When most people think about AI hardware, Nvidia is the name that comes to mind. But in the quieter corners of the semiconductor supply chain, Micron Technology has been putting up growth numbers that make Nvidia’s triple-digit gains look almost pedestrian.
In its fiscal third quarter of 2026, Micron posted revenues of approximately $41.5 billion, a 346% increase year over year. For context, Nvidia’s fiscal Q2 2027 revenue came in at roughly $96.2 billion, representing 106% year-over-year growth. Nvidia still earns more in absolute terms, but the rate at which Micron is accelerating has caught Wall Street’s full attention.
The HBM factor
Micron’s explosive growth traces back to one product category: high-bandwidth memory, or HBM. These are the specialized memory chips that sit directly on top of, or adjacent to, AI accelerators in data center servers. Without them, Nvidia’s GPUs can’t move data fast enough to train and run the large language models that companies are spending billions to deploy.
That supply-demand imbalance has pushed Micron’s gross margins to roughly 84.6%, a figure that would make most semiconductor companies blush. Its adjusted earnings per share hit $25.11 in the same quarter, reflecting a business that isn’t just growing fast but growing profitably.
To lock in that advantage, Micron has secured multi-year customer agreements valued at approximately $100 billion, extending through 2030. These contracts include floor pricing terms, which essentially guarantee a minimum price for Micron’s products regardless of market fluctuations.
Wall Street’s projections tell the story
Analyst consensus estimates project Micron’s fiscal 2027 revenue growth at around 84%, compared to roughly 42% for Nvidia.
Micron’s stock has significantly outperformed Nvidia’s shares over the past twelve months, showing hundreds of percent in gains for Micron versus lower double-digit returns for Nvidia.
The broader semiconductor industry is projected to generate profits of approximately $700 billion in 2027, with Micron and Nvidia expected to be among the largest contributors.
Why memory is the new bottleneck
Every new generation of AI accelerator, whether from Nvidia, AMD, or custom chip efforts at hyperscalers like Google and Amazon, requires more HBM per unit. Micron, alongside SK Hynix and Samsung, is one of only three companies in the world capable of producing HBM at scale. That oligopoly structure, combined with surging demand, creates pricing power that the memory industry hasn’t traditionally enjoyed. The $100 billion in multi-year contracts with floor pricing is a direct result of customers scrambling to guarantee supply.
There’s an important caveat here. Memory markets are cyclical, and the industry has a long history of overbuilding capacity during boom times, only to crash when demand softens. The floor pricing terms in Micron’s contracts mitigate this risk to some degree, but they don’t eliminate it entirely.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.
