Nvidia and AMD Positioned for Compute Demand Surge
Piper Sandler identifies Nvidia and AMD as primary beneficiaries of rising AI compute costs, citing structural shifts in training and inference requirements that favor established market leaders.
Rising compute costs are reshaping the AI hardware landscape, with GPU prices increasing by at least 25% year-to-date due to a self-reinforcing cycle of model training and agentic AI inference. Piper Sandler analysts, led by David O’Connor, identify Advanced Micro Devices and Nvidia as the top sector picks, asserting that compute capacity rather than model design is the primary bottleneck constraining industry scaling.
The firm rates both companies Overweight, viewing the current market structure as favorable for established players who can deliver the necessary infrastructure. This assessment reflects a shift where demand for general-purpose processors and specialized GPUs is outpacing supply, creating a favorable environment for companies with deep customer relationships and robust product roadmaps.
Nvidia Dominates AI Compute Market
Nvidia holds an 80% market share in AI compute, maintaining its position as the sector leader despite recent stock underperformance relative to peers. O’Connor characterizes this weakness as a temporary mispricing rather than a fundamental flaw, pointing to the company’s rapid product innovation as a key differentiator. The analyst sees large-scale gigawatt enterprise deals as the primary near-term catalyst for valuation expansion.
Nvidia trades at approximately 14 times forward earnings, near the low end of the 14x to 22x range assigned to the broader compute group. This valuation suggests the stock remains inexpensive relative to its growth trajectory, particularly given management’s guidance for $108 billion in sales for the current financial quarter with a 74% gross margin.
AMD Expands CPU and GPU Share
Advanced Micro Devices is gaining market share in CPU server chips as agentic AI systems require more general-purpose processing for multi-step task orchestration. This trend benefits AMD, which is taking share from Intel, while also seeing inference gains against Nvidia through its Helios GPU line. The company has secured anchor customers including OpenAI, Meta, and Anthropic for these chips.
AMD guides for revenue of about $13 billion in the current quarter, representing a 41% year-over-year increase. However, analysts note that the company needs to broaden its core customer base before it can realistically secure its own gigawatt-scale enterprise compute deals, a milestone that remains a key hurdle for further valuation growth.
Valuation Metrics Support Growth Thesis
Piper Sandler sets a $300 price target for Nvidia, implying 34% upside, and a $600 target for AMD, representing 15% upside. These figures reflect the firm’s confidence in the structural demand for compute resources, with both companies positioned to benefit from the increasing computational intensity of modern AI applications.
The analysis underscores that while both companies face different market dynamics, the overarching trend of rising compute demand provides a strong tailwind. The divergence in their current market positions highlights the importance of product mix and customer diversification in capturing the expanding AI infrastructure market.
