I keep hitting the buy button on Advanced Micro Devices (NASDAQ:AMD | AMD Price Prediction) even though the trailing P/E stares back at me at 207, and I want to explain why in plain terms. My conviction here rests on hyperscalers refusing to let one vendor own 85% of the AI accelerator market forever. The anti-monopoly math only requires AMD to be the credible number two, and the receipts say it already is.
The Anti-Monopoly Math
The consensus $11.50 to $16.00+ EPS target for 2027/2028 only requires AMD to hold a 7% to 12% merchant accelerator share, and roughly 12% to 15% of the multi-GPU rack-scale tier, while the total addressable market keeps expanding and EPYC keeps taking server CPU share. Hyperscalers have a self-interest to fund exactly that outcome, because a single-vendor supply chain is a boardroom liability. So they are writing the checks. Meta committed to up to 6 GW of AMD Instinct GPUs. OpenAI signed on as a core preferred partner for 6 gigawatts. Oracle is building a 50,000-GPU AI supercluster on the AMD Helios rack design. That is the second-source demand curve showing up in ink.
The Data Behind the Conviction
Q1 FY2026 revenue landed at $10.25 billion, up 37.9% YoY, with Data Center alone at $5.775 billion, up 57% YoY. Non-GAAP EPS of $1.37 beat the $1.29 consensus, extending a streak in which four of the last five quarters cleared the bar. FY2025 free cash flow reached $5.519 billion, up 129.48%, and Q2 2026 guidance calls for revenue of roughly $11.2 billion, about 46% YoY growth. The balance sheet backs the ambition: net debt/EBITDA of -0.16 and interest coverage of 28.2x. That is a net-cash company funding its own hyperscaler pursuit.
Why Not the Obvious Alternative
NVIDIA (NASDAQ:NVDA) is the reflex trade. I own it too. I am not adding to it here because the anti-monopoly thesis is the mirror image of NVIDIA’s dominance, and the growth math already sits on a larger base. NVIDIA’s stock is up 27.13% over one year. AMD’s is up 256.99% over the same window, because the market is repricing the second source. Intel (NASDAQ:INTC) is the other name people mention, and I pass because AMD’s net margin of 12.5% and FCF growth of 252.96% YoY sit on the opposite side of Intel’s well-known margin struggles, while EPYC keeps winning sockets at AWS, Google Cloud, Azure and Tencent.
The Real Risk
China export controls are the real one. The MI308 restrictions cost AMD roughly $800 million in Q2 25 inventory charges and about $440 million net for FY2025. That is not a rounding error. I keep buying anyway because the Meta, OpenAI, and Oracle commitments are non-China demand, and the domestic hyperscaler pipeline is what the valuation is discounting.
Forward Conviction
The forward P/E of 70 is rich, and I do not pretend otherwise. But if AMD merely holds its second-source seat while the AI TAM keeps compounding, the earnings arrive. Until a hyperscaler cancels a gigawatt, my buy button stays active.
Contact [email protected] for any questions or corrections.
Top Gaining Stocks
Top Losing Stocks