If you’ve noticed RAM prices going insane lately, you’re not imagining it. DRAM contract prices have risen more than 170% over the last year, and DDR5 chips used in AI servers have jumped nearly 500% since September 2025.
We’re living through what some are already calling “RAMmageddon,” and Micron just announced the most aggressive counter-move in the history of the memory industry.
Micron Technology has announced an investment plan of up to $200 billion to expand production capacity and tackle what it calls the most severe memory chip shortage in the last four decades, fueled by the artificial intelligence boom. Yes, $200 billion. That’s not a typo.
The AI hunger that broke the memory market
The root of the problem is pretty clear: AI is eating everything. Firms like Alphabet, Amazon, Microsoft, and Meta are set to spend about $650 billion on AI infrastructure in 2026 alone, up from around $360 billion last year.

All those Nvidia GPUs powering those data centers are absolutely starving for high-bandwidth memory, and that’s pulling supply away from everyone else.
As Micron’s EVP of Operations Manish Bhatia put it, high-bandwidth memory for AI chips is “consuming so much of the available capacity across the industry that it’s leaving a tremendous shortage for the conventional side“, meaning your phone, your laptop, your car.
The ripple effects are very real. Apple has warned the shortage will weigh on iPhone margins, and Elon Musk said Tesla may need to build its own fab to avoid hitting “the chip wall.”
What Micron is actually building
In Boise, Idaho, the $50 billion project will more than double Micron’s existing 450-acre campus with two new factories, each covering 600,000 square feet, with construction crews operating 24/7. Workers have already detonated over 7 million pounds of dynamite just to prepare the site. That’s the kind of commitment you don’t walk back.
The first Boise plant is expected to begin production in mid-2027, manufacturing high-bandwidth DRAM, a critical component for advanced computing. Both plants should be fully operational by end of 2028.

The plan also extends to New York State, where Micron opened a $100 billion complex near Syracuse, the largest private investment in the state’s history. A plant in Hiroshima, Japan adds another $9.6 billion to the total.
Micron’s Chief Business Officer Sumit Sadana summed it up well: “Memory has gone from being a system component to being a strategic asset.” And with the company currently fulfilling only 50 to 66% of customer orders, all HBM4 and HBM3e inventory sold out through December 2026, the urgency couldn’t be more obvious.
The company’s stock has reflected all of this, gross margins jumped from 18.5% in early 2024 to 56% in the most recent quarter, with expectations of hitting 68% this period. Business is booming, but so is the pressure to deliver.
Relief isn’t coming overnight. Intel’s CEO Lip-Bu Tan bluntly told attendees at the Cisco AI Summit: “There’s no relief until 2028.” So buckle up.
What do you think, is Micron’s $200B bet the right move, or is this too little too late? Drop your take in the comments, we want to hear from you!

