IN THE UNITED STATES DISTRICT COURT FOR THE NORTHERN DISTRICT OF CALIFORNIA SAN FRANCISCO DIVISION
October 12, 2011
IN RE NVIDIA CORPORATION
The opinion of the court was delivered by: Richard Seeborg United States District Judge
ORDER GRANTING MOTION TO DISMISS SECOND CONSOLIDATED AMENDED CLASS ACTION COMPLAINT
Lead plaintiffs Roberto Cohen and the New Jersey Carpenters Pension Fund and New Jersey Carpenters Annuity Fund bring this putative class action for securities fraud against defendants NVIDIA and its founder and CEO Jen-Hsun Huang. On July 2, 2008, NVIDIA issued a press 21 release disclosing that it was taking a $150 million to $200 million charge against its cost of revenue 22 to cover expenses arising from defective packaging of its semiconductor chip sets. Plaintiffs allege 23 that NVIDIA was aware the problem affected a significant amount of products previously sold to 24 customers including Hewlett-Packard (HP) and Dell and therefore the company should have 25 disclosed the potential losses earlier. Plaintiffs are suing on behalf of all persons who purchased 26
NVIDIA common stock between November 8, 2007 and July 2, 2009 (the "class period"). On October 19, 2010 the Court dismissed plaintiffs' Consolidated Amended Class Action Complaint (CAC) with leave to amend. In re NVIDIA Corp. Secs. Litig., No. 08-CV-04260 RS, 2010 U.S. NO. C 08-04260 RS Dist. LEXIS 114230 (N.D. Cal. Oct. 19, 2010). Plaintiffs subsequently filed a Second Consolidated Amended Class Action Complaint (SCAC). Defendants presently move to dismiss this complaint.
For the reasons stated below, defendants' motion to dismiss the SCAC is granted without leave to amend.
NVIDIA is a semiconductor company that designs and sells graphics processing units (GPUs) and media and communications processors (MCPs). Original equipment manufacturers (OEMs) incorporate these semiconductor chips into computers that they assemble for sale to 9 consumers. The NVIDIA products interact with other components in the computer to enable 10 graphics display and other communications functions. The two products account for approximately 80% of NVIDIA's revenues. As a routine matter, the company discloses to investors that the products carry some inherent risk that they will fail. The company records a quarterly reserve based on historical return rates to offset the estimated cost of product failures.
In manufacturing the GPU and MCP products, the semiconductor chip or "die" is mounted onto a substrate that ultimately is incorporated into a computer's motherboard. The die is connected to the substrate with "bumps" of solder. These bumps are also surrounded with a support material, referred to as underfill or paste.
According to the SCAC, NVIDIA observed problems with its GPUs and MCPs in late 2006.
Some units experienced cracks in the solder bumps when subjected to pressure testing. At that time, NVIDIA was using a "eutectic" solder-a material with relatively low lead content-and a eutectic 21 paste. As a result of that pressure-related problem, NVIDIA switched to a high-lead solder, which 22 was more malleable. It continued to use the same eutectic paste. Plaintiffs contend that the 23 mismatch in thermal properties between these two materials contributed to new problems.
After this switch, plaintiffs allege that NVIDIA's customers, including HP and Dell, began to complain to the company about problems with its chips. According to plaintiffs, HP engineers 26 observed physical damage to NVIDIA's parts, specifically cracking of the solder bumps connecting 27 the die to the substrate (the "material set" problem). By the summer of 2007, HP allegedly had 28 determined that the root cause of the GPU failures was associated with heat cycling. As a result, in November 2007, HP issued a software system update, or Basic Input/Output System (BIOS) update, 2 that was directed toward continuously operating the cooling fans on affected laptops.
On May 22, 2008, NVIDIA disclosed in its quarterly report that it had received customer claims from one of its OEMs for incremental costs due to notebook failures. The company stated 5 that it was unable to estimate the amount of cost that it might incur beyond its normal reserve for 6 product warranty accrual. At that time, it did not record any additional related costs for the quarter.
On July 2, 2008, NVIDIA announced the $150 million to $200 million charge. The stock dropped 31% from its high on July 2 to the opening price on July 3 (from $18.78 to $12.98). Based on these 9 events, plaintiffs initiated this suit for violation of the Securities and Exchange Act of 1934.
III. LEGAL STANDARD
Under Federal Rule of Civil Procedure 8(a)(2), a complaint must present "a short and plain statement of the claim" demonstrating that the plaintiff is entitled to relief. Fed. R. Civ. P. 8(a)(2).
If this standard is not met, the defendant may move to dismiss for failure to state a claim upon
which relief can be granted. Fed. R. Civ. P. 12(b)(6). Under Rule 12(b)(6), dismissal is appropriate 15 if the claimant either does not raise a cognizable legal theory or otherwise fails to allege sufficient facts to support a cognizable claim. Balistreri v. Pacifica Police Dep't, 901 F.2d 696, 699 (9th Cir. 1988). Thus, while a legally sufficient complaint does not require "detailed factual allegations," it 18 must contain more than "unadorned" assertions of harm or bare legal conclusions without factual 19 support. Ashcroft v. Iqbal, 129 S. Ct. 1937, 1949 (2009) (citing Bell Atlantic Corp. v. Twombly, 550 U.S. 544, 570 (2007)). Furthermore, Federal Rule of Civil Procedure 9(b) requires that "[in] 21 allegations of fraud or mistake, a party must state with particularity the circumstances constituting 22 fraud or mistake." To satisfy the rule, a plaintiff must plead the "who, what, where, when, and how" 23 of the alleged misconduct. Vess v. Ciba-Geigy Corp. USA, 317 F.3d 1097, 1106 (9th Cir. 2003). In 24 claims governed by the Private Securities Litigation Reform Act (PSLRA), plaintiffs must meet 25 additional pleading requirements beyond those imposed by the Federal Rules, as elucidated below.
In their first and second claims for relief, plaintiffs aver that NVIDIA and Huang violated section 10(b) of the Securities Exchange Act and Rule 10b-5 promulgated thereunder. Section 10(b) makes it unlawful for "any person . . . [t]o use or employ, in connection with the purchase or sale of 2 any security registered on a national securities exchange . . . any manipulative or deceptive device or 3 contrivance in contravention of such rules and regulations as the Commission may prescribe as 4 necessary or appropriate in the public interest or for the protection of investors." 15 U.S.C. § 78j(b). 5
Rule 10b-5 further provides: "It shall be unlawful for any person . . . [t]o engage in any act, practice, 6 or course of business which operates or would operate as a fraud or deceit upon any person, in 7 connection with the purchase or sale of any security." 17 C.F.R. § 240.10b-5(c). In their third 8 claim, plaintiffs assert that Huang violated section 20(a) of the Exchange Act, which makes certain
In order to maintain a section 20(a) claim, a plaintiff must sufficiently plead an underlying section 10(b) violation. See Zucco Partners, LLC v. Digimarc Corp., 552 F.3d 981, 990 (9th Cir. 2009)
A claim for violation of Rule 10b-5 includes five elements: "(1) a material misrepresentation or omission of fact, (2) scienter, (3) a connection with the purchase or sale of a security, (4) transaction and loss causation, and (5) economic loss." Id. (quoting In re Daou Sys., Inc. Secs. 15 plaintiffs must "specify each statement alleged to have been misleading, the reason or reasons why the statement is misleading, and, if an allegation regarding the statement or omission is made on 18 information and belief . . . state with particularity all facts on which that belief is formed." 15 U.S.C. 78u-4(b)(1). With respect to scienter, the PSLRA requires that a plaintiff must "state with 20 particularity facts giving rise to a strong inference that the defendant acted with the required state of 21 mind." 15 U.S.C. § 78u-4(b)(2). Under the law of this circuit, to state a claim for an alleged 22 violation of section 10(b), a complaint "must allege that the defendants made false or misleading 23 statements either intentionally or with deliberate recklessness." Daou, 411 F.3d 1006, 1015 (9th Cir. 2005) (citing In re Silicon Graphics Secs. Litig., 183 F.3d 970, 974 (9th Cir. 1999)).
"controlling" individuals also liable for violations of section 10(b) and its underlying regulations.*fn1
Litig., 411 F.3d 1006, 1014 (9th Cir. 2005)) (internal quotation marks omitted). Under the PSLRA,
Defendants move to dismiss the section 10(b) claims for failure to plead the elements of a material 2 misrepresentation or omission, scienter, and loss causation.
A. Material Misrepresentation or Omission
Plaintiffs allege that NVIDIA made a number of misleading statements because it knew, but did not disclose, that key customers including HP and Dell were complaining about problems with 6 the company's main products. In the securities context, an omitted fact is material if there is a 7 substantial likelihood that a reasonable investor would have viewed the disclosure as "significantly 8 alter[ing] the 'total mix' of information made available." Basic Inc. v. Levinson, 485 U.S. 224, 231-9 (1988) (internal quotation marks and citation omitted). As a general matter, a company is not 10 obligated to disclose publicly "any and all" material information. See Matrixx Initiatives, Inc. v.
Siracusano, 131 S. Ct. 1309, 1321 (2011). Disclosure is necessary, if under the circumstances, omitting the material information would render a company's statement misleading. See id. at 1321-22 (quoting 17 C.F.R. § 240.10b-5(b)). "Silence, absent a duty to disclose, is not misleading under Rule 10b-5." Basic, 485 U.S. at 239 n.17. With respect to its quarterly reports and annual report, 15 plaintiffs contend that silence was not an option, because NVIDIA was under an affirmative duty based on Item 303 of Regulation S-K to disclose the risk that the material set problem would lead to a substantial charge to the company.*fn2
In particular, plaintiffs identify sixteen passages that appear in documents NVIDIA publicly filed during the class period.*fn3 For instance, in a November 8, 2007 press release, the company 20 states: "[W]e believe this [our first billion dollar quarter] is just the beginning. . . . Our core 21 businesses are continuing to grow as the GPU becomes increasingly central to today's computing 22 experience in both the consumer and professional market segments." On March 21, 2008, the company filed its annual report, which included the following: "[W]e believe that close relationships 2 with OEMs, ODMs [original design manufacturers] and major system builders will allow us to 3 better anticipate and address customer needs with future generations of our product." On May 8, 2008, NVIDIA issued a press release reporting: "The growth of GPUs continues to outpace the PC 5 market. We shipped 42 percent more GPUs this quarter compared to the same period a year ago, 6 resulting in our best first quarter ever." Plaintiffs rely on these and other statements that omit 7 mentioning NVIDIA's ongoing investigation into complaints regarding its chips. Defendants argue 8 that the statements identified by plaintiffs constitute vague or generalized representations on which 9 no reasonable investor would rely. They also contend that various accused statements are forward 10 looking, accurate historical statements, risk disclosures, or Sarbanes-Oxley certifications that are non-actionable.
Ultimately, on May 22, 2008, in the quarterly report filed six weeks prior to NVIDIA's admission that it was taking a $150 million to $200 million charge, the company disclosed the potential defect:
During the first quarter fiscal year 2009, one of our customers asserted claims for incremental repair and replacement costs related to an alleged die/packaging material set defect in one of our notebook MCP products. This product was included in a significant number of the customer's notebook products that have been sold to end users, and has also been shipped to other of our customers in significant quantities. We are evaluating the potential scope of this situation, including the nature and cause of the alleged defect and the merits of the customer's claim, and to what extent the alleged defect might occur with other of our products. We are currently unable to estimate the amount of costs that may be incurred by us beyond the normal product warranty accrual that we have taken related to this claim and the alleged defect and, therefore, we have not recorded any additional related costs or a liability in our Condensed Consolidated Financial statements as of, and for the three months ended, April 27, 2008.
With respect to this disclosure, defendants admit that the statement is of the type that could be 25 actionable, but argue that the falsity of the statement is insufficiently pleaded.
Although plaintiffs do not need to prove at the pleading stage that this statement is false or misleading, pursuant to Rule 9(b) and the PSLRA, they must plead with enough particularity to 28 make the falsity or the misleading character of the statement plausible. In the above statement, NVIDIA reports that it had shipped the allegedly defective MCP to other customers in significant 2 quantities. It also states that it was evaluating the extent to which the defect "might occur" in other 3 products. These statements do not expressly reveal, as plaintiffs contend was true, that more than 4 one customer had complained of defective chips or that the company had received complaints 5 regarding its GPU, as well as the MCP. Plaintiffs rely, among other allegations, on statements made 6 by Michael Hara, Vice President of Investor Relations at NVIDIA, at a conference in September 7 2008. In particular, Hara stated that the company had "been working on this problem with the 8 customers for well over a year, going all the way back to August of ." Plaintiffs also point to 9 the observations of confidential witness (CW) 7. He/she worked at Dell in 2007 and was involved 10 in testing prototypes. According to CW 7, Dell experienced problems with NVIDIA's GPU in early 2007. CW 7 claims he/she communicated directly with NVIDIA about the issue.
Taking the allegations in the SCAC to be true, NVIDIA was already working with Dell on related issues and knew that the alleged defect also occurred in the GPU at the time it released its May 22, 2008 quarterly report. Even if, in good faith, NVIDIA did not believe its potential liability 15 would exceed its normal reserve, it is substantially likely that a reasonable investor would consider material the fact that the alleged defect was observed across customers and products. Moreover, its omission, under the circumstances, rendered NVIDIA's statement that the defect "might occur" in 18 other products misleading. Thus, plaintiffs have adequately pleaded at least one material 19 misrepresentation and the SCAC is not subject to dismissal for failure to identify any statement that 20 satisfies this element.
B. Scienter 22
In Silicon Graphics, the Ninth Circuit considered the "deliberate recklessness" standard and explained that "it [is] a form of intentional or knowing misconduct." 183 F.3d at 976. Thus, a 24 strong inference of scienter requires factual allegations supporting more than that the defendant 25 acted with "mere recklessness" or possessed the motive and opportunity to commit fraud. See Zucco Partners, 552 F.3d at 991. Instead, a plaintiff must plead "a highly unreasonable omission, 27 involving not merely simple, or even inexcusable negligence, but an extreme departure from the 28 standards of ordinary care, and which presents a danger of misleading buyers or sellers that is either known to the defendant or is so obvious that the actor must have been aware of it." Id. (internal 2 quotation marks and citations omitted). Furthermore, the Supreme Court clarified that a "strong 3 inference" of the required intent means that a complaint will survive a motion to dismiss "only if a 4 reasonable person would deem the inference of scienter cogent and at least as compelling as any 5 opposing inference one could draw from the facts alleged." Tellabs, Inc. v. Makor Issues & Rights, Ltd., 127 S. Ct. 2499, 2510 (2007).
In this case, the issue is whether NVIDIA and Huang made material misrepresentations or omissions regarding the financial impact of the material set problem with the intent to mislead, or 9 with deliberate recklessness of misleading, investors about the value of the company's stock.
Plaintiffs rely essentially on five categories of allegations: (1) information obtained from persons acting as CWs; (2) information from industry experts or commentators; (3) NVIDIA's conduct directed toward addressing the product defect; (4) other lawsuits; and (5) Huang's sale of NVIDIA stock and the departures of NVIDIA employees. In addition to arguing that the factual allegations are sufficient to establish defendants possessed the requisite scienter, plaintiffs also contend that it 15 may be inferred from the scope of the problem and the importance of the products to NVIDIA's business. See South Ferry LP v. Killinger, 542 F.3d 776, 785 (9th Cir. 2008) (explaining that "management's role in a company may be relevant and help to satisfy the PSLRA scienter 18 requirement"). Finally, plaintiffs rely on the collective scienter doctrine in arguing that scienter is 19 adequately pleaded. See Glazer Capital Mgmt., LP v. Magistri, 549 F.3d 736, 744 (9th Cir. 2008) (recognizing that "in certain circumstances, some form of collective scienter pleading might be 21 appropriate"). Under the law of the Ninth Circuit, the court engages in a "dual inquiry" into 22 whether plaintiffs' allegations in turn or collectively satisfy the element of scienter. See Zucco Partners, 552 F.3d at 992 (citing Tellabs, 127 S. Ct. 2499). Specifically, the court first "will 24 determine whether any of the plaintiff[s'] allegations, standing alone, are sufficient to create a strong 25 inference of scienter; second, if no individual allegations are sufficient, [it] will conduct a 'holistic' 26 review of the same allegations to determine whether the insufficient allegations combine to create a 27 strong inference of intentional conduct or deliberate recklessness." Id.
problem. The complaint uses a number of CWs in an attempt to establish that NVIDIA knew of the 4 severity of the problem more than a year prior to its public disclosure in July 2008. In order to rely 5 on CWs, a complaint "must pass two hurdles to satisfy the PSLRA pleading requirements." Zucco
Partners, 552 F.3d at 995. First, a CW must be described with sufficient particularity to establish 7 his or her reliability and personal knowledge. Id. Second, the statements supplied by CWs "with 8 sufficient reliability and personal knowledge must themselves be indicative of scienter." Id. 9
The witnesses are particularly described in that their titles and the time they held their respective positions are identified. The main issue with plaintiffs' CWs, however, is that their statements are not "indicative of scienter." Only six of the CWs worked at NVIDIA during the class period and none of these worked on the defective chip. Moreover, none of the seventeen CWs report any contact with Huang or other senior management on the scope of the material set defect. Examples 15 of the general nature of most of the CWs' allegations include the observations of CW 10, a former
NVIDIA IT manager who left in March 2007. He/she states, "I know I had all sorts of overheating issues that required the replacement of graphics systems . . . No one from the chain of command told 18 me formally that it was our product and we were at fault or anything like that, but I can tell you that
I certainly noticed a pattern of failures that were related to overheating in the GPU." CW 11, a 20 senior engineering manager who left in March 2006, reports that NVIDIA operated with "a failure 21 to recognize real problems. They just won't admit it." CW 12, a sales director at a company that 22 sold components to NVIDIA and who left in 2004, claims that NVIDIA was notorious for blaming 23 other entities for product related problems: "A lot of times it's easy for them to just blame [a 24 customer's component] and not the graphics chips. So their first area of default was always to say, 25
'This is your problem. You fix it.' Some of the time they were right in saying that and some of the time 26 they were wrong, but their default was always to say, 'This is not our problem. This is your problem.'"
CW 1 is the only witness who specifically discusses testing any of the components at issue
in this suit. He/she was a former manager with HP's product engineering group and was part of the
1. Confidential Witnesses
Plaintiffs contend that scienter may be inferred from NVIDIA's response to the material set
In this case, plaintiffs' complaint incorporates statements from seventeen different CWs. team that did the root cause analysis of the defective chip. According to CW1, communications 2 between HP and NVIDIA began in 2006, with HP asking NVIDIA to conduct failure analysis. HP 3 routinely shared its data with NVIDIA, but the latter continued to deny that it was at fault. CW 1 4 claims that communications were conducted at a high level, stating "these were executive level 5 guys." CW 1 reports that "NVIDIA was on-site with HP, into the factories, having conversations at 6 the director level, toward the end of the first quarter of 2007." By early 2007, HP "had 7 overwhelming data demonstrating root cause." According to CW 1, the particular thermal profile 8 that produced cracking in the solder bumps of NVIDIA's chips had been identified as the root cause 9 by HP in early 2007 and reproduced by NVIDIA by the middle of that year. 10
it merely reiterates that NVIDIA knew of a problem by early 2007. CW 1 does not provide details about specific communications or statements made by NVIDIA. Thus he/she cannot know what conclusions the employees at NVIDIA made in 2007 regarding the source or scope of the problem or the company's potential liability. In short, none of the statements provided by the CWs support a 15 strong inference that Huang or NVIDIA omitted discussing the material set problem in order to mislead investors.
2. Information from Other Industry Participants
Plaintiffs also rely on statements from other individuals to support their contention that
NVIDIA knew of the risk of substantial exposure from the defective chips. In particular, plaintiffs 20 refer to statements made in a declaration by Richard Hunt Hodge as part of a lawsuit in which HP is 21 separately involved, over the loss of wireless capability in affected laptops. Hodge was the director 22 of engineering and quality for HP's notebook division from 2003 to July 2009. Plaintiffs select 23 portions of this declaration to support a conclusion that HP and NVIDIA understood the material set 24 problem by November 2007. Hodge states, "Eutectic soldering and eutectic soldering paste react to 25 temperature fluctuation in harmony and, therefore, eutectic based C-51 chips [from NVIDIA] are 26 not susceptible to the heat cycling issue that was identified in late October 2007." As Hodge 27 explains, NVIDIA transitioned from high-lead solder in the problem chips back to eutectic solder in 28 or around July 2008.
While the information that CW 1 provides is more specific than that of the remaining CWs,
Plaintiffs also refer to an email Hodge sent in May 9, 2007 in which he states that HP was working on development of a new laptop. "The first few months of production have a very high 3 failure rate in the field (about 130 k units). Most of the failures were traced to an issue identified 4 and corrected during early warning . . . There are other issues on the product that involve NVIDIA 5 and a change of the solder composition at the famous L1 joint between the die and the substrate. In 6 this case, NVIDIA damaged some of the ICs during their factory test. Part of the solution however 7 was to move away from Eutectic solder." Plaintiffs argue that this email reflects the fact that, after 8 observing cracking during pressure testing, NVIDIA made a hasty change from eutectic solder to 9 high-lead solder and therefore knew or deliberately disregarded the fact that cracks in the solder*fn4 bumps were due to its own 2006 manufacturing change. While plaintiffs sufficiently allege that HP and NVIDIA knew of the solder bump cracks and were working on a fix, their allegations do not provide a basis to infer that Huang or others appreciated the severity of the problem or recognized during the class period that any extraordinary liability was likely to result.
For the same reason, additional statements on which plaintiffs rely demonstrate that NVIDIA knew about the defect, but do not contribute to an inference that defendants possessed the intent to defraud. For instance, plaintiffs point to the remarks from Hara in September 2008, in which he admits that NVIDIA knew about the chip defects and had been working with their customers for 18 well over a year. He reports the company had "already put many corrective procedures in place." 19
Plaintiffs note that at the hearing on defendants' motion to dismiss the CAC, defendants' counsel 20 admitted that NVIDIA was talking to "at least two of their most important OEMs" in August 2007. 21
Plaintiffs also obtained a technology consultant, John Rigg, with nearly twenty-five years of experience in graphics architecture and integrated circuits. Plaintiffs offer his statements to buttress 23 their contention that NVIDIA would not have needed a long time to understand the origin and scope of the defects. Rigg suggests that failures of the GPU are easily identifiable, because they result in 2 visual problems with a computer's display. Rigg also claims that post-manufacture "fixes" to 3 reduce thermal stresses in laptops are not feasible solutions. He states that changing the speed of the 4 laptop's fan will likely only delay discovery of the problem. 5
Plaintiffs also rely on comments by Charlie Demergian, a reporter with twenty-five years of experience who covered the NVIDIA chip defects and purportedly spoke to "dozens" of people 7 about the problem. According to Demergian, an HP engineer stated that one of NVIDIA's chips 8 had failure rates as high as 50%. By the summer of 2007, Demergian maintains that HP employees 9 knew NVIDIA's GPUs were having problems associated with heat cycling.
Plaintiffs additionally point to various articles about the problems at NVIDIA, which they suggest provide "context and information regarding industry-known principles." Specifically, plaintiffs contend that defendants deliberately disregarded the obvious fact that their change from high lead to eutectic solder was the cause of the chip defect. In August 25, 2008, TG Daily published an article noting: "What supports the theory that a high-lead solder bump in fact is at fault 15 is the fact that NVIDIA ordered an immediate switch to use eutectic solders instead of high-lead versions in the last week of July ." Plaintiffs also suggest that NVIDIA must have known the severity of the problem for several months prior to its return to high-lead solder in July 2008. In September 10, 2008, another TG Daily article on NVIDIA stated: "Transition to a different solder 19 bump material requires substantial engineering effort, testing and qualification and that may have 20 been the case here." Finally, on September 28, 2008, TG Daily reported results of a scientific study 21 completed at UCLA comparing the combination of high-lead solder and eutectic paste (used by NVIDIA in its problem chips) with a material set incorporating eutectic solder and eutectic paste.
The article stated that "eutectic solder joints may have a longer life than the high-lead joints."
In short, plaintiffs have raised a number of statements from sources outside NVIDIA to support their contention that the company must have known the scope of the problem much earlier 26 than they disclosed. While these allegations suggest that NVIDIA might have acted differently in 27 hindsight, the company's response to the defective products does not represent "an extreme 28 departure from the standards of ordinary care." Thus, these allegations from industry experts and participants, together with plaintiffs' other averments, do not support the conclusion that Huang or NVIDIA acted with fraudulent intent or deliberate recklessness.
3. NVIDIA's Response to the Alleged Defect
Plaintiffs argue that NVIDI was motivated to delay its acceptance of responsibility for the defect until the affected chips were near the end of their product life. HP issued its BIOS update to 6 reduce temperature cycling in November 2007 and Dell also released a similar update in February 2008. According to plaintiffs, NVIDIA participated in preparing these BIOS updates. Thus, 8 plaintiffs allege that NVIDIA had knowledge of the source of the problem and acted with disregard 9 of its responsibility. A contrary inference, however, is that both NVIDIA and its customers believed 10 that the software updates could reduce the product failures. Thus, NVIDIA's participation in the BIOS updates, again in combination with the other information advanced by plaintiffs, does not warrant the inference of an intent to delay recognizing financial loss from the defects.
Along similar lines, plaintiffs also submit that steps defendants took to rectify the problem are indicative of their intent to deceive. In NVIDIA's July 2, 2008 press release, at the same time 15 the company announced it was taking the charge against its cost of revenue, it also reported that "[a]ll newly manufactured products and all products currently shipping in volume have a different and more robust material set." Plaintiffs thus suggest that NVIDIA took "extraordinary, secret 18 steps" for at least eleven months prior to informing investors of the material set problem. The 19 switch to high-lead solder in products going forward does suggest that the company knew its 20 materials at least contributed to defects observed by its customers. This response, however, does 21 not create or contribute to a strong inference that NVIDIA knew the scope of its potential financial 22 liability at the time it was redesigning the chips and hoped to hide that fact. 23
4. Other Lawsuits
Plaintiffs also rely on allegations from other lawsuits to support scienter. Two of NVIDIA's insurers filed declaratory judgment actions against the company. Those insurers both argue that the 26 product damage is not covered by their respective policies because NVIDIA knew of the damage 27 before the effective dates of coverage. The fact that other third parties, with money at stake, would 28 join in accusing NVIDIA of not disclosing information does not add support to a showing of scienter under the standards of the PSLRA. The same goes for a consumer class action settled by NVIDIA in August 2010. Whether the company was ultimately liable for defective chips, again, 3 does not reflect on Huang or NVIDIA's intent to deceive investors. Finally, HP was sued on November 14, 2007, only a few days into the class period, by consumers who were experiencing 5 problems related to the defective NVIDIA chips. Plaintiffs contend that NVIDIA knew of the suit 6 and that HP would seek indemnification. The apparent source of this information is the complaint 7 from one of the insurance actions. Thus, the SCAC does not identify with particularity who at NVIDIA knew HP would seek indemnification, or when or how the company allegedly became 9 aware of that fact.
Plaintiffs also raise Huang's sale of NVIDIA stock as evidence of scienter. In particular,Huang allegedly exercised options for NVIDIA common stock and sold 186,000 shares in March 2008 that netted three million dollars. These sales were made pursuant to a 10b5-1 plan, which Huang adopted in August 2007. While this date is prior to the class period, plaintiffs attempt to 15 suggest that Huang was in possession of material nonpublic information about problems with the company's core products at that time. The prior Order rejected Huang's stock sales as evidence of scienter, as he was a net acquirer of stock during the class period. See NVIDIA, 2010 U.S. Dist. LEXIS 114230, at *41. Huang also sold more stock in each of the three quarters prior to the 19 beginning of the class period. Id. Thus, his trading in March 2008 is neither unusual nor 20 suspicious.
they mention are indicative of scienter. Id. at *42-43. On March 27, 2007, NVIDIA's CFO Marvin 2008, NVIDIA's senior director and head of internal audit, Alex Garcia left the company. David Kirk, NVIDIA's chief scientist left his position in January 2009, but remained with the company as 26 a Fellow. As noted previously, these bare facts without any specifics regarding the circumstances of 27 the departures add nothing to the inference of scienter.
5. Huang's Stock Trades and Employee Departures
The Court also previously rejected plaintiffs' suggestion that the few employee departures
Burkett stepped down from that position, but remained as a senior advisor to the company. In June statements intentionally or with deliberate recklessness, they claim that scienter can be attributed to Huang because GPUs and MCPs were indisputably NVIDIA's core business. Such averments may 5 be considered "along with other allegations" in assessing the totality of the circumstances. South 6
6. Core Operations and Collective Scienter
Even if plaintiffs are unable to plead with particularity that defendants made misleading
Ferry, 542 F.3d at 785. In addition, the Ninth Circuit recognizes that in narrow situations, "bare 7 allegations of falsely reported information" may be probative of scienter. Zucco Partners, 552 F.3d 8 at 1000. In particular, the Court describes two exceptions to the general rule that knowledge of 9 certain facts may not be attributed to the company and its officers merely because the information is 10 critical to core operations. Id. (citation omitted). In one exception, general allegations about the
importance of the corporate information can create a strong inference of scienter if the plaintiffs
include "detailed and specific allegations about management's exposure to factual information
within the company." Id. (quoting South Ferry, 542 F.3d at 785) (internal quotation marks omitted).
The second exception applies where the company reports information of which the falsity would be omitted).
"patently obvious" to senior management, based on operation of the company. Id. at 1001 (citations
The prior Order dismissing the CAC rejected plaintiffs' argument that scienter could be
inferred from the importance of the GPU and MCP chips to NVIDIA's business. See NVIDIA, 2010
NVIDIA's knowledge of the material set problems, these do not warrant a different conclusion with 21 respect to scienter. Plaintiffs allege that "it would strain credulity to suggest" that Huang did not 22 know about the material set problem prior to the class period. Knowledge of the problem, however, 23 is insufficient to infer that he acted with the intent to defraud or with deliberate recklessness in not 24 reporting the issue publicly. The SCAC includes no specific allegations that anyone, let alone 25
Huang, concluded that the likely financial impact of the material set problem would exceed the 26 company's normal reserve. Without such a determination, the falsity of NVIDIA's statements 27 cannot be considered "patently obvious." As a result, the requisite scienter cannot be inferred even 28 if, as plaintiffs contend, Huang would have had intimate knowledge of the defective products.
U.S. Dist. LEXIS 114230, at *38-39. While the SCAC adds additional averments with respect to
Plaintiffs also argue that scienter with respect to the company is demonstrated "collectively,"
2 even if the SCAC fails to plead sufficiently that any particular employee acted with intent or 3 reckless disregard. The Ninth Circuit has acknowledged that "there could be circumstances in 4 which a company's public statements were so important and so dramatically false that they would 5 create a strong inference that at least some corporate officials knew of the falsity upon publication." 6
Glazer Capital Mgmt., LP, 549 F.3d at 744. Again, plaintiffs' averments only suggest that NVIDIA 7 had received complaints from important customers and was investigating the material set problem 8 during the class period. Even if accepted as true, these facts do not create a strong inference that 9
NVIDIA was aware of an unusually large financial liability and deliberately hid that fact from 10 investors. In short, plaintiffs' averments add no support for the inference that NVIDIA's public statements which failed to mention the defective products and potential financial risk were "dramatically false."
As noted above, after assessing each category of allegations raised by plaintiffs, as well as their reliance on the core operations and collective scienter doctrines, the SCAC must be viewed "holistically" to determine if a strong inference arises of intent or deliberate recklessness on the part of Huang or NVIDIA. See Tellabs, 551 U.S. at 326. Altogether, plaintiffs' allegations suggest no more than that NVIDIA knew of the material set problem, discussed it with customers, worked on 18 implementing a software fix, and developed next generation chips that would not exhibit the same 19 defect. While NVIDIA omitted discussion of the alleged defect from its public statements until the May 2008 quarterly report, a more reasonable, competing inference is that the company was 21 investigating the scope of the issue. In fact, plaintiffs raise some allegations in order to suggest that, 22 at other times in the past, NVIDIA sought to avoid taking early responsibility for potential 23 problems. Thus, even accepting plaintiffs' assertion that the partial disclosure of the problem in May 2008 was misleading, NVIDIA still may have underappreciated the likelihood of incurring 25 extraordinary financial liability. Such behavior, at worst, reflects recklessness in the ordinary sense 26 of the word with respect to customer relations and potential financial risk. It does not, however, add 27 support for an inference at the pleading stage that NVIDIA omitted discussion of the allegedly 28 defective products in order to mislead investors that is "at least as compelling" as "plausible, nonculpable explanations." See Tellabs, 551 U.S. at 323-24. As plaintiffs fail sufficiently to plead 2 scienter as to Huang or NVIDIA, their claims for violation of section 10(b) and Rule 10b-5 must be 3 dismissed.*fn5 Moreover, as plaintiffs fail to plead a section 10(b) violation, their section 20(a) claim 4 against Huang must similarly be dismissed. 5
Despite prior opportunity to amend, plaintiffs fail to plead the element of scienter that is necessary to state a claim pursuant to section 10(b), Rule 10b-5, or section 20(a). Accordingly, the SCAC must be dismissed without leave to amend.
IT IS SO ORDERED.