Micron discontinues Crucial: a memory brand available to consumers and businesses who build their own PCs. The allure of massive AI profits proves too strong.
Why produce DRAM for PC users when the same production line can churn out custom memory chips for AI? Micron could no longer answer that question and is discontinuing its Crucial memory brand. With this move, the memory manufacturer exits the consumer market, which will have adverse effects on the price of computer memory.
Micron is the world’s fourth-largest memory manufacturer, with a 23 percent market share. Samsung (32 percent) and SK Hynix (38 percent) are larger. Micron is exiting the consumer market to focus on the enterprise market. Meaning: Micron will maximize its production capacity to build lucrative memory for AI chips.
read also
AI Makes DDR4 more Expensive than DDR5
The manufacturer will continue to sell Crucial products until February 2026 and will also honor its warranty obligations. Micron still plans to sell DRAM memory directly to PC manufacturers.
Even Higher Prices
However, this step is another setback for everyone except RAM manufacturers, AI chip makers, and large players like Meta, who are massively deploying these chips in data centers to develop new LLMs. The memory market is already facing shortages and high prices. The end of Crucial will further exacerbate these shortages. Demand for competitors’ products will rise, but they are unable to meet it and, moreover, are eager to sell to AI developers themselves.
The AI boom is putting significant strain on the PC component supply chain. After the demand for AI accelerators drove up GPU prices, memory prices will now also continue to skyrocket. Counterpoint expects a doubling, and PC manufacturers share that fear.
read also
Reduced Memory, Higher Prices: HP and Other PCs to Become More Expensive in 2026
SSDs are not faring much better. The NAND chips in SSDs are also made by Micron, Samsung, and SK Hynix, and there too, availability is decreasing and prices are rising as the focus shifts to AI.
Consumed by Hyperscalers
A laptop with 64 GB RAM is already generously equipped. A single Nvidia HGX B300 system with eight Blackwell Ultra GPUs can easily have more than 4 TB of memory on board. Nvidia has millions of orders for Blackwell GPUs to process, which themselves require large amounts of HBM memory and fit into memory-rich systems. Buyers are primarily large cloud providers and AI developers, including AWS, Microsoft, and Google.
The demand for AI thus creates a complex situation, where the big tech giants greedily absorb the available production capacity for PC components. Those who need PCs or servers on a more modest scale must increasingly dig deeper into their pockets.
Micron’s decision to now bury Crucial is yet another sign of the distorted market situation caused by the boundless hunger for more and larger AI systems. Ironically, the fact that hyperscalers are buying up the memory market makes the offerings of those same hyperscalers more competitive against on-premises hardware, which is gradually becoming less affordable.