Flying in Formation: Why a DRAM Price Case Survived Its First Test

  • Tech
  • November 16, 2026
  • 0 Comments

On Nov 14 a federal judge in California declined to throw out a consumer lawsuit accusing the three largest makers of memory chips of fixing prices, and the decision turned a complaint into a case with a schedule. Judge Noël Wise, sitting in the Northern District of California, found that the plaintiffs had pleaded enough to proceed. Samsung, SK Hynix and Micron had asked her to dismiss the action jointly. They now face discovery.

The most quoted line in the order describes how the three companies expanded capacity and set prices: in formation, like eagles. It also records that memory prices had risen by roughly 500% over the period the plaintiffs describe. Those two observations carry the weight the ruling needs. A price-fixing claim that survives a motion to dismiss usually does so because the court accepts, for now, that the pattern is too convenient to be coincidence.

That is a lower bar than it sounds and a higher one than these defendants usually face. Most antitrust class actions end at this stage, dismissed for want of a plausible agreement. Clearing it means the case acquires subpoenas, depositions, expert reports and a calendar measured in years. Companies settle here more often than they win here.

The plaintiffs’ theory is not that the three firms met in a room. It is that each announced capacity plans and price increases in a sequence that let the others read the signal, and that none broke ranks by building aggressively while margins were at records. That pattern is what the phrase about eagles is meant to capture.

A DRAM market with three suppliers is a market where the argument is available in a way it would not be with ten. Prices are set by how much capacity the incumbents choose to build, and capacity takes quarters to arrive. Each firm knows that adding supply lowers the price for everyone. Restraint pays whether or not anyone coordinates it.

Lawyers for the three companies declined to comment on the ruling, which is the standard response and also the only one available to them. The case will now be argued in filings rather than in statements.

The cross-border mechanics of the litigation will matter as much as the legal theory. Two of the defendants are Korean, one is American, and the documents that will decide the dispute sit in Seoul, Icheon and Boise. Requests for records held in Korea meet a second set of rules at home, and in comparable cases that has meant months of argument over what may leave the country and in what form. Discovery in San Jose will run at the speed of the slower jurisdiction.

The three companies also answer to competition authorities outside the United States, and an order holding that their conduct is plausible enough to litigate gives those authorities a document they did not have before. Regulators rarely lead on price coordination in components. They follow, and they cite.

There is a version of this case in which the defendants are right. Memory is a capital-intensive business with violent cycles, and the period the plaintiffs describe was one in which every producer had reason to be careful about building. A firm that under-invests during a shortage loses share; a firm that over-invests as demand turns loses its balance sheet. Moving in step can be the rational response to a shared signal rather than evidence of an agreement.

The plaintiffs will have to defeat that explanation twice: at class certification, where they must show that damages can be calculated across a class of buyers, and at trial. Certification is the harder gate. Memory chips end up in laptops, phones and servers sold by companies that are not parties to the suit, and tracing an overcharge through three layers of distribution is the kind of problem that has ended antitrust cases before.

The damages, if they arrive, would be large. American antitrust law allows an award to be trebled, and a 500% price move over a defined period gives an economist a base to work from. The exposure the three companies carry is not a fine that a general counsel can reserve for. It is an unquantified liability attached to the most profitable part of the semiconductor industry.

For buyers of memory, the ruling changes less than it appears to. DRAM prices are set by supply and demand, not by courtrooms, and the current shortage will ease when the capacity now under construction comes online.

What the ruling changes is the cost of being one of three. The companies will spend several years answering questions under oath about how they decided to build, when they decided it, and who they spoke to first. A market with that few participants can be managed quietly. It cannot be managed quietly while a judge is reading the emails.

That is the part worth watching. The suit will probably settle, and the settlement will probably be sealed. Discovery happens first, and discovery is where a concentrated industry’s internal reasoning becomes a public record.

Related Posts

  • November 20, 2026
  • 8 views
Micron Has Sold 75% of Its 2027 Memory Output Before the Plant Opens

Micron Technology has committed 75% of the high bandwidth memory it expects to produce in 2027, a year for which most of that capacity has not been built. The company…

  • November 20, 2026
  • 7 views
Tesla Opens a Truck Charging Stop and Sketches a Car With No Pedals

A charging site beside a highway outside a Texas city is an odd place to begin a vehicle program. On Nov. 19 Tesla opened its first Semi charging station in…