Quick Answer: A federal judge in the Northern District of California denied Samsung, SK Hynix and Micron's motion to dismiss a proposed class action accusing them of conspiring to restrict conventional DRAM supply. The judge found the complaint plausibly alleged an agreement under Section 1 of the Sherman Act. The allegations are unproven, and the case moves forward with answers due December 7, 2026.
A federal judge just decided that a DRAM price fixing lawsuit against the world's three largest memory makers will go forward. Judge Noël Wise ruled on November 14, 2026, and her reasoning matters more than the headline. She didn't just say the complaint was plausible. She explained which facts carry weight and which ones don't. If you buy memory for a living, or you follow the chip market, that's the part worth understanding.
What the Court Actually Ruled
Judge Wise denied the joint motion to dismiss in full. The suit was filed June 25, 2026, by 17 consumers and small businesses from five states. It targets conventional DRAM, the standardized memory built to industry specifications, including DDR5 modules used in computers.
The companies argued this was a rerun of DRAM II, the case where the Ninth Circuit affirmed dismissal in 2022 of claims over a 2016–2018 supply restriction involving the same three firms. Litnews reported that Wise rejected that framing outright: "The Court finds that it is not."
The allegations are accepted as true at this stage. Nothing has been proven, and the companies will continue to deny them.
The Core Problem: Parallel Conduct Without Direct Evidence
Most antitrust cases over price coordination have no smoking gun. No one finds a memo that says "let's cut supply." So plaintiffs rely on parallel conduct, where competitors make similar moves without an explicit agreement. Courts don't treat parallel conduct as illegal on its own. Competitors are allowed to watch each other.
The plaintiffs needed plus factors, additional evidence that points toward a prior agreement rather than independent decisions. Under Ninth Circuit precedent, the key test is whether a move was so self-damaging that "no reasonable firm would make the challenged move without such an agreement."
Here's where the case gets interesting. Wise said a pivot to HBM alone "might not be sufficient." Taken alone, moving wafer capacity to high bandwidth memory for AI accelerators is a rational business decision. HBM requires roughly three to four times the silicon wafer area per gigabyte compared to DDR5, so every wafer moved to HBM is one not making commodity chips. The complaint's own numbers show why that trade looks attractive for each firm.
The problem for the companies is what came next.
Why Long-Term Contracts and Fab Decisions Tipped the Scale
Wise found the related conduct "tips this factor over the edge." According to the complaint, that conduct includes:
- All three companies signed long-term contracts locking in HBM supply.
- Samsung and SK Hynix reportedly refused long-term agreements with major conventional DRAM customers.
- Micron announced on December 3, 2025, that it would close Crucial, its 30-year-old consumer brand.
- Samsung and Micron allegedly delayed or halted new DRAM plants.
Wise called several of these moves "difficult if not impossible to unwind." That's the core of her analysis. Fab capacity decisions take years and billions of dollars. A company that walks away from a consumer brand or refuses long-term supply deals is making choices it can't easily reverse.
She leaned on the Ninth Circuit's 1990 Petroleum Products decision, which noted that "Changes in refinery capacity are not readily reversible." She applied the same logic to fabs: "The Court views the commitment to build or not build manufacturing capacity as analogous."
Most people stop here, and that's a mistake. The next piece is what separates a plausible claim from a weak one.
The Counter-Intuitive Point: Rising Prices Aren't the Smoking Gun
The popular take goes like this: prices spiked, so someone must be cheating. Wise didn't rest on that, and you shouldn't either. Price increases alone are what a competitive market does when demand surges.
The complaint alleges conventional DRAM prices rose 171.8% in one year by Q3 2025, and 697% over 18 months. Those numbers are dramatic. But the court's real interest was in what the companies did while prices rose. The complaint says none of the three "expanded or announced plans to meaningfully expand commodity DRAM production capacity."
That's the difference. A high price with capacity expansion looks like a market responding. A high price with deliberate restraint, even as margins balloon, looks like a choice. Wise called the failure to expand "the sort that at the pleading stage, it appears no reasonable firm would make without a prior agreement."
The companies offered an innocent explanation: booming AI demand. Wise rejected it, pointing out that the customer-vetting practices all three allegedly adopted in January 2026 came after the AI boom and HBM pivot had already started. She said those near-identical practices "strongly support a plausible inference of conspiracy."
Where the Court Gave the Companies a Win
Wise didn't rule for plaintiffs on everything. Three of their arguments got zero weight, and the reasons are instructive for anyone building an antitrust complaint.
- Investor statements about "supply discipline." Micron's 2022 reference to supply discipline looked too similar to statements DRAM II had already found insufficient.
- Executives meeting at industry conferences. DRAM II "all but forecloses" relying on trade-show contact unless plaintiffs show what was actually exchanged or agreed. Wise said the allegations here went further than before, but still didn't clear that bar.
- Market structure. Three dominant firms controlling over 91% of global DRAM revenue as of Q3 2025 sounds damning. Wise called the oligopolistic structure "neutral in the analysis." Concentration is a precondition for coordination, not proof of it.
Wise also rejected the companies' argument that the 1998–2002 cartel was too old to count. The complaint says the companies and two other manufacturers fixed DRAM prices then. Samsung paid a $300 million fine, Hynix paid $185 million, and Micron received immunity as a cooperating whistleblower. The DRAM II court had also refused to discount that history, and Wise followed it.
The companies' best line at the hearing came from the late Judge Posner, who said competitors in concentrated markets "watch each other like hawks." Wise's reply was pointed: "their contentions are more akin to hawks flying in formation – an activity that hawks ordinarily do not do."
What Happens Next and What to Watch
The companies must answer the complaint by December 7, 2026. The initial case management conference is set for January 12, 2027, at 9 a.m. The timeline was pushed back about three weeks after the court granted an unopposed administrative motion to continue early deadlines.
The plaintiffs are represented by Bathaee Dunne LLP. Samsung is represented by O'Melveny & Myers LLP, SK Hynix by Cleary Gottlieb Steen & Hamilton LLP, and Micron by Gibson, Dunn & Crutcher LLP.
Here's the caveat. A denied motion to dismiss doesn't mean liability. Plaintiffs still have to prove an agreement, and DRAM II shows how a case can die later. Discovery will determine whether internal documents show a plan or only independent responses to a shared market.
Frequently Asked Questions
What is the DRAM price fixing lawsuit about?
It's a proposed class action by 17 consumers and small businesses against Samsung, SK Hynix and Micron. They allege the three companies conspired to cut conventional DRAM supply and shift capacity to HBM, raising prices. The court found the claim plausible at the pleading stage, but the allegations have not been proven.
How can I tell if a price fixing case is strong?
Look for "plus factors," evidence beyond parallel conduct. Strong signals include moves that are costly and hard to reverse, like long-term supply lock-ups, plant delays, and identical policies adopted at the same time. Weak signals include industry conference attendance and concentrated market structure alone.
Why did the judge reject the companies' argument that this was like DRAM II?
Judge Wise found the complaint differed in key ways. It added allegations about long-term HBM contracts, the Crucial closure, fab delays and near-identical customer vetting. Some arguments, like investor statements, were too similar to DRAM II, and got no weight.
When will the DRAM case move forward?
The companies must answer by December 7, 2026. The initial case management conference is set for January 12, 2027, at 9 a.m. in the Northern District of California.
Does a motion to dismiss denial mean the companies lost?
No. It only means the case can proceed to discovery. The companies still deny the allegations, and plaintiffs must prove an agreement later.
Watch the next filings. Discovery is where a case like this lives or dies, and the documents that come out will tell you more than any judicial summary can. If you buy memory, track contract pricing and capacity announcements from the three firms over the next year. Pass this to someone who is trying to understand why memory prices stayed high while factories stayed quiet.