Nvidia is a victim of the compute marketplace it created
TechCrunch
β’Thu, 09 Jul 2026 17:06:11 +0000
π° What Happened
Nvidia's stock price has fallen 15% from its May 2026 peak despite growing revenue, making it cheaper than the S&P average on projected earnings. Meanwhile, memory company Micron has nearly tripled in value as high-bandwidth memory becomes the new bottleneck for AI data centers, shifting investor focus from GPUs to memory solutions.
π The Backstory
Nvidia dominated the AI hardware market through its powerful GPUs and CUDA programming platform, becoming one of the world's most valuable companies. However, easing GPU shortages and the rising importance of HBM (high-bandwidth memory) have shifted market dynamics. Memory components, which previously received less attention, are now critical to AI performance as data centers scale up.
π― Why It Matters
The shift signals a maturing AI hardware market where different components cycle in and out of being the primary bottleneck. Nvidia's relative decline suggests the AI infrastructure buildout is entering a new phase where memory, networking, and specialized accelerators all play crucial roles, not just raw GPU compute.
Nvidia's stock price has fallen 15% from its May 2026 peak despite growing revenue, making it cheaper than the S&P average on projected earnings. Meanwhile, memory company Micron has nearly tripled in value as high-bandwidth memory becomes the new bottleneck for AI data centers, shifting investor focus from GPUs to memory solutions.
Nvidia dominated the AI hardware market through its powerful GPUs and CUDA programming platform, becoming one of the world's most valuable companies. However, easing GPU shortages and the rising importance of HBM (high-bandwidth memory) have shifted market dynamics. Memory components, which previously received less attention, are now critical to AI performance as data centers scale up.
The shift signals a maturing AI hardware market where different components cycle in and out of being the primary bottleneck. Nvidia's relative decline suggests the AI infrastructure buildout is entering a new phase where memory, networking, and specialized accelerators all play crucial roles, not just raw GPU compute.