
Gathered from external sources. Rights to this text belong to whoever originally published it.