A large-scale lawsuit has begun in the United States against the world's largest semiconductor manufacturers: Samsung, SK Hynix, and Micron. The companies are accused of intentionally limiting the supply of standard dynamic random-access memory (DRAM) to artificially drive up prices amid the AI technology boom.
What Happened
A class-action antitrust lawsuit (Case No. 3:26-cv-06345) is being heard in the U.S. District Court for the Northern District of California against Samsung, SK Hynix, and Micron. Plaintiffs allege that the manufacturers coordinated to limit the supply of standard DRAM. This led to a sharp price surge of nearly 700% since 2022, creating a situation dubbed the "RAMpocalypse."
Context
Manufacturers justified the reduction in DRAM production by citing the need to reallocate capacity toward producing High Bandwidth Memory (HBM), which is critical for AI chips. However, plaintiffs believe that the shift to HBM was merely a pretext to create an artificial shortage in the standard segment. It is worth noting that these companies have a history of DRAM market price-fixing cartels in the early 2000s.
Why It Matters for the Industry
For the industry, a confirmation of these allegations could result in colossal fines and forced changes to production capacity management strategies. This would create high legal and market uncertainty for suppliers and could radically change the economics of building infrastructure for neural network training and inference.
Why It Matters for Users
For end consumers and businesses, the outcome of the case is critical: a victory for the plaintiffs could lower the cost of PC and server components and pave the way for large-scale compensation for overpaying for RAM in recent years. In the short term, the market will suffer from high hardware price volatility.
What Is Not Yet Known / Limitations
There are various assessments of the consequences: ranging from purely legal risks to the direct impact on hardware availability for local computing and the economics of AI startups.
Sources
Author
Look at AI, Editorial Staff