Key takeaways
-
Nvidia and Micron Technology are drawing fresh optimism from analysts tied to continued demand for AI infrastructure.
-
For Nvidia, Raymond James analyst Simon Leopold is focused on the potential for AI hardware supply constraints to ease and on sustained GPU pricing power; for Micron, Melius Research’s Ben Reitzes points to long-term strategic customer agreements supporting memory margins.
-
Both analysts frame their views around AI data-center expansion—Nvidia via GPUs and Micron via high-bandwidth memory used to move data inside AI systems.
-
The key implication for investors is that AI infrastructure spending could remain a multi-year theme, but it is closely linked to how quickly manufacturers can expand capacity for constrained components.
AI infrastructure remains a central focus for investors as analysts continue to highlight the durability of demand for compute and memory components used in AI-accelerated data centers. Nvidia and Micron Technology are at the center of two separate bullish cases, with each argument centered on different parts of the AI stack: GPUs on the one hand, and high-bandwidth memory on the other.
What drove the analysts’ bullish view
On Nvidia, Raymond James analyst Simon Leopold said the company’s momentum reflects both the performance of its GPUs and the company’s strong position supplying enterprise AI-accelerated data centers. Leopold’s thesis also hinges on a key bottleneck: supply. Even as contract manufacturers work to expand production capacity, shortages of certain hardware components have persisted across multiple layers of the AI supply chain. In his view, securing additional manufacturing capacity could allow Nvidia to better convert demand into shipments and sales.
Leopold also argued that growth from non-hyperscalers could outpace growth from hyperscalers. If that holds, it could extend demand beyond the largest cloud buyers and support Nvidia’s pricing power for its GPUs, according to the analyst’s framework.
For Micron Technology, Melius Research’s Ben Reitzes is leaning on Micron’s critical role as a memory and storage supplier to AI systems, particularly through high-bandwidth memory (HBM). HBM is packaged alongside GPUs to enable high-speed data transfers, which is a key requirement for training large language models.
Reitzes tied his bullish outlook to both demand and pricing dynamics. He pointed to unusually strong demand for HBM that has translated into elevated gross margins. He also emphasized a more structural support: Micron’s strategic customer agreements (SCAs). According to Reitzes, Micron announced it had secured at least $100 billion in SCAs through 2030 for its memory solutions.
Locking in supply arrangements at known terms is particularly relevant in a sector that has historically swung toward commoditization after periods of premium pricing, Reitzes said. He noted that ongoing supply constraints for HBM, DRAM, and NAND flash could help Micron sustain its margin profile by limiting the ability of excess supply to quickly erode pricing.
Why Nvidia and Micron are being treated as part of the same AI build-out
Although Nvidia and Micron operate in different areas of the AI hardware stack, the underlying investment logic presented by the analysts is aligned: large-scale AI models require both compute capacity and rapid data movement. Nvidia’s GPUs are central to the compute layer, while Micron’s HBM is designed to sit close to GPUs to support the bandwidth needs of training and inference workloads.
That interdependence is why capacity constraints matter to both theses. If component availability improves—through expanded manufacturing capacity for the compute and supply chain inputs needed for AI servers—analysts expect companies to capture a greater share of the spending cycle. Conversely, if bottlenecks persist, the market may continue to treat near-term revenue conversion as limited even when end demand is strong.
Market implications investors will likely focus on
-
Supply normalization versus persistent constraints: Both bullish cases depend on the ability of manufacturers and component suppliers to keep pace with AI system demand.
-
Pricing power linked to demand breadth: Nvidia’s upside case includes expectations that demand from non-hyperscalers can support sustained GPU pricing, while Micron’s case emphasizes margin durability supported by long-term agreements.
-
Long-term contracting as a stabilizer: For Micron, the SCAs through 2030 are positioned as a way to reduce uncertainty around future supply demand at favorable terms, potentially supporting margins in a historically cyclical memory market.
-
AI capex cycle remains the dominant driver: In both arguments, data-center expansion is the primary tailwind; company-specific execution largely determines how much of that spending translates into earnings.
Looking ahead, investors are likely to watch whether AI server production and critical component supply can expand without interruption—especially as new system builds move from R&D deployment into broader rollout. Key read-throughs will include updates on manufacturing capacity, progress on supply constraints across the AI hardware chain, and additional disclosures around long-term supply arrangements that can affect visibility for memory pricing and margins.
