If your hardware budget depends on memory, the DRAM antitrust case is now a planning variable, not just a news item. A federal judge has let a price-fixing suit against the three dominant memory makers proceed, which means the legal fight over the recent price surge will run for months or longer, while the prices buyers pay are unlikely to change because of a pleading-stage ruling.

What happened

According to Tom’s Hardware, Judge Noël Wise denied a motion by Micron, Samsung and SK hynix to dismiss a proposed class action. The defendants argued the suit repeats an earlier antitrust case from 2018. The judge rejected that, citing new evidence that makes collusion plausible. TechPowerUp reports the ruling came on November 14 in the Northern District of California and found a plausibly alleged agreement under Section 1 of the Sherman Act.

The plaintiffs are 17 individuals and small businesses. Tom’s Hardware says they point to production cuts from October 2022 to April 2023, a pivot to HBM, DDR4 wind-down announcements within three months, stricter customer vetting, refusals to sign long-term DRAM contracts, Micron closing its Crucial consumer brand and no capacity expansion despite prices up as much as 500% in a year. TechPowerUp puts the conventional DRAM rise at 697% over 18 months and says the three firms held over 91% of DRAM revenue in Q3 2025.

The companies say these are independent, rational moves in a concentrated market. The judge wrote that the alleged conduct looks more like “hawks flying in formation.” Nothing is proven. The defendants must respond to the complaint by December 7, and a case management conference is set for January 2027.

Why it matters

The ruling matters mainly for what it signals. Courts at this stage ask only whether the allegations are plausible, yet the judge accepted that parallel behavior here may go beyond ordinary rivalry. Her distinction between the defendants’ earlier win and this case rests on newer conduct, such as the customer vetting and contract practices. That suggests discovery could expose internal communications about supply decisions, which is where cases like this are won or lost.

It also reframes a shortage many buyers treat as pure market physics. Part of the supply squeeze has an AI-driven explanation, since capacity moving to HBM leaves less for conventional parts. Whether that shift was independent or coordinated is the legal question, and the answer will take a long time. Our earlier piece on memory costs hitting hardware budgets through 2027 shows why buyers cannot wait for it.

What it means for you

Do not budget for relief. A lawsuit at this stage gives no refund and no price cut, so keep your 2027 forecasts for servers, laptops and devices on current memory pricing, with a contingency line for further increases.

Ask vendors direct questions. Check how much of a quoted price is memory, how long the quote holds and whether they can lock supply. If suppliers are not offering long-term contracts for conventional DRAM, as the plaintiffs allege, expect shorter price validity and more spot exposure through your OEMs and distributors.

Review your own exposure. If you bought DRAM or memory-heavy hardware directly, talk to counsel about whether you could fall inside a future class and what records to keep, such as invoices, quotes and contract terms. Small businesses are among the named plaintiffs, so this is not only a hyperscaler issue.

Reduce dependence where you can. Right-size memory in new designs, extend refresh cycles for machines that still do their job and avoid standardizing on DDR4 parts that suppliers are winding down. Our hardware coverage tracks related supply pressures.

What to watch

  • The defendants’ December 7 response, which will show whether they challenge the complaint again or answer it.
  • The January 2027 case management conference and the discovery schedule it sets.
  • Any announced DRAM capacity expansion, since the plaintiffs cite its absence as evidence.
  • Whether suppliers offer long-term contracts to conventional DRAM buyers.

Sources