The hottest margin in AI hardware no longer belongs to the company selling the GPUs. It belongs to the company selling the scarce memory those GPUs cannot run without.
Micron Technology, Inc. (NASDAQ:MU) posted an 84.9% non-GAAP gross margin in its fiscal third quarter. NVIDIA Corporation (NASDAQ:NVDA) reported a 75% adjusted gross margin in its July quarter and expects that figure to fall toward 71% to 72% in the January quarter because memory costs are surging.
The comparison is not really perfect because the companies sell different products and account for them differently. But the larger point stands: the supplier controlling the scarcest component has the most room to raise prices.
Godlikeart/Shutterstock.com
Memory Is Eating the AI Stack
Micron’s quarterly revenue reached $41.46 billion, while its adjusted free cash flow hit $18.3 billion. The company has begun high-volume production of HBM4 designed for Nvidia’s Vera Rubin platform, while management expects tight DRAM and NAND conditions beyond 2027, with supply gradually improving in 2028. Under 16 strategic customer agreements, Micron Technology, Inc. (NASDAQ:MU) projects $22 billion of cash deposits and related financial commitments. Fourteen represent about $100 billion of cumulative revenue at minimum prices over their remaining terms.
Its 36-gigabyte HBM4 stack delivers more than 2.8 terabytes per second of bandwidth with over 20% better power efficiency than Micron’s HBM3E, according to the company.
NVIDIA Corporation (NASDAQ:NVDA) remains the far larger profit engine. Quarterly revenue doubled to $96.2 billion, including $89 billion from data centers. Yet management’s gross-margin reset shows that even the AI king must pay the memory toll.
Some of Nvidia’s largest customers have reportedly been told that prices for servers containing its AI chips will rise by more than 15% on systems shipping in early 2027. Management also plans on additional increases early next fiscal year but even that relief will not prevent the near-term margin slide.
The Bull and Bear Cases
The Micron bull case hinges on scarcity: HBM demand, long agreements and customer deposits could keep utilization and pricing unusually strong. The bear case is cyclicality. Customers are advancing cash because supply is short today; if that shortage breaks, 85% margins can normalize violently.
For Nvidia, rising revenue and its Rubin roadmap show that demand remains extraordinary. But lower margins reveal a real vulnerability: memory inflation and product complexity can absorb more economics even while GPU shipments soar.
Hedge Funds Are Chasing Both Winners
Micron’s hedge-fund holder count jumped from 154 to 185 in Q2, a 20.1% increase. Nvidia’s count rose from 277 to 286, up 3.2%. Short interest remains modest for both: 2.66% of Micron’s float and 1.23% of Nvidia’s float were short as of August 14.
Investors do not need Micron to replace Nvidia, just that the memory remains the scarce component. Right now, Nvidia’s own margin guidance is making that case for them.
READ NEXT: NVIDIA (NVDA): What Foxconn and Super Micro Are Telling Us about the AI Boom and Pony AI Is Scaling Robotaxis Fast—Can the Stock Reach BofA’s $17 Target?
Disclosure: None. Follow Insider Monkey on Google News.
Hedge Fund Resource Center