Meet this pick-and-shovel AI stock that just joined Meta, Tesla and Broadcom as the newest member of the trillion-dollar club

Meet this pick-and-shovel AI stock that just joined Meta, Tesla and Broadcom as the newest member of the trillion-dollar club
Meet this pick-and-shovel AI stock that just joined Meta, Tesla and Broadcom as the newest member of the trillion-dollar club

At the turn of the century, investors might have considered the idea of ​​a company reaching a $1 trillion market capitalization quite far-fetched.

Today, there are more than a dozen companies making this claim, including the artificial intelligence (AI) chip giant. NVIDIA It now trades with a market capitalization of more than $5 trillion. Unless valuations pull back sharply as AI hits a major roadblock, more companies are poised to surpass trillion-dollar market caps.

Did Nvidia miss out in 2009? This rare signal flashes again. In 2009, a “double down” signal appeared for a little-known chipmaker called Nvidia. For the first time in years, that same “Total Conviction” signal is flashing for a company that is 1/100th the size of Nvidia. Continue “

In fact, one just did. Meet This Pick-and-Shovel AI Stock That Just Joined Metaplatforms, teslaand Broadcom in the billion dollar club.

Image source: Getty Images.

Understanding Micron’s role in AI

What an incredible career the computer memory company Micron technology (NASDAQ:MU) has been on. The stock is up more than 227% this year and more than 867% last year (as of May 28).

The big reason Micron and some other related stocks have risen is because of the tremendous demand for memory, which has become a big part of the AI ​​story. Chipmakers like Nvidia need memory to sit directly in their graphics processing units (GPUs), which are key to training large language models. Memory is needed to feed the GPUs with data.

As Nvidia and other chipmakers have built more advanced GPUs and scaled GPU clusters, they need more memory to process data faster. There are different types of memory, but Micron makes a type called dynamic random access memory (DRAM), which was commonly used in older technologies such as computers and cell phones.

However, more recently, Micron has been developing a more complex memory product called high-bandwidth memory (HBM), which essentially stacks multiple DRAM chips. This is better for GPUs because it reduces the distance data must travel to reach them.

In March, Micron announced that it had designed a new HBM model called HBM4 with 36 gigabytes of 12-Gi memory, made specifically for Nvidia’s Vera Rubin GPU platform.

According to market research company Counterpoint, Micron held 23% of the global DRAM market share by revenue at the end of 2025, behind only Samsung (36%) and SK Hynix (32%). In the global HBM market, Micron had a 21% market share at the end of last year, still behind SK Hynix (57%) and Samsung (22%).

However, Micron’s share had more than doubled from just 9% at the end of 2024.

Source link