Just days after the U.S. Supreme Court agreed to take up the Facebook/Cambridge Analytica securities case concerning risk factor disclosures (as discussed here), the Court has now agreed to take up yet another securities case, this time in a case involving Nvidia and involving the standards for pleading scienter and falsity under the PSLRA. The NVIDIA case involves alleged fraud in connection with the company’s disclosures concerning its sales of graphics processing units (GPU) to cryptocurrency companies as a component of its overall GPU sales. The specific questions the case presents to the Supreme Court concern what and how a plaintiff must plead when pleading scienter and falsity. Because the case involves the PSLRA’s “exacting pleading requirements,” the case potentially could prove to be very significant. A copy of the Court’s June 17, 2024 Order granting the petition for writ of certiorari can be found here.


Nvidia made GPUs intended for gaming applications. In or about 2017, cryptocurrency companies began using its GPUs for crypto mining operations. Sales of its GPUs reached record levels in 2017 and most of 2018 and the company told investors it was working hard to meet demand. However, when the company learned of excess supply of certain of its GPUs in certain distribution channels, it disclosed the excess supply of the GPUs, its share price declined, and plaintiff shareholders filed securities class action lawsuits against the company and certain of its directors and officers.

In their complaint, the plaintiff shareholders alleged that Jensen Huang, the company’s CEO, knowingly understated the extent to which demand for certain of the company’s GPUs was being driven by cryptocurrency miners (as opposed to demand from gamers). In their complaint, the plaintiffs did not, for example, cite internal documents reflecting GPU demand by crypto miners as opposed to gamers; rather, plaintiffs relied on an expert witness to analyze public data about activities of the crypto mining companies, using assumptions about the amount of computing power needed to facilitate the disclosed activities, and from that estimating the number of GPUs needed for the activity and what percentage of those GPUs would have been NVIDIA’s. From this “chain of assumptions,” the expert purported to determine what percentage of NVIDIA’s revenue was derived from crypto miners, the basis on which the plaintiffs alleged that the company’s actual dependence on crypto miners differed from the company’s disclosures.  

The plaintiffs also cited in their complaint statements from confidential witnesses – former employees — that the company had internal reports analyzing GPU sales and usage data. However, none of the former employees described the content of the reports or the data the CEO actually viewed. Instead, the company alleges, the plaintiff “speculates” what the report would have shown.

The District Court granted the defendants motion to dismiss. However, the Ninth Circuit, in a split opinion, written over the dissent of Judge Gabriel Sanchez affirmed in part and reversed in part the lower court’s dismissal.

NVIDIA filed a petition for a writ of certiorari to the U.S. Supreme Court. In their petition, NVIDIA said that the case presents two questions:

1. Whether plaintiffs seeking to allege scienter under the PSLRA based on allegations about internal company documents must plead with particularity the contents of those documents.

2. Whether plaintiffs can satisfy the PSLRA’s falsity requirement by relying on an expert opinion to substitute for particularized allegations of fact.

In its petition, NVIDIA argued that there is a split between the circuits on both questions. The company argued that five circuits (the Second, Third, Fifth, Seventh, and Tenth) have held that plaintiffs seeking to plead scienter based on internal company documents must plead with particularity the actual contents of those documents, while the First Circuit (and now the Ninth Circuit) allowed plaintiffs to proceed past a motion to dismiss in a securities fraud action based on allegations that internal company reports existed, combined with speculative allegations about what those reports might have said.

The company also contended in its petition that there was a split on the circuits with respect to whether plaintiffs could rely on experts’ opinions in pleading falsity. The Second Circuit and the Fifth Circuit have held that a plaintiff’s expert opinion could not substitute for particularized factual allegations of falsity, a necessary element of a securities fraud action.  The Ninth Circuit’s opinion in this case, in relying on an expert opinion to answer the critical question of whether the defendant’s statements were false, created a circuit split on this issue as well.

The company noted that on both of the questions, the positions of the Second and Ninth Circuit’s views diverged; given the extent to which securities litigation activity is weighted to these circuits, the company argued, the court’s consideration of these issues is warranted.

In their opposition to the company’s petition, the plaintiffs argued that their complaint was carefully researched and investigated, and indeed served as a “essential supplement” to the SEC’s charges against the company on the same allegations. (The company agreed to pay a $5 million settlement to the SEC on a neither admit-nor-deny basis.) The plaintiffs also argued that the company had mischaracterized their allegations, particularly with respect to the detailed testimony from the confidential witnesses, but also with respect to the content and purposes of the expert report to which they referred in their complaint. The petitioners also argued that there is no circuit split as described by the company.

A number of amicus briefs were filed in connection with the petition, including one by a long list of former SEC officials urging the court to take up the case. As the officials argued in their brief, “the court of appeals stripped the scienter requirement of any significant force and eroded the PSLRA’s heightened pleading standard.  Under the Ninth Circuit’s approach, a plaintiff can plead a Section 10(b) claim through a chain of speculative allegations and assumptions.”

In its June 17, 2024, order, the Court granted the company’s petition for a writ of certiorari. The case will now be heard during the Court’s 2024-2025 term.


It is an unusual development for the U.S. Supreme Court to agree to take up a securities case. So it is quite extraordinary that in the space of just a few days, the Court has agree to take up two securities cases to be heard in its next term, the Facebook/Cambridge Analytica case and now the NVIDIA case. The presence of these cases on the Court’s docket holds significant potential for important securities law developments during the next term.

The NVIDIA case could prove to be particularly important. The case involves the fundamental requirements for pleading securities fraud, including the requirements for pleading two bedrock pleading elements, scienter and falsity. Almost every motion to dismiss in a securities class action lawsuit challenges the sufficiency of the plaintiff’s allegations of both scienter and falsity. What the court has to say in the NVIDIA case on these two concerns could be of great significance to future motions to dismiss in securities cases.

While I am sure the plaintiffs are griped that the Court granted cert in this case, the company did make a valid point that it would create quite a mess in the securities litigation arena if there were to be a split between the Second and the Ninth circuits on issues as important as pleading scienter and falsity. The company correctly pointed out in its petition that in fact most of the securities litigation that is filed is filed in either the Second or Ninth Circuits. Diverging views on key securities issues between these two circuits could be quite disruptive and presents the prospect of diverging outcomes in similar cases simply because of the circuit in which the case was filed.

Indeed, in a particularly interesting aspect of the pleadings filed in connection with the company’s petition, Stanford Law Professor Joseph Grundfest, in an amicus brief filed in support of the company’s petition, argued that a statistical analysis is possible to support the court’s consideration of whether a split between the circuits in sufficiently important to warrant the Court’s review of a question. Grundfest argued that given the Second and Ninth Circuit’s “market share” predominance in securities litigation, the split between the two Circuits on the questions here particularly justified the Court’s taking up the case.

It will in any event be interesting to watch see how this case and the Facebook/Cambridge Analytica case proceed. At a minimum, the next Supreme Court term is sure to be an eventful one in the world of securities class action litigation.