Bernstein sees something in AMD beyond Nvidia's shadow
Advanced Micro Devices (AMD)'s AI story might be getting a lot bigger in a place investors haven't been watching closely enough.
According to TheFly, Bernstein's latest note on the AI chip giant goes beyond GPUs, where it has been trying to close the gap with Nvidia.
It's about CPUs becoming a lot more important as agentic AI workloads spread across data centers.
That's the big step-change for Bernstein as AMD looks to rewrite its growth story.
It argues AI systems will need more general-purpose computing around accelerators, making server CPUs a larger part of the infrastructure buildout than investors may have assumed.
The firm now sees the 2030 server CPU market at $223 billion, comfortably above its prior estimate of $137 billion.
For AMD, that could turn EPYC from a strong side business into a critical part of the AI bull case.
What Bernstein sees in AMD beyond GPUs
Bernstein's latest AMD call saw the firm bumping its target on the stock to $600 from $525 and keeping an Outperform rating.
More AI:
- Goldman Sachs has blunt message for AI stock investors
- Microsoft CEO sends a blunt warning on AI and the tech ecosystem
- The next AI infrastructure race has nothing to do with chips
The firm sees AMD gaining from a broader shift in AI infrastructure.
The big phrase from the note was"CPU renaissance."
Bernstein's argument is that agentic AI entails far more general-purpose computing on accelerators, meaning CPUs may become a larger part of the AI data center buildout than previous assumptions.
That gives AMD another leg up on the competition, as it already has a stronger position in server CPUs through EPYC, giving it another way to benefit from AI spending.
Earlier this year, Bernstein was still cautious, keeping AMD at Market Perform even as it raised its target from $225 in January to $265 in April.
Though there was bullishness pertaining to improving server assumptions, but there were concerns over valuation, PC demand, customer concentration and AI execution.
That said, the new target suggests Bernstein now sees AMD as a structural AI infrastructure winner, not just Nvidia's backup plan.
Wall Street price targets for AMD stock
- Baird: $625. Baird analyst Tristan Gerra raised AMD's target from $300, citing stronger AI server positioning and likely gains in the data center chip space.
- Citi: $575. Citi analyst Atif Malik upgraded AMD to Buy from Neutral and raised the target from $460, saying Wall Street is underestimating AMD's GPU upside and Meta-linked AI opportunity.
- Bank of America: $560. BofA analyst Vivek Arya raised AMD's target from $500, calling AMD a top CPU pick as EPYC Venice, AI inference, and server CPU TAM growth strengthen the bull case.
- Barclays: $500. Barclays raised its AMD target, framing the stock as a compelling AI play as XPU compute shortages and rising CPU relevance make AMD more than just a GPU challenger.
- Goldman Sachs: $450. Goldman upgraded AMD to Buy and raised its target from $240, arguing AMD is becoming a major beneficiary of enterprise agentic AI and server CPU demand.
Why agentic AI could change AMD's server CPU story
Agentic AI changes the debate for AMD around GPUs.
Traditional AI demand has been dominated by training large models and running inference, with accelerators receiving most of the attention.
Bernstein's view is that the next phase of AI involves more autonomous systems that can plan tasks, call tools, search databases, execute workflows, and coordinate multiple steps without constant human prompts.
That kind of workload still needs GPUs, but it also requires more CPU capacity to manage everything from orchestration, memory movement, and general-purpose compute.
Bernstein argues that this could tighten the CPU-to-GPU ratio in AI systems, potentially moving from older assumptions of 1 CPU for every 4 to 8 GPUs toward 1:1 or higher in some inference-heavy setups.
That is important for AMD because EPYC already has strong server credibility. If agentic AI makes CPUs a larger part of AI infrastructure spending, AMD's server business becomes central to the bull case.
What has to happen next for AMD stock
For AMD stock, Bernstein's call raises the bar as much as it raises expectations.
The market is no longer just asking whether AMD can sell more AI GPUs.
It is asking whether EPYC server CPUs can become a larger part of the AI infrastructure trade as agentic AI lifts demand for general-purpose compute.
That comes at a time when tech stocks are under duress.
According to the BBC, the Fed held rates steady yesterday at 3.50% to 3.75%, while new projections showed inflation risk still sticky and nearly half of policymakers open to a 2026 hike.
That is not ideal for expensive growth stocks, because higher-for-longer rates make future earnings less valuable and push investors to demand cleaner execution.
According to Seeking Alpha, AMD stock is trading at 69 times forward non-GAAP earnings, 182% above the sector median.
AMD has the data to support the story. Q1 revenue rose 38% to $10.3 billion, and Data Center revenue jumped 57% to $5.8 billion.
Moreover, for perspective, AMD stock has sharply outperformed the broader market for the bulk of the year, rising 158.68% over 6 months and 139.30% YTD, compared with just 10.39% and 8.39% for the S&P 500.
However, it now has to prove that CPU demand, AI accelerator ramps, and large customer deployments can turn into sustained earnings growth.
Related: Qualcomm eyes $10 billion AI shortcut as smartphone growth slows
The Arena Media Brands, LLC THESTREET is a registered trademark of TheStreet, Inc.
This story was originally published June 18, 2026 at 10:07 AM.