The three storage giants have been sued.

On June 25, a proposed class action lawsuit against Samsung, SK Hynix, and Micron was filed in the California federal court. The plaintiffs allege that the three companies used the AI and HBM transition as an excuse to shift more production capacity to high-margin HBM, while keeping supply of regular DRAM tight, ultimately driving up the price of ordinary memory significantly over four years.
Then on the same day, Apple announced price increases for iPad, MacBook, and other products, citing the rapid rise in memory and storage chip costs. Apple also said it had never seen component prices rise so much in such a short time.
This makes the lawsuit narrative very impactful. Upstream storage manufacturers say this is a structural supply-demand imbalance driven by AI demand, while downstream consumer electronics manufacturers say costs have become unbearable and they have no choice but to raise prices.
Consumers and businesses are beginning to question whether this is a normal cyclical price increase, or whether the oligopolists have completed a round of supply control by leveraging the HBM transition.
Of course, filing a lawsuit is not a conviction. The storage industry is inherently a highly concentrated, high capital expenditure, and highly cyclical industry. Synchronized capacity cuts do not necessarily equal collusion. There have been previous similar DRAM lawsuits that were dismissed due to insufficient evidence.
But this incident at least sends a signal: when storage prices have risen so high that Apple has to publicly raise prices, this round of AI-driven storage cycle is no longer just a story for the capital markets; it is starting to transmit to the real consumer side.
For investors, this indicates both that storage manufacturers have strong pricing power, and that there will be more regulatory scrutiny, lawsuits, and downstream backlash pressure ahead. The storage boom is still ongoing, but the narrative along this line will gradually shift from supply shortage to whether the price increases are sustainable and how long downstream can endure.
At the end of the day, this is still about resource repricing driven by the continuous expansion of AI capital expenditure. The training side causes HBM to seize DRAM capacity, while the inference side drives demand for server DRAM and enterprise SSDs, eventually spilling over and pushing up the costs of ordinary DRAM and NAND/SSD.