AMD’s stock has become one of the loudest success stories on Wall Street in 2026, and the reason boils down to a single word: inference. The chipmaker’s shares are trading around $506, up 136% year-to-date, and the latest catalyst behind that run is a new AMD AI inference partnership with Microsoft that puts AMD hardware inside some of the most advanced AI systems in operation today. The deal has reignited debate over whether AMD can finally close the gap with Nvidia, or whether its stock has simply run ahead of the fundamentals.
Summary
Key takeaways
- AMD’s data center revenue jumped 107% year-over-year in Q2 FY26, pushing total revenue up 50% to $11.54 billion.
- Microsoft is deploying AMD’s Helios AI rack platform for frontier model inference, a major vote of confidence from a top enterprise AI buyer.
- CEO Lisa Su expects data center revenue to more than double again in 2027, driven largely by inference workloads.
- Among Wall Street analysts, a Strong Buy consensus prevails, with an average price target reaching $647, suggesting roughly 28% upside.
- AMD trades at 68 times forward earnings, well above Nvidia’s 25x, keeping valuation concerns front and center.
AMD’s Strong Financial Growth Fueled by AI Demand
AMD’s latest quarterly numbers explain why investors have piled into the stock. The company’s data center division generated $6.7 billion in Q2 FY26 revenue, a jump that reflects how deeply AI infrastructure spending has reshaped AMD’s business in a single year.
Q2 FY26 Revenue Surge
Total revenue climbed 50% year-over-year to $11.54 billion, with data center sales up 107% over the same period. That segment now makes up 58% of everything AMD sells, according to figures reported by Crypto Briefing, a sharp shift for a company that spent years playing catch-up to Nvidia in AI silicon. Server CPU revenue alone grew more than 70% year-over-year, adding another layer to the growth story beyond GPUs.
Data Center Sales Growth and Future Projections
CEO Lisa Su isn’t treating the current quarter as a peak. She has projected data center sales will exceed $13 billion in Q3 2026, with the full segment expected to more than double again in 2027. Server revenue is also on track to grow more than 80% annually in the back half of FY26. The driving force behind that forecast is a shift AMD says is already underway: AI inference workloads are set to overtake training demand for the first time this year, a transition that changes which chips actually matter most to buyers.
Microsoft’s AMD AI Inference Partnership Reshapes the Competitive Picture
Microsoft has agreed to deploy AMD’s Helios rack platform specifically for frontier model inference, a decision that signals real confidence in AMD’s hardware for advanced AI workloads rather than just cheaper commodity computing.
Helios AI Rack Platform Deployment
Helios is AMD’s all-in-one AI rack system, bundling GPUs, CPUs, networking, and software into a single package designed to let customers move away from Nvidia-based infrastructure entirely. Microsoft’s decision to run frontier model inference on that platform is notable because frontier models are the most advanced AI systems in existence, and choosing AMD hardware for them suggests Microsoft sees the technology as genuinely competitive, not just a budget alternative.
Significance of Frontier Model Inference
Inference and training are fundamentally different workloads. Training demands raw compute power, while inference leans more on cost efficiency and reliability once a model is already built. That distinction matters because as the initial wave of AI training slows down, inference is expected to make up the majority of total compute demand going forward. Some estimates point to an inference market that could roughly double the size of the AI training market by 2032, reaching into the trillions of dollars. If AMD can capture a meaningful slice of that shift through deals like the one with Microsoft, the company would move from being a distant number two to a legitimate rival to Nvidia over time. This is one of the clearest signs yet that AMD’s strategy is less about beating Nvidia head-on in training and more about winning the next phase of AI computing outright.
AMD has been building out that strategy on multiple fronts. Its Instinct MI350, MI400, and MI450 GPUs are moving through development and deployment, with the MI400 series expected to ramp volume into 2027, while its EPYC “Venice” and “Verano” processors are built to complement those chips in AI-optimized servers. Phillip Securities analyst Yik Ban Chong, who rates the stock a Buy with a $755 price target, has flagged that AI startup Anthropic is expected to deploy 2 gigawatts of MI450 GPUs through the Helios platform starting in the first half of 2027, a rollout he estimates could add roughly $30 billion to AMD’s revenue. Meta has also signed a multi-year GPU deal with AMD covering up to 6 gigawatts of compute capacity, adding another large customer to the list.
Market Valuation and Analyst Perspectives
Wall Street remains broadly bullish on AMD even as the stock’s price tag draws skepticism. Of 33 analysts covering the company, 27 rate it a Buy and six are neutral, with an average price target of $647 implying about 28% more upside from current levels.
Wall Street Consensus and Price Targets
Vivek Arya, analyst at Bank of America, maintained his Buy rating and set a $620 price target, calling AMD the best-positioned CPU vendor in the market. He his projection for the 2030 server CPU total addressable market was increased to $210 billion, compared to the previous $170 billion previously, against a market that was worth just $35 billion in 2025. Arya pointed specifically to the Zen 6 Venice CPU, which offers up to 256 cores and 512 threads, ahead of Intel’s Diamond Rapids at 192 cores and 192 threads and Nvidia’s Vera chip at 88 cores and 176 threads.
Valuation Comparison Between AMD and Nvidia
Not every analyst is convinced the stock deserves its current price. AMD trades at roughly 68 times forward earnings, compared with Nvidia’s 25 times, even though Nvidia is still growing faster and holds a far larger share of the AI accelerator market. Chong recently cut his FY26 profit estimate by 8%, citing weak demand in AMD’s client and gaming units, which are being squeezed by elevated memory costs.
The bigger question for investors is whether the Microsoft partnership alone justifies AMD’s premium multiple. Analysts covering the stock generally agree that this deal is a meaningful step, but that AMD will need more major client wins beyond Microsoft and Anthropic to grow into its current valuation rather than rely on future promises. The anticipated MI450 deployment for Anthropic, expected to begin in the first half of 2027, will likely serve as the next real test of whether AMD’s inference bet is paying off or whether the market has already priced in more good news than the company can deliver.
FAQ
Why did AMD’s stock rise significantly in 2026?
AMD’s stock rose 136% year-to-date driven by strong AI data center demand, including a 50% revenue increase in Q2 FY26 and 107% growth in data center sales.
What is the significance of Microsoft’s partnership with AMD?
Microsoft partnered with AMD to deploy the Helios AI rack platform for frontier AI model inference, showing confidence in AMD’s technology for advanced AI workloads.
How does AMD’s valuation compare to Nvidia’s?
AMD trades at a 68x forward earnings multiple, which is much higher than Nvidia’s 25x, raising concerns about AMD’s premium valuation.
What are AMD’s expectations for data center revenue in the near future?
AMD expects its data center revenue to double in 2027, reflecting strong growth driven by AI inference workloads.
Article produced with the assistance of artificial intelligence and reviewed by the editorial team.
