The economic impacts of the COVID-19 pandemic are manifold. The precipitous climb in COVID-19 cases in the United States and around the world in the spring of 2020 led to massive layoffs and the hollowing out of densely-populated urban cores. The federal government responded with a massive stimulus program intended to boost an economy that had lost 23 million jobs between January and April 2020. And as jobs bounced back, the tightening of the labor market, combined with a sharp increase in the costs of many goods and services, disruption to supply chains, and other factors, led to a spike in inflation beginning in mid-2021. To combat inflation, which has yet to abate fully, the Federal Reserve (“the Fed”) began increasing interest rates to levels not seen since the 1980s.

Fortunes have been made and lost during this period. For example, on December 16, 2020, Upstart Holdings, Inc. (“Upstart”), a consumer lending startup completed its initial public offering (“IPO”) at $20.00. In less than a year, its stock rose to a peak of $401.49; its early investors- including the investment firm, Third Point LLC (“Third Point”)-executives, and shareholders profited immensely by selling high. But as the pandemic stimulus funding came to an end and interests rates began to rise, increasing the costs of borrowing and hampering demand for loans, Upstart's business flagged. By November 8, 2022, its stock had dropped to $19.04 per share; investors who had bought in during the glory days in the fall of 2021 were left with massive losses.

Plaintiffs, Universal-Investment-Gesellschaft mbHaft, a German investment firm, Emmanuel Sebag, and Kathy Brooks, now bring a putative securities fraud class action, alleging violations of the Securities Exchange Act of 1934 (“the Exchange Act”), 15 U.S.C. §§ 77a-78pp, and the regulations promulgated thereunder, against Defendants Upstart; David J. Girouard, cofounder and chief executive officer (“CEO”) of Upstart and chair of its board; Sanjay Datta, the chief financial officer (“CFO”) of Upstart; Paul Gu, co-founder and Senior Vice President of Product and Data Science of Upstart and a member of its board of directors; Anna Counselman, co-founder and Senior Vice President of People and Operations at Upstart; Third Point; Third Point Ventures LLC, the Menlo Park-based venture capital arm of Third Point; Daniel S. Loeb, the CEO of Third Point; and Robert Schwartz, the Managing Partner of Third Point Ventures and a former member of the Upstart board. The Upstart Defendants and the Third Point Defendants have now moved to dismiss Plaintiffs' Consolidated Amended Complaint (hereinafter, “the Complaint”) (ECF No. 45). Plaintiffs oppose, and additionally seek to strike a portion of the Third Point Defendants' reply brief.

The Upstart Defendants consist of Upstart, Girouard, Datta, Gu, Counselman and Schwartz. Plaintiffs dispute the inclusion of Schwartz in the Upstart Defendants, arguing that he is instead a Third Point Defendant. This Court rejects that characterization, for the reasons described infra.

The Third Point Defendants consist of Third Point, Third Point Ventures, and Loeb.

For the reasons set forth more fully below, this Court Grants in part and denies in part the Upstart Defendants' Motion to Dismiss (ECF No. 58) and GRANTS the Third Point Defendants' Motion to Dismiss (ECF No. 60). Additionally, Plaintiffs' Motion to Strike, or, in the Alternative, for Leave to File a Sur-Reply (ECF No. 65) (hereinafter “Motion to Strike”) (ECF No. 65) is Granted in part and denied in part.

II. BACKGROUND

A. Disrupting the FICO Standard

Upstart was founded in 2012 by Girouard, Gu, and Counselman with the goal of disrupting the unsecured personal loan market. (See Compl. ¶ 58, ECF No. 45). Traditionally, lenders have evaluated a prospective borrower's creditworthiness by looking at her “FICO” score, which was developed by the Fair Isaac Corporation in 1989; lenders have also relied on the FICO score to determine the interest rate to charge on loans. Some broad contours of what goes into a FICO score are known, though the precise details are not: the score is intended to incorporate an individual's payment history, amount owed, length of credit history, new credit accounts, and types of credit used. See Jason Steele, How FICO Scores Are Calculated, INVESTOPEDIA (Apr. 30, 2022), https://perma.cc/52PK-HNXE. The Fair Isaac Corporation has created software used by the three major credit bureaus, Equifax, Experian, and TransUnion, who calculate credit scores and run credit reports.

Girouard, Gu, and Counselman posited that the traditional FICO score had some serious limitations-that it considered only a subset of the relevant variables and therefore failed to identify accurately a prospective borrower's credit risk or price loans accurately. (See Compl. ¶ 58, ECF No. 45). An artificial intelligence (“AI”) underwriting model could overcome these limitations by taking into account a far greater volume of variables while also adapting on the fly to changing conditions. (See id. ¶¶ 59, 126). And the model could divine a prospective borrower's financial potential, and not rely on their (potentially outdated) financial history. (Id. ¶ 59).

That's where Upstart comes into play, with its “cloud-based AI lending platform.” (Id. ¶ 22). Upstart's platform matches would-be borrowers of unsecured personal and auto loans and lenders, and charges a fee for facilitating the connection. The platform is powered by an AI underwriting model, which looks at more than 1,600 variables and was trained on more than nine (9) million repayment events. (Id. ¶ 59). Prospective borrowers, typically seeking a personal loan between $1,000 to $50,000, arrive at the Upstart platform either directly through Upstart's website (Upstart.com) or through Upstart's bank partners. (Upstart IPO S-1 at 8, ECF No. 58-4). To receive a loan, they apply on the Upstart platform, the AI model evaluates the borrower and the request, and the model then decides whether to issue the loan and at what rate. (See Compl. ¶ 3, ECF No. 45). If the request is approved, one of Upstart's banking partners originates the loan. Normally, the banking partner would then hold on to the loan. But Upstart also had an agreement with its partners, pursuant to which the banks were not required to retain the loans they originated: they could choose to keep only those loans that they wanted, and sell any undesired loans back to Upstart. (See id. ¶ 65). Upstart would then turn around and sell the loans to institutional investors, as whole loans, pass-through certificates, or re-packaged as asset-backed securities (“ABSs”). (See id. ¶¶ 61, 66, 73).

In other words, Upstart's business model is to “partner[] with traditional banks and . . . not [to] use its own balance sheet.” (Id. ¶ 63). As Jim Cramer, host of CNBC's “Mad Money” program summarized: “Upstart uses data and its AI technology to facilitate [loans]....They don't lend money themselves. They're a marketplace that matches borrowers and lenders. They take a referral fee.” (Id.) (alteration in original) (emphasis omitted). The referral fee is paid by the banks in exchange for access to Upstart's AI lending platform, and its loan application data, underwriting, fraud detection, and documentation delivery services. (Id. ¶ 62).

In theory, this setup provides substantial value to everyone involved. Upstart claimed that its AI underwriting model would generate “better” loans for Upstart's bank partners than traditional FICO-based models-with “higher approval rates and lower interest rates at the same loss rate.” (Id. ¶ 126). It would also insulate Upstart from the credit risk associated with the loans themselves, since Upstart would simply facilitate lending between borrowers and its banking partners, without holding any long-terms loans on its balance sheet. (Id. ¶¶ 3, 60, 63).

B. Upstart's Stock Skyrockets Post-IPO

In 2015, Upstart closed a $35 million Series C fundraising round led by Third Point, at a valuation of $145 million. (See id. ¶ 67). Upstart continued to grow and raise more capital in subsequent funding rounds; Third Point continued to invest. (See id.). In total, Third Point committed $66 million to Upstart prior to the IPO. (Id.). And it was not a hands-off investor. Rather, Plaintiffs allege that Third Point actively advised Upstart on various aspects of corporate strategy, including its organizational development, staffing, and business plans. (See id. ¶ 68). Schwartz, the Managing Partner of Third Point Ventures, took a seat on Upstart's board of directors. (Id. ¶¶ 69-70). Third Point also contributed in other ways, purchasing loans generated by the Upstart platform. (Id. ¶¶ 72-73).

Despite the immediate economic downturn caused by the onset of the COVID-19 pandemic in March and April 2020, the second half of the year was rife with major technology startup IPOs. Companies like DoorDash and Airbnb went public in early December; Upstart soon followed suit. (See id. ¶ 83). Almost all of Upstart's revenue, at this point, was generated through the fees it was charging its banking partners for access to the platform, loan referrals, and servicing the loans. (See id. ¶ 64). It held very few loans: approximately 22% of the loans generated by the Upstart platform were funded (i.e., held) by Upstart's banking partners, and 76% by credit investors. (Id.). In other words, Upstart held on to only 2% of the loans generated via its platform, despite the buyback arrangement it had with its banking partners. Demand for its loans from credit investors remained robust, so it could move any repurchased loans off its books quickly. (See id. ¶¶ 61, 66, 73). Revenue generated from interest payments on the loans it retained represented a negligible portion of its overall revenue. (Id.).

Upstart debuted on the Nasdaq exchange on December 16, 2020, at $20.00 per share under the symbol “UPST.” (Id. ¶ 83). Third Point immediately purchased 1.2 million additional shares of Upstart stock. (Id.). Analysts were bullish on Upstart stock from the beginning, as they were impressed with Upstart and its platform-especially the “best-in-class” AI underwriting model and the low rate of loans held by Upstart. (See id. ¶¶ 85-86). These attributes suggested that Upstart would be able to continue keeping risk low (by virtue of having few volatile liabilities, like personal loans), while maintaining strong fee revenue.

External factors also contributed to Upstart's early post-IPO success. In response to the effects of the pandemic on the economy, the federal government pulled a series of levers to stimulate the economy, including interest rate cuts, monetary benefits (including direct cash payments, subsidies, and tax credits) to individuals, loans to businesses, and a moratorium on mortgage foreclosures and loan repayments. (See id. ¶¶ 74-81). Because of the stimulus, the economy was “awash in money”: individuals had extra cash in hand, were not obligated to service student loan payments immediately, and enjoyed easy access to credit.

Immediately after the passage of the American Rescue Plan Act (“ARPA”) in early April 2021, which consisted of a $1.9 trillion stimulus package, Upstart announced a secondary public offering (“SPO”), i.e., the sale of further shares of the company on the public market. (Id. ¶ 92). This time, Upstart marked its share price at $120.00 per share. (Id.). Upstart, its executives, and directors once again touted the strength of the Upstart business model, platform, and AI model, reiterating in both public disclosures filed with the Securities and Exchange Commission (“SEC”) and its first public earnings call that the AI underwriting model was nimble and “able to respond to the sort of macroeconomic changes on really a dime,” and that Upstart carried very little loan risk because demand for Upstart-generated loans from banks and investors remained strong. (Id. ¶ 94; see id. ¶¶ 91, 93-96).

Upstart's stock price continued to climb after the SPO, reaching a high of $401.49 per share on October 15, 2021. (See id. ¶ 97). The continued optimism was driven at least in part by the lack of credit exposure that Upstart continued to maintain and by the outstanding reputation of the AI model. (See id. ¶¶ 98-99). The Upstart Defendants continued to tell investors, through public filings and statements, that both remained true, even as the stimulus programs came to a close and interest rates began to rise. It was also driven by impressive increases in Upstart's revenue, rising from $86.7 million in the fourth quarter of 2020, to $194 million for the second quarter of 2021, $228 million for the third quarter, and $305 million for the final quarter of the year.

Behind the scenes, however, Upstart executives and insiders had begun to sell their stock. The executives had been subject to a 181-day lock-up period after the IPO. (See IPO S-1 at 80, 210, ECF No. 58-4). Upon the expiration of the lock-up period, Girouard, Gu, and Counselman all sold substantial amounts of their Upstart stock. (See Compl. ¶ 275, ECF No. 45). Counselman sold 608,355 shares on August 19, 2021, raking in over $122 million in proceeds. (Id.). Girouard sold approximately 90,000-140,000 shares at the beginning of each month from September 2021 to May 2022; in total, he sold approximately one (1) million shares for nearly $204 million. (Id.). And Gu also sold shares on a regular schedule: 155,000 shares in the middle of the month in August, September, and October 2021, and smaller chunks in November 2021 and the March 2022. (Id.). His stock sales generated $143.6 million. (Id.).

Third Point also exited its Upstart position around the same time. In total, it had acquired nearly 20% control of Upstart through its early investments and the IPO. It sold around one (1) million shares over three days in August 24, then sold the remaining 12 million shares it owned between November 15 and December 20, 2021. (Id.). Third Point earned over $2.2 billion from fully liquidating its position in Upstart. (Id.).

C. The Stock Plummets as Interest Rates Rise

In retrospect, the fall of 2021-when Girouard, Gu, and Counselman began selling their shares in Upstart stock, and Third Point fully sold its position-was when Upstart's stock peaked. On November 9, 2021, Upstart announced its financial results for the third quarter of 2021 (“Q3 2021”). While total revenue and total fee revenue had both risen dramatically, so too had the amount of loans that Upstart retained on its balance sheet. (See id. ¶ 298). Upstart's stock price fell by over 18% in the wake of the announcement, from $313.71 at closing on November 9, 2021, to $256.59 the following day. (Id. ¶ 299). Girouard explained on the earnings call that followed that this increase was not a concern, because Upstart's bank partners continued to have faith in the AI underwriting model and demand for the loans remained strong (despite the fact that Upstart was holding on to more loans). (See id. ¶ 104). The bulk of Third Point's stock sales occurred in the month following the release of Q3 financial results. (See id. ¶¶ 106-108).

In the months that followed, Upstart's financial results continued to worsen as loans on the balance sheet continued to grow, from $130 million at the end of Q3 2021 to $252 million at the end of Q4 2021, and then to $604 million at the end of Q1 2022. (Id. ¶¶ 108, 112). During the February 2022 earnings call, Upstart executives maintained that the increase in loans held during Q4 2021 was not a cause for concern, that the loans were priced correctly, and that the AI underwriting model had no problem accounting for rapidly increasing interest rates. (See id. ¶¶ 110-111). Datta explained the increase in loans as a “temporary” issue, associated entirely with research and development for a planned expansion into the automobile lending sector. (Id. ¶ 109). The Upstart executives also waved away concerns about rising loan default rates, which had been abnormally low as the stimulus package buoyed the economy but had begun to creep up to “normal,” pre-stimulus levels. (See id. ¶ 110).

It turned out that there was in fact cause for concern. After Upstart announced the massive increase in loans held during Q1 2022, Datta explained on an earnings call that Upstart had started to use its balance sheet as a “funding buffer” to hold on to loans that it could not immediately offload. (See id. ¶ 112). That is, Upstart was being forced to retain loans that had been sold back to it by its banking partners and that it was unable to repackage and sell to investors. (Id. ¶ 114). This, in turn, exposed the company to interest and credit risk.

The market responded negatively in the wake of the earnings report and call. The price of Upstart stock dropped a further 56%, and analysts suggested that Upstart's value proposition as a marketplace lender-i.e., as a facilitator between borrowers and lenders-was untenable if it could not pass on the loans generated by its platform to external funding partners. (See id. ¶¶ 115-117). The stock price continued to drop after Upstart cut its earnings guidance for Q2 2022 on July 7, 2022, and Girouard admitted that Upstart was “funding-constrained” vis-a-vis the loans, and had converted some loan on its balance sheet into cash at a loss. (See id. ¶¶ 118-119).

On November 8, 2022, Upstart announced a 31% decrease in revenue for Q3 2022 compared to the prior year, and a net loss of $58.1 million. (Id. ¶ 305). This was attributed primarily to a substantial decrease in loans generated by Upstart's platform, due to the Fed's decision to continuing raising interest rates, which in turn increased the costs of borrowing and depressed demand for loans. (See id. ¶ 305). Previously, Girouard had indicated that the AI model was “rate agnostic,” but he and Datta, in discussing Upstart's Q3 2021 performance, acknowledged that higher interest rates had led to the AI model underwriting pricing loans at higher annual percentage rates (“APRs”), i.e., higher loan interest payments. (Id. ¶¶ 111, 305-06).

The next day, Upstart stock dropped to a low of $17.06. (Id. ¶ 122).

D. Procedural Background

In the wake of the events described above, several individuals and investment firms filed suit in this Court and in the U.S. District Court for the Northern District of California alleging violations of federal securities laws and regulations. The plaintiffs in each of these cases asserted, on behalf of themselves and all others similarly situated, that Upstart, its executives and other insiders, and Third Point had engaged in securities fraud, including the dissemination of material misstatements. Plaintiffs in the Northern District of California voluntarily dismissed their cases. (See Unopposed Mot. for Appointment as Lead Counsel at 2 n.3, ECF No. 17). Plaintiffs in this Court consolidated their efforts behind Universal-Investment-Gesellschaft mbHaft as Lead Plaintiff, pursuant to 15 U.S.C. § 78u-4(a)(1). (See generally id.). Separately, several individuals also filed derivative actions nominally on behalf of Upstart in this Court; those cases have been consolidated and stayed pending a decision in this case. (See also ECF, No. 2:22-cv-02961).

Upstart is co-headquartered in San Francisco, within the jurisdiction of the Northern District of California, and in Columbus, within the jurisdiction of this Court.

On December 5, 2022, Lead Plaintiff Universal-Investment-Gesellschaft mbHaft and Plaintiffs Sebag and Brooks submitted their Consolidated Amended Complaint (ECF No. 45). The complaint asserts a putative class action, involving three counts: (1) violations of § 10(b) of the Exchange Act and SEC Rule 10b-5 against all defendants; (2) violations of § 20(a) of the Exchange Act against Girouard, Gu, Counselman, Datta, Schwartz, Third Point, and Third Point Ventures; and (3) violations of § 20A of the Exchange Act against Girouard, Gu, Counselman, Third Point, Third Point Ventures, and Loeb. (See generally id.). Plaintiffs propose a Class Period of December 16, 2020, through November 8, 2022. (Compl. ¶ 1, ECF No. 45). The Upstart Defendants and the Third Point Defendants have each filed a motion to dismiss. Plaintiffs oppose, and separately move to strike a portion of the Third Point Defendants' reply brief in support of their motion to dismiss. All motions are now ripe for review.

The Third Point Defendants incorporated by reference the Upstart Defendants' motion to dismiss.

III. STANDARD OF REVIEW

A motion to dismiss pursuant to Rule 12(b)(6) operates to evaluate the sufficiency of the complaint and permits dismissal of a complaint for “failure to state a claim upon which relief can be granted.” FED. R. CIV. P. 12(b)(6). Such a motion “is a test of the plaintiff's cause of action as stated in the complaint, not a challenge to the plaintiff's factual allegations.” Golden v. City of Columbus, 404 F.3d 950, 958-59 (6th Cir. 2005) (citation omitted). Accordingly, the Court must “construe the complaint in the light most favorable to the plaintiff, accept its allegations as true, and draw all reasonable inferences in favor of the plaintiff.” In re Travel Agent Comm'n Antitrust Litig., 583 F.3d 896, 903 (6th Cir. 2009) (quoting Jones v. City of Cincinnati, 521 F.3d 555, 559 (6th Cir. 2008)). Although the court's primary focus will be on the allegations in the complaint, the court may also consider “any exhibits attached thereto, public records, items appearing in the record of the case and exhibits attached to the defendant's motion so long as they are referred to in the Complaint and are central to the claims contained therein.” Bassett v. Nat'l Collegiate Athletic Ass'n, 528 F.3d 426, 430 (6th Cir. 2008) (citation omitted).

To survive a motion to dismiss, “the plaintiff must allege facts that, if accepted as true, are sufficient to raise a right to relief above the speculative level and to state a claim to relief that is plausible on its face.” Hensley Mfg. v. Pro Pride, Inc., 579 F.3d 603, 609 (6th Cir. 2009) (internal quotation marks omitted) (citing Bell Atlantic Corp. v. Twombly, 550 U.S. 544, 570 (2007)). A claim is considered plausible on its face “when the plaintiff pleads factual content that allows the court to draw the reasonable inference that the defendant is liable for the misconduct alleged.” Id. (quoting Ashcroft v. Iqbal, 556 U.S. 662, 678 (2009)). And though the court “must accept all well-pleaded factual allegations in the complaint as true,” the court “need not ‘accept as true a legal conclusion couched as a factual allegation.'” Id. (quoting Twombly, 550 U.S. at 555).

But when a plaintiff's claims sound in fraud, she must satisfy a more stringent burden. Rule 9(b) requires that allegations of fraud “state with particularity the circumstances constituting fraud or mistake.” FED. R. CIV. P. 9(b); see also United States ex rel. SNAPP, Inc. v. Ford Motor Co., 532 F.3d 496, 504 (6th Cir. 2008) (noting that this heightened requirement “reflects the rulemakers' additional understanding that, in cases involving fraud and mistake, a more specific form of notice is necessary to permit a defendant to draft a responsive pleading” (internal quotation marks and citation omitted)). Particularity, in this context, requires at minimum allegations of “the time, place, and content of the alleged misrepresentation on which he or she relied; the fraudulent scheme; the fraudulent intent of the defendants; and the injury resulting from the fraud.” Bennett v. MIS Corp., 607 F.3d 1076, 1100 (6th Cir. 2010) (quoting Yuhasz v. Brush Wellman, Inc., 341 F.3d 559, 563 (6th Cir. 2003)).

Additionally, the Private Securities Litigation Reform Act of 1995 (“PSLRA”), Pub. L. No. 104-67, 109 Stat. 737 (codified as amended in scattered sections of 15 U.S.C.), sets out additional standards applicable specifically to allegations of securities fraud. The PSLRA requires plaintiffs to “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 information and belief, . . . [to] state with particularity all facts on which that belief is formed.” 15 U.S.C. § 78u-4(b)(1)(B). She must also “state with particularity facts giving rise to a strong inference that the defendant acted with the required state of mind”-i.e., with the requisite scienter. Id. § 78u-4(b)(2)(A); see also Tellabs, Inc. v. Makor Issues & Rights, Ltd., 551 U.S. 308, 313 (2007). This “strong inference” of scienter “must be cogent and at least as compelling as any opposing inference of nonfraudulent intent.” Tellabs, 551 U.S. at 314; see also id. at 324. Taken together, “Rule 9(b) and the PSLRA require a complaint to allege the ‘who, what, where, when, and why' of the fraudulent statements.” City of Taylor Gen. Emps. Ret. Sys. v. Astec Indus., 29 F.4th 802, 810 (6th Cir. 2022) (citation omitted).

IV. LAW & ANALYSIS

A. Preliminary Matters

This Court first addresses Plaintiffs' Motion to Strike (ECF No. 65) before turning to the motions to dismiss. In the Third Point Defendants' opening brief in support of their motion to dismiss, they argue inter alia that the stock sales that Plaintiffs allege are evidence of insider trading were, in fact, conducted in the ordinary course of business for a venture capital firm like Third Point. (See Third Point Memo. in Supp. of Mot. to Dismiss (“TP MTD”) at 5, 12, ECF No. 60-1). Instead, they write, the stock sales “strongly support the competing inference of non-fraudulent conduct.” (Id. at 12). Plaintiffs opposed this argument, countering that the Third Point Defendants had failed to present a non-fraudulent explanation for the stock sales, which Plaintiffs view as “anything but normal or routine” and as “suspiciously timed.” (See Resp. in Opp'n to Third Point Mot. to Dismiss (“TP Opp'n”) at 15, ECF No. 62). The Third Point Defendants reply, explaining that the stock sales were routine, referencing the firm's “exit[] [from] several other technology companies around the same time as it exited Upstart.” (Third Point Reply Br. at 14, ECF No. 64). They suggest that their SEC filings corroborate this trading pattern.

Plaintiffs now assert that this line of argument and the incorporation of the December 2021 and March 2022 Form 13F documents “impermissibly attempt[] to offer new legal arguments and information from new extrinsic materials.” (Mot. to Strike at 1, ECF No. 65). Their motion poses three questions: (1) whether the Third Point Defendants have improperly raised an argument for the first time on reply; (2) whether this Court may take judicial notice of Third Point's Forms 13F; and (3) whether Plaintiffs should be permitted to file a sur-reply if this Court decides not to strike the argument and the forms. This Court addresses each question in turn.

First, it is well-settled that a movant “cannot raise new issues in a reply brief; [s]he can only respond to arguments raised for the first time in [the nonmovant's] brief” or issues already discussed in her opening brief. United States v. Jerkins, 871 F.2d 598, 602 n.3 (6th Cir. 1989) (citation omitted); In re FirstEnergy Corp. Sec. Litig., 316 F.Supp.2d 581, 599 (N.D. Ohio 2004) (same). This rule is not meant to prevent a movant from elaborating on her prior arguments, proposing new examples that illustrate her point, or parrying her opponent's thrusts in her reply brief. Rather, it bars only those arguments that are genuinely new. See, e.g., United States v. Campbell, 279 F.3d 392, 401 (6th Cir. 2002) (disposing of an argument that the litigant “failed to raise, even perfunctorily . . . in his original brief”). And this makes sense. The rule is in place to ensure that the non-movant is on notice and has an opportunity to respond. See also Knighten v. Comm'r of Internal Revenue, 702 F.2d 59, 60 n.1 (5th Cir. 1983), cert. denied, 464 U.S. 897 (1983). When a movant is simply discussing a variation on a theme in her reply brief rather than a new melody, the non-movant has already been placed on notice and there is no prejudice.

That is the case here. The allegedly new argument (that Third Point's stock sales coincided with its exit from other holdings) is a continuation of an issue already discussed in the Third Point Defendants' opening brief (that Third Point's stock sales were consistent with its ordinary business dealings), and is directly responsive to the counterarguments raised in Plaintiffs' opposition (that the opposite is true). That is, the Third Point Defendants' opening brief clearly signaled their disagreement with the inferences Plaintiffs drew from the stock sale, and their reply brief provided further explanation for that disagreement. Cf. In re First Energy, 316 F.Supp.2d at 599 (dismissing an attack on scienter that “Defendants failed to even passingly raise . . . in their original motion”). Hence, the challenged argument was not improperly raised for the first time in a reply brief.

Second, courts are generally prohibited from considering materials outside the pleadings on a motion to dismiss, but can consider “other sources courts ordinarily examine when ruling on Rule 12(b)(6) motions to dismiss, in particular, documents incorporated into the complaint by reference, and matters of which a court may take judicial notice.” Tellabs, 551 U.S. at 322 (citation omitted). Courts may take judicial notice of “a fact that is not subject to reasonable dispute because it: (1) is generally known within the trial court's territorial jurisdiction; or (2) can be accurately and readily determined from sources whose accuracy cannot reasonably be questioned.” FED. R. EVID. 201(b). While documents filed with the SEC are typically considered public records subject to judicial notice, see Lovelace v. Software Spectrum Inc., 78 F.3d 1015, 1018 (5th Cir. 1996) (allowing for judicial notice of “public disclosure documents which (1) are required to be filed with the SEC, and (2) are actually filed with the SEC”), it is important to remember that courts can “not consider the statements contained in [an SEC filing] for the truth of the matter asserted [] at the motion-to-dismiss stage.” In re Omnicare, Inc. Sec. Litig., 769 F.3d 455, 467 (6th Cir. 2014) (citations omitted); see also Lovelace, 78 F.3d at 1018.

The documents the Third Point Defendants seek to introduce-Forms 13F-were filed by Third Point with the SEC. Thus, they fall into the category of public documents that can be subject to judicial notice. But the Third Point Defendants rely on the Forms 13F for the truth of the matter asserted therein. They suggest that the stock transactions listed therein, taken as true, support their argument that their trades vis-a-vis Upstart were not suspicious. (See Third Point Reply Br. at 14, ECF No. 64). Because this is an improper use of the documents, this Court strikes these Forms 13F exhibits. See In re First Energy, 316 F.Supp.2d at 592; In re Omnicare, 769 F.3d at 467; see also Khoja v. Orexigen Therapeutics, Inc., 899 F.3d 988, 999 (9th Cir. 2018) (“Just because the document itself is susceptible to judicial notice does not mean that every assertion of fact within that document is judicially noticeable for its truth.”); Passa v. City of Columbus, 123 Fed.Appx. 694, 697 (6th Cir. 2005) (collecting cases).

By contrast, it would be permissible to rely on the contents of SEC filings to demonstrate what statements or disclosures were made by corporate defendants in assessing whether they have violated their duty to disclose under Rule 10b-5. See, e.g., In re Direct Gen. Corp. Sec. Litig., 398 F.Supp.2d 888, 893 (M.D. Tenn. 2005).

Finally, Plaintiffs ask to submit a sur-reply if this Court finds that striking the Third Point Defendants' argument is unwarranted. (Mot. to Strike at 7-8, ECF No. 65). Although the Federal Rules of Civil Procedure do not explicitly contemplate the filing of sur-replies, this Court's Local Rules permit additional memoranda “upon leave of court for good cause shown.” S.D. OHIO CIV. R. 7.2(a)(2). The Rules do not define good cause, but the Sixth Circuit has noted that additional filings “may be allowed in the appropriate circumstances, especially ‘[w]hen new submissions and/or arguments are included in a reply brief, and a non-movant's ability to respond to the new evidence has been vitiated.'” Key v. Shelby Cnty., 551 Fed.Appx. 262, 265 (6th Cir. 2014) (alteration in original) (quoting Seay v. Tenn. Valley Auth., 339 F.3d 454, 481 (6th Cir. 2003)). But a party's desire to point out mischaracterizations or misrepresentations of caselaw, even if helpful, does not provide grounds for the filing of sur-replies. See Little Hocking Water Ass'n, Inc. v. E.I. Du Pont de Nemours & Co., No. 2:09-cv-01081, 2014 WL 12651139, at *2 (S.D. Ohio Oct. 31, 2014).

Because this Court concludes that the Third Point Defendants' alleged “new argument” was not an improper new argument, contrary to Plaintiffs' assertions, see supra, Plaintiffs have failed to demonstrate “good cause” justifying the filing of a sur-reply. Moreover, even a cursory review of Plaintiffs' proposed sur-reply reveals that the filing is dedicated almost entirely to distinguishing the Third Point Defendants' cited cases and explaining how they have improperly characterized the reasoning in those decisions. (See generally Proposed Sur-Reply, ECF No. 651). Permission to file a sur-reply is not warranted under such circumstances.

Accordingly, Plaintiffs' Motion to Strike (ECF No. 65) is Granted in part and denied in part. This Court Strikes the Forms 13F proffered by the Third Point Defendants. This Court will nevertheless consider all arguments raised in their reply brief, but not those contained in Plaintiffs' proposed sur-reply.

B. Securities Fraud (Count I)

Section 10(b) of the Exchange Act prohibits “directly or indirectly . . . us[ing] or employ[ing], in connection with the purchase or sale of any security registered on a national securities exchange . . . any manipulative or deceptive device or contrivance in contravention of such rules and regulations as the [Securities and Exchange] Commission may prescribe as necessary or appropriate in the public interest or for the protection of investors.” 15 U.S.C. § 78j(b). Pursuant to this directive, the SEC has promulgated Rule 10b-5, which prohibits three types of actions: (1) “employ[ing] any device, scheme, or artifice to defraud,” (2) “mak[ing] any untrue statement of a material fact or [] omit[ting] to state a material fact necessary in order to make the statements made, in the light of the circumstances under which they were made, not misleading,” and (3) “engag[ing] in any act, practice, or course of business which operates or would operate as a fraud or deceit upon any person.” 17 C.F.R. § 240.10b-5.

The Complaint purports to advance theories under each of the three categories of conduct prohibited by Rule 10b-5. The required elements for the three categories are nearly identical:

(1) a material misrepresentation or omission [or . . . a ‘deceptive or manipulative act']; (2) scienter, that is, a wrongful state of mind; (3) a connection with the purchase or sale of a security; (4) reliance; (5) economic loss; and (6) loss causation, a causal connection between the misrepresentation and the loss.

In re First Energy Corp. Sec. Litig., 2:20-cv-03975/04287, 2022 WL 681320, at *8 (S.D. Ohio Mar. 7, 2022) (citations omitted) (alteration in original); see also Gordon v. Royal Palm Real Est. Inv. Fund I, LLLP, 320 F.Supp.3d 910, 923 (E.D. Mich. 2018) (applying the misstatement test to scheme liability).

This Court first considers the “misstatement theory” of securities fraud liability in this subsection, pursuant to Rule 10b-5(b), before turning to the “scheme liability” theory in Part IV.C, supra. In total, Plaintiffs identify dozens of allegedly false and misleading statements made by Upstart and its executives during the Class Period. (See generally Compl. ¶¶ 123-259, ECF No. 45). This covers statements made: (1) in Upstart's IPO and SPO Registration Statements filed with the SEC; (2) in Upstart's annual and quarterly SEC filings (i.e., Forms 10-K and 10-Q, respectively), on which Upstart reported its financial performance; (3) by executives during quarterly earnings calls; and (4) by executives in other settings, such as podcast interviews, industry events and conferences. (Id.).

As an initial matter, the Third Point Defendants are not exposed to misstatement liability under Section 10(b)/Rule 10b-5 because they never “made” any of the challenged statements. A bit of background first: the parties dispute whether Schwartz is an “Upstart Defendant” or a “Third Point Defendant.” (See TP MTD at 4 n.1, ECF No. 60-1; TP Opp'n at 3-5, ECF No. 62). Schwartz is the Managing Partner of Third Point Ventures; he took a seat on the board of directors of Upstart concurrent with Third Point's investment in Upstart. Plaintiffs argue that Schwartz should be classified as a Third Point Defendant because his actions as an Upstart director (including his participation in board meetings and his sign-off on Upstart's SEC filings) were done in his capacity as an employee of Third Point Ventures. (See TP Opp'n at 4, ECF No. 62).

But while courts must credit plaintiffs' versions of the facts at the motion to dismiss stage, see Mediacom Se. LLC v. BellSouth Telecomms., Inc., 672 F.3d 396, 400 (6th Cir. 2012), they cannot ignore legal principals and formalities in doing so. It is axiomatic that a corporate director owes fiduciary duties of care and loyalty to the corporation and its shareholders, not to anyone else. See Gantler v. Stephens, 965 A.2d 695, 708-09 (Del. 2009); see also In re Amcast Indus. Corp., 365 B.R. 91, 104 (Bankr. S.D. Ohio 2007). That is, directors “cannot be presumed to have been acting at the direction of their outside employers in their capacities as [] directors.” In re Glob. Crossing, Ltd. Sec. Litig., No. 02 Civ. 910, 2005 WL 1907005, at *4 (S.D.N.Y. 2005); cf. Williams v. Clerac, LLC, 635 F.Supp.3d 607, 613 (N.D. Ohio 2022) (noting that the respondeat superior doctrine only makes an employer vicariously liable for torts committed by the employees within the scope of employment). As applied here, Schwartz's actions as a director of Upstart cannot be conflated with his role as a Third Point Ventures employee-unless he was violating his fiduciary duties to Upstart and acting on behalf of Third Point Ventures while carrying out his director duties. The complaint is bereft of any allegations that that was the case. This Court therefore categorizes Schwartz as an Upstart Defendant and not as a Third Point Defendant.

For this reason, the Third Point Defendants are not subject to liability for false or misleading statements, which is limited to the “maker” of the false or misleading statement. Janus Cap. Grp., Inc. v. First Derivative Traders, 564 U.S. 135, 142 (2011). The “maker” is defined as someone with “ultimate authority over the statement, including its content and whether and how to communicate it.” Id. Merely publishing or helping to prepare a statement, but not speaking or signing the statement, therefore, is not enough to expose a defendant to liability. Id.; see also Id. at 143 (“This rule might best be exemplified by the relationship between a speechwriter and a speaker. Even when a speech writer drafts a speech, the content is entirely within the control of the person who delivers it. And it is the speaker who takes credit-or blame-for what is ultimately said.”).

None of the Third Point Defendants is adequately alleged to have “made” any of the challenged statements. (See generally Compl. ¶¶ 1-345, ECF No. 45). Neither Plaintiffs' speculative claims that the Third Point Defendants were involved in deciding which disclosures Upstart would make, unadorned with any plausible factual allegations, nor their attempt to group Schwartz with the Third Point Defendants alters this analysis. It is undisputed that the challenged statements made at conferences, during interviews, and on podcasts were spoken-i.e., “made”- by Upstart executives. See Janus, 564 U.S. at 143. It is similarly undisputed that the challenged SEC filings were signed-i.e., “made”-by Upstart executives and directors (including Schwartz, who is not alleged to have been acting in his capacity as a Third Point Ventures employee when he signed). See In re Smith Barney Transfer Agent Litig., 884 F.Supp.2d 152, 163-64 (S.D.N.Y. 2012). Janus compels the conclusion that, if Upstart executives and directors made the statements, then the Third Point Defendants did not. In re Galena Biopharma, Inc. Sec. Litig., 117 F.Supp.3d 1145, 1187 (D. Or. 2015) (“[W]here legally distinct entities are involved, only one entity has the final say in what, if anything, is published.”). Accordingly, because the Third Point Defendants did not “make” any of the challenged statements, there is no basis for sustaining a misstatement claim against them. Their motion to dismiss is Granted on this claim.

This Court next turns to the claims against the Upstart Defendants, who challenge only the first and second elements of a Rule 10b-5: they assert that none of the challenged statements was false or misleading and that Plaintiffs have failed to plead a strong inference of scienter.

1. False or Misleading

Adhering in part to the taxonomy suggested by the Upstart Defendants, this Court groups the alleged misstatements into three broad categories: (1) statements about Upstart's proprietary AI model; (2) statements about Upstart's business model and the retention of loans; and (3) statements made as part of Sarbanes-Oxley Act (“SOX”) certifications. (See Upstart Defs.' Memo. in Supp. of Mot. to Dismiss (“Upstart MTD”) at 10-11, ECF No. 58-1). The Upstart Defendants suggest that the statements in the first and second categories were simply puffery, opinion statements, forward-looking statements subject to the PSLRA safe harbor, or not false when made, and that the statements in the third categories were not false.

Plaintiffs argue that the Upstart Defendants failed to provide arguments about several of the challenged statements and thus have waived the issue. (Resp. in Opp'n to Upstart MTD (“Plaintiffs' Opp'n”) at 13-15, ECF No. 61) (referencing ¶¶ 203, 216-18, 230-31, 233 of the Amended Complaint). This includes statements about the impact of the end of the stimulus on Upstart's business, the pricing of loans, and the AI underwriting model being “rate agnostic” and doing the right things. But Plaintiffs rely on an overly restrictive understanding of waiver. Each of these statements falls within one of the categories proposed by the Upstart Defendants; the Upstart Defendants explicitly argued that “none” of the challenged statements were false or misleading; and they directly addressed portions of several of the statements. They were not required to discuss every single sentence of each of the 61 paragraphs in the Complaint describing allegedly false or misleading statements.

Before addressing the merits of these arguments, this Court first provides a brief review of some generally relevant background principles. First, defenses like puffery go to Section 10(b)'s materiality requirement, which is intended “to filter out essentially useless information that a reasonable investor would not consider significant, even as part of a larger ‘mix' of factors to consider in making his investment decision.” Basic, Inc. v. Levinson, 485 U.S. 224, 234 (1988) (citation omitted). Thus, statements involving material facts-i.e., facts where there “is a substantial likelihood that a reasonable shareholder would consider it important in deciding how to vote” or invest, id. at 231 (quoting TSC Indus., Inc. v. Northway, Inc., 426 U.S. 438, 449 (1976))-are actionable, but “[i]mmaterial statements [like] vague, soft, puffing statements or obvious hyperbole” are not. In re K-tel Int'l, Inc. Sec. Litig., 300 F.3d 881, 897 (8th Cir. 2002); see also In re Ford Motor Co. Sec. Litig., 381 F.3d 563, 570-71 (6th Cir. 2004) (“Courts everywhere ‘have demonstrated a willingness to find immaterial as a matter of law a certain kind of rosy affirmation commonly heard from corporate managers and numbingly familiar to the marketplace-loosely optimistic statements that are so vague, so lacking in specificity, or so clearly constituting the opinions of the speaker, that no reasonable investor could find them important to the total mix of information available.'” (quoting Shaw v. Digit. Equip. Corp., 82 F.3d 1194, 1217 (1st Cir. 1996))).

But the Sixth Circuit has cautioned district courts to “tread lightly” when assessing materiality at the motion to dismiss stage, in recognition of the federal judiciary's “limited understanding of investor behavior and the actual economic consequences of certain statements.” In re Omnicare, Inc. Sec. Litig., 769 F.3d at 472. Of course, granting a motion to dismiss on the basis of immateriality is sometimes warranted; examples abound where courts have done so. See, e.g., id.; In re Ford Motor Co., 381 F.3d at 461. But it should not be done lightly, and certainly not without first taking into careful consideration the context of the alleged misstatements. After all, a statement that appears to be puffery or opinion statements at first glance may turn out to be material upon closer examination of the circumstances in which it was made-if it was “used to emphasize and induce reliance,” Casella v. Webb, 883 F.2d 805, 808 (9th Cir. 1989), for example, or if a reasonable investor would “understand [the] opinion statement to convey facts about how the speaker has formed the opinion.” Omnicare, Inc. v. Laborers Dist. Council Constr. Indus. Pension Fund, 575 U.S. 175, 188 (2015).

Second, the PSLRA includes a safe harbor provision for forward-looking statements, like revenue projections, future objectives, future economic performance, etc. 15 U.S.C. § 78u-5(c); see also id. § 78u-5(i)(1) (defining “forward-looking statement”); H.R. REP. No. 104-369, at 45 (1995) (Conf. Rep.). If a “forward-looking statement” is accompanied by meaningful cautionary language, “a defendant's statement is protected regardless of the actual state of mind.” Miller v. Champion Enters. Inc., 346 F.3d 660, 672 (6th Cir. 2003). If it is “not accompanied by meaningful cautionary language, actual knowledge of [its] false or misleading nature is required.” Id. (citing 15 U.S.C. § 78u-5(c)(1)(B); Helwig v. Vencor, Inc., 251 F.3d 540, 552 (6th Cir. 2001) (en banc), cert. denied, 536 U.S. 935 (2002). To be considered meaningful, cautionary language must be specific and substantive enough that it “realistically could cause results to differ materially from those projected in the forward-looking statement.” Helwig, 251 F.3d at 558-59 (citation omitted); see also Asher v. Baxter Int'l, Inc., 377 F.3d 727, 732-33 (7th Cir. 2004) (deeming boilerplate warnings like “all businesses are risky” not meaningful, while noting that “the cautions need not identify what actually goes wrong and causes the projections to be inaccurate”).

Against this background, this Court now turns to the specific statements made by the Upstart Defendants.

a. Statements about the AI Model

At various points, the Upstart Defendants touted the strength of the AI underwriting model. This is the centerpiece of Upstart's business: it is what improves upon the traditional FICO-based system for determining a potential borrower's creditworthiness and makes Upstart's platform appealing for lenders who think that the old approach prices loans inaccurately. There are a few sub-categories of statements about the model.

First: statements about the superiority of the AI model. Some of these statements were in actionable puffery. Puffery, as noted earlier, refers to vague, optimistic statements that, “both on their own terms and in context, lack[] a standard against which a reasonable investor could expect them to be pegged.” City of Monroe Emps. Ret. Sys. v. Bridgestone Corp., 399 F.3d 651, 671 (6th Cir. 2005). Courts frequently dismiss assertions about a product's “value,” “strength,” or “quality” as puffery. See In re Trans Digm Grp., Sec. Litig., 440 F.Supp.3d 740, 763-64 (N.D. Ohio 2020); see also In re Cable & Wireless, PLC, Sec. Litig., 321 F.Supp.2d 749, 767 (E.D. Va. 2004) (collecting cases). And here, statements that turbulent times presented an “opportunity for a new platform to shine,” that the platform offered “compelling and valuable [benefits],” and that the AI model was a “fairly magical thing” capable of “doing the right things” are all “loosely optimistic statements” that cannot be objectively verified. Bridgestone, 399 F.3d at 669; see, e.g., San Leandro Emergency Med. Grp. Profit Sharing Plan v. Philip Morris Cos. Inc., 75 F.3d 801, 807, 811 (2d Cir. 1996); Colby v. Hologic, Inc., 817 F.Supp. 204, 211 (D. Mass. 1993). There is no measuring stick by which the veracity of these optimistic statements can be evaluated-no objective way for reasonable investors (or this Court) to assess whether Upstart's AI model was, in fact, “magical” or whether it “shined.” (See, e.g., Compl. ¶¶ 132, 138, 203, ECF No. 45). These statements are inactionable.

Other statements made by the Upstart Defendants about the superiority of the AI model are more specific, material, and verifiable. Consider, for example, statements about the “significant advantage” of the AI model over traditional FICO-based models. The Upstart Defendants had explained what those purported advantages are: “higher approval rates and lower interest rates at the same loss rate.” (See, e.g., Compl. ¶ 126, ECF No. 45; id. ¶ 145). They further emphasized that this superiority would be especially noticeable in elevated risk environments. (See, e.g., id. ¶ 132) (“[T]he level of risk in the environment went up a lot, and that really meant that it really stressed out competing models. And we feel like we have pretty significant advantage in an environment where there's elevated levels of risk.”). Whether the model did in fact provide a “significant advantage” during times of elevated risk can be objectively assessed by looking to these metrics. Cf. In re Cable & Wireless, 321 F.Supp.2d at 767 (finding generalized statements about “competitive advantage” to be immaterial where the “advantage” is undefined). And Plaintiffs have adequately pled that the Upstart model did not provide these verifiable advantages, in general or in times of macroeconomic turbulence. (See, e.g., Compl. ¶ 110, ECF No. 45) (noting the rising default rate for Upstart loans as the Fed hiked interest rates). These statements are actionable material misstatements under the first step of the Rule 10b-5 analysis.

Second: statements that the AI model accounted for the pandemic and stimulus. The Upstart Defendants explained on investor calls and in SEC filings that they were regularly updating the AI model and, more specifically, that the model had been “refined and updated to account for the COVID-19 pandemic,” which presented “unique circumstances” that Upstart had “really built into our model.” (Id. ¶¶ 137, 145). But Plaintiffs fail to plead adequately that these statements were false or misleading: there are no allegations that Upstart was not regularly updating its model, or that Upstart did not make adjustments to account for the unusual economic conditions created by the pandemic and the federal stimulus packages. Rather, the crux of the allegations is simply that the AI model was not able to adapt “dynamically” to the end of the stimulus, see infra. That, however, poses a distinct question from whether the AI model adapted at all or whether Upstart asked the model to consider certain factors. As such, Plaintiffs fail to plead that statements about Upstart's model being updated to account for the pandemic or the stimulus are false. Therefore, these statements are inactionable. (See Compl. ¶¶ 137, 154, 171, 206, 222, ECF No. 45).

Third: statements that the AI model had the ability to “respond very dynamically” to macroeconomic changes. One of the main advantages of the AI model, according to the Upstart Defendants, is its ability to “very quickly adjust itself to handle the macro situation that is evolving out there,” even when macroeconomic trends in the broader economy are going through substantial “disruptions” and “dislocations.” (Id. ¶ 161; see also id. ¶ 145). Rising interest rates represented the key macroeconomic change during the Class Period. On that front, Girouard asserted that the AI model is a “rate agnostic system. As interest rates go up our system adjusts to that but we're not overly rate sensitive like you might see elsewhere.” (Id. ¶ 203).

Plaintiffs also argue that statements that the AI was automated were false or misleading, claiming that Datta admitted the model required manual on a May 2022 earnings call. (See, e.g., Plaintiffs' Opp'n at 47, ECF No. 61). During the call, Datta said:

[W]hen interest rates in the economy change quite quickly, I think it would be fair to say that our platform, its ability to react to the new market-clearing price, it's probably not as nimble as we would like. It's somewhat manual. It requires a bunch of conversations and phone calls. And so when interest rates smooth and investors are-so each deciding what their new return hurdles are, there can be a gap or a delay in responding to funding.

(Compl. ¶ 241, ECF No. 45). But, as the Upstart Defendants explain and as is evident from the statement itself, Datta was discussing the process by which Upstart identified investors to sell their loans to, which had become more manual as interest rates rose. (See Upstart MTD at 16, ECF No. 58-1). He was not discussing the underwriting model, which continued to assess risk and price loans without manual intervention. In other words, Plaintiffs' only allegation that the model was not automated is predicated on an alleged “admission” about a different topic.

Whether these statements are inactionable puffery presents a close call. They are certainly difficult to verify: for example, how quickly must a model react for the statement to be true? On the other hand, they are certainly more specific than vague affirmations about the “benefits” of the model or its ability to “do the right thing”; they are about a specific attribute of the model-its speed. Moreover, the purported ability of the AI model to react quicker and more accurately than traditional FICO-based models is at the core of Upstart's pitch to lenders and investors. See Bridgestone, 399 F.3d at 672 (“What might be innocuous ‘puffery' . . . standing alone may be actionable as an integral part of a representation of material fact when used to emphasize and induce reliance upon such a representation.” (quoting Casella, 883 F.2d at 808)). Statements about the speed and adaptability of the model, in other words, would likely be considered essential to a reasonable investor. See also Brumbaugh v. Wave Sys. Corp., 416 F.Supp.2d 239, 250-51 (D. Mass. 2006). Thus, they are material and are not mere puffery.

Nor are these statements inactionable opinions merely because the Upstart Defendants prefaced them with “we believe” or “we feel.” (See, e.g., id. ¶¶ 132, 230). Typically, opinion statements are deemed inactionable “unless the statement (a) expresses an opinion the speaker did not actually hold, (b) contains embedded statements of fact that are untrue, or (c) omits material facts about the issuer's inquiry into or knowledge concerning a statement of opinion such that the statement as a whole is misleading to a reasonable investor.” In re FirstEnergy Corp., 2022 WL 681320, at *10 (cleaned up) (citation omitted). The challenged statements about the ability of the AI model to “react quickly” in this case contain “embedded statements of fact”-i.e., that the AI model can adapt quickly-that Plaintiffs have alleged are untrue. Omnicare, 575 U.S. at 185; cf. id. at 193 (noting that, because “those magic words [‘we believe' or ‘we think'] can preface nearly any conclusion, and . . . remain perfectly capable of misleading investors,” they should not be construed to give companies “virtual carte blanche to assert opinions in registrations statements free from worry about” liability).

The same applies to statements about the advantages of Upstart's AI model over traditional models, which imply a factual assertion that the AI model produces higher approval rates, at lower interest and loss rates.

The Upstart Defendants' final critique, that these statements were not false when made, implicates the “fraud by hindsight” doctrine. (See Upstart MTD at 16, ECF No. 58-1). “Fraud by hindsight” suggests that courts should view with skepticism cases “where a plaintiff alleges that the fact that something turned out badly must mean defendant knew earlier that it would turn out badly.” La. Sch. Emps.' Ret. Sys. v. Ernst & Young, LLP, 622 F.3d 471, 484 (6th Cir. 2010) (citation omitted). In other words, “fraud by hindsight” is a warning for courts, to not be blinded by their post hoc knowledge in analyzing alleged misstatements. As applied here, the Upstart Defendants suggest that statements about the AI model's ability to adapt quickly were true at the time they were made, even if the model later struggled to adapt to unprecedented rate hikes.

But courts must be cautious when considering fraud by hindsight arguments at the motion to dismiss stage because there has been little if any discovery and the court has no visibility into the evidence that was available to the corporate managers when they made the allegedly-false statements. Moreover, “at this stage, the allegation that a forecast has turned out to be inaccurate provides at least some affirmative evidence that fraud might have occurred.” Mitu Gulati, Jeffrey J. Rachlinski & Donald C. Langevoort, Fraud by Hindsight, 98 NW. L. REV. 773, 787 (2004). This is especially true where the allegedly-false statement is about a fundamental aspect of the business-a topic that the corporate managers have discussed and studied extensively. Cf. In re Cardinal Health Inc. Sec. Litigs., 426 F.Supp.2d 688, 724 (S.D. Ohio 2006). The AI model's ability “quickly [to] adjust itself to handle the macro situation” and its sensitivity to interest rates are both fundamental to Upstart's business. The Fed's interest rates, after all, dictate the cost of borrowing and therefore have a massive impact on demand for loans. It is not credible that executives at a lending platform would not have studied and been intimately aware of how their underwriting model accounts for rate hikes or cuts, i.e., the rate-sensitivity of the underwriting model. The same goes for the model's ability to “very quickly adjust” to changing conditions, where that ability is one of its defining advantages over traditional approaches. (See generally IPO S-1, ECF No. 58-4) (comparing the Upstart AI model favorably to the traditional FICO model because of its ability to improve in real time and respond quickly to changing circumstances). This Court therefore rejects the Upstart Defendants' arguments that the statements about the AI model being rate-agnostic and adjusting rapidly were puffery, opinion statements, or not false when made.

Note what this conclusion does not include: alleged omissions by the Upstart Defendants for failing to disclose that the AI model could not adapt quickly to changing macroeconomic conditions. (See, e.g., Compl. ¶ 148, 152, ECF No. 45). Instead, it includes only affirmative statements by the Upstart Defendants about the ability of the AI model to adapt quickly. Cf. Helwig, 251 F.3d at 561 (differentiating between the “duty to speak” and “liability for not having spoken enough”); id. (“[A] company may choose silence or speech elaborated by the factual basis as then known-but it may not choose half-truths.”). In this case, the alleged omissions were disclosed in Upstart's public filing, including its 2020 10-K, which stated that its AI model “may not be able to effectively account for . . . macroeconomic conditions.” (FY20 10-K at 30, ECF No. 58-3). Plaintiffs cannot plausibly claim that the Upstart Defendants' omissions are material when the same information was disclosed publicly in SEC filings. See In re Every Ware Glob., Inc. Sec. Litig., 175 F.Supp.3d 837, 873 (S.D. Ohio 2016).

Accordingly, this Court finds that the third subcategory of statements-about the AI model's ability to adapt quickly and its rate-insensitivity-are actionable material misstatements under the first step of the Rule 10b-5 analysis. (See Compl. ¶¶ 125, 161, 203, ECF No. 45).

b. Statements about Business Model and Retention of Loans

The Upstart Defendants also made various statements during the Class Period about demand and funding for loans that Plaintiffs now assert were false or misleading. The statements implicate a fundamental question about the nature of Upstart's business-namely, whether it is merely a platform for connecting borrowers and lenders, or if it is itself a lender. As with the statements about the AI model, there are a few different sub-categories of statements at issue here.

First: statements about robust demand for Upstart-powered loans. Throughout the Class Period, the Upstart Defendants repeatedly stated that demand for Upstart-powered loans was robust. (See Compl. ¶¶ 133, 139, 151, 161, 165-66, 201-02, 214, ECF No. 45) (positive statements about demand from March 17, 2021, to February 15, 2022). Several of these statements were not false when made because they referred to the existence of demand at a particular point in time, but Plaintiffs have not pled contemporaneous facts demonstrating that there was a lack in demand at that time. (See id. ¶ 133) (discussing the robustness of the demand pipeline “today”).

Demand in this context refers to interest from lenders in assuming the loans that Upstart's platform had generated, not demand for loans from prospective borrowers. (See, e.g., Compl. ¶ 151, ECF No. 45) (discussing the “supply chain of money that funds Upstart loans”).

Others are forward-looking statements subject to the PSLRA's safe harbor provision. The challenged statements include projections for how demand for loans will look in the future, i.e., the Upstart Defendants' expectations about demand after the then-existing distortion in the market settled down. (See, e.g., id. ¶ 202). The PSLRA, as discussed earlier, precludes liability for forward-looking statements if they were “accompanied by meaningful cautionary statements.” See Miller, 346 F.3d at 671-72. Here, the Upstart Defendants disclosed in SEC filings that “[d]uring periods of economic slowdown or recessions, [Upstart's] current and potential investors in our loan funding programs may reduce the number of loans or interests in loans they purchase.” (FY20 Form 10-K at 33, ECF No. 58-3).

Plaintiffs suggest that this cautionary language was not meaningful for two reasons, though neither is particularly persuasive. As an initial matter, it is not the case that cautionary language must reduce the risk of deception to zero, as Plaintiffs claim without citation to any authority. Such a formulation is neither reasonable nor rooted in the caselaw, which requires only that “cautionary statements must convey substantive information . . . that realistically could cause results to differ materially from those projected in the forward-looking statements.” Helwig, 251 F.3d at 558-59. This is a less substantial burden than requiring cautionary statements to eliminate any and all risk of deception; note, for example, the language indicating possibility, and not certainty, used by the Sixth Circuit in Helwig. Id. (“could”). Nor does cautionary language need to be quite as narrow as Plaintiffs would like. They argue that, for the Upstart Defendants' statements to be protected, the cautionary language must have warned of the risk that the Upstart Defendants' “already knew that the Company's AI system could not adjust to macroeconomic changes quickly.” (Plaintiffs' Opp'n at 31-32, ECF No. 61) (emphasis in original). But there is no requirement that corporate officers word their cautions in just the way that Plaintiffs would find most helpful. The Upstart Defendants satisfied their burden: they issued cautionary language, providing substantive information about a risk of declining demand for loans in times of economic slowdowns-exactly what the Plaintiffs complain happened. And this caution was meaningful, as it disclosed a specific risk unique to the Upstart business, not a generic word of caution applicable to any company. See Asher, 377 F.3d at 733; see also Southland Sec. Corp. v. INSpire Ins. Sols., Inc., 365 F.3d 353, 372 (5th Cir. 2004).

Plaintiffs also suggest that courts should not consider the issue of whether cautionary language was meaningful at this stage, because doing so requires consideration of “questions of fact that are inappropriate for determination on [a] Motion to Dismiss.” (Plaintiffs' Opp'n at 32, ECF No. 61) (quoting In re Prison Realty Sec. Litig., 117 F.Supp.2d 681, 690 (M.D. Tenn. 2000)). The Sixth Circuit has rejected such an approach. See Helwig, 251 F.3d at 554 (“Nevertheless, Congress apparently intended the applicability of the safe harbor to be addressed even on a motion to dismiss.” (citing 15 U.S.C. § 78u-5(e))); see, e.g., In re Cardinal Health, 426 F.Supp.2d at 746.

Note that the warning Plaintiffs argue the Upstart Defendants should have given is just a variation of the warnings they did issue.

Nevertheless, there are two statements in this sub-category that are do not fall into the forward-looking or not false when made categories. The Upstart Defendants stated that “demand is there” for loans during an earnings call on November 9, 2021, and that Upstart had “deep and diverse sources of liquidity” available from investors during an earnings call on February 15, 2022. (Compl. ¶¶ 201, 214, ECF No. 45). Plaintiffs have pled sufficient contemporaneous facts alleging that these statements were false when made (in contrast to other, similar statements that were made earlier in 2021). Such facts include the precipitous increase in loans retained on Upstart's books between July 1, 2021, and September 30, 2021, and the continued increase through the end of 2021 and into 2021. (See Compl. ¶¶ 108, 112, ECF No. 45). Drawing inferences in Plaintiffs' favor, the sharp uptick in loans held strongly suggests that demand for Upstart loans was not “there” and that liquidity from investors was neither deep nor diverse-contrary to what the Upstart Defendants stated. Otherwise, one would assume, Upstart would have continued to find funding readily for its loans and not had to hold on to so many loans. But that did not happen. Plaintiffs have adequately pled that these statements were false and misleading.

Second: statements about not being a bank or a lender. The Upstart Defendants repeatedly stated that Upstart does not make loans, is not a lender, and is not a bank. (See Compl. ¶¶ 131, 147, 159, 161, 176, 235, 237, ECF No. 45). These statements go to the nature of Upstart's business model: whether it is a loan platform, connecting lenders and borrowers, or a lender itself. The Upstart Defendants maintain that these statements are true-that it was not and is not in the loan business. (See Upstart MTD at 21, ECF No. 58-1). Instead, the Upstart Defendants explain, its banking partners originated all loans generated by the platform; if the partner did not want to keep the loan, the agreement between Upstart and the partner stipulated that the partner could sell the loans back to Upstart. (See id.). And that, apparently, is what happened in increasing numbers in late 2021 and early 2022, leading to the ballooning volume of loans held by Upstart. But, even then, Upstart held the loans only as a “market-clearing mechanism.” (Compl. ¶ 241, ECF No. 45). That is, Upstart held the loans on its balance sheet while figuring out how to sell them to banks and investors. (See Upstart MTD at 21, ECF No. 58-1). At no point did Upstart intend to keep any non-R&D loans for the long-term.

Even as Upstart “increasingly add[ed] loans it could not sell to its balance sheet” during the third and fourth quarters of 2021, its fundamental business model did not change. (See Plaintiffs' Opp'n at 20-21, ECF No. 61) (citing Compl. ¶¶ 199, 240, ECF No. 45). Plaintiffs' argument relies on the proposition that, Upstart became a lender (counter to statements made by the Upstart Defendants) during this time period simply because it was holding on to substantial volumes of loans. But the reality is that Upstart remained a lending platform that intended to keep risk low by offloading loan liabilities to banks and investors. It did not begin originating any loans; instead, the loans it held were bought back from its banking partners, who continued to do the origination. It did not switch to a business model centered on interest revenue; it continued to focus on generating fees. (See, e.g., Compl. ¶ 213, ECF No. 45) (reiterating Upstart's intent to maintain a core business model centered on fees on February 15, 2022). And even though Upstart had deviated from its capital-light business model, that was always intended as a temporary stopgap, it did not intend to hold on to the loans for the long-term. (See also id. ¶¶ 244-45). The only thing that changed during this time period was Upstart's ability to execute their fee-based platform model well. Given this context, it is apparent that statements about Upstart not being a lender or a bank were not false.

This Court acknowledges that one dictionary definition of “lender” is “[o]ne who lends.” See Lender, OXFORD ENGLISH DICTIONARY, available at: https://perma.cc/T7J6-PU3U. But another definition is “one who makes a business of lending money at interest.” Id. Moreover, Rule 10b-5 looks at statements from the perspective of a reasonable investor, see Basic, 485 U.S. at 234, who is able to distinguish between the two.

Third: statements about the retention of loans. Plaintiffs challenge statements made by the Upstart Defendants about the retention of loans on Upstart's balance sheet, including the existence of loans retained, the purposes for retaining loans, and the length of time the loans were expected to remain on the balance sheet. (See, e.g., Compl. ¶¶ 197, 212-13, ECF No. 45). At the outset, it is important to note that the Upstart Defendants disclosed the retention of loans on its balance sheet, repeatedly and publicly. There is no serious dispute that they did so. (See, e.g., Form 10-Q at 7, ECF No. 58-15).

In some of these disclosures, the Upstart Defendants stated that loans were being retained for R&D purposes. Plaintiffs have adequately pled that the November 9, 2021, and February 15, 2022, statements attributing the increasing retention of loans on the balance sheet to the auto loan program were false when made. (Compl. ¶¶ 200, 212, ECF No. 45). The fact that the Upstart Defendants first explained this trend as part of an R&D effort, but later clarified that loans retained had increased as a “market-clearing mechanism” suggests that the first explanation was false or misleading. Moreover, the sheer volume of the increases in loans held contemporaneous with the statements undermines the Upstart Defendants' argument that the “market-clearing mechanism” did not kick in until a later time.

As to the statements that the loan retention problem was “a temporary thing” and that the loans were expected to be sold to banks and investors “over the next quarter or so,” the Upstart Defendants argue they are subject to the PSLRA's safe harbor. (Compl. ¶¶ 197, 213, ECF No. 45; Upstart MTD at 21-22, ECF No. 58-1). But Plaintiffs have pled that the statements were made without meaningful cautionary language and with actual knowledge of their falsity, i.e., that the Upstart Defendants knew that the loans were of such poor quality that they could not be sold easily. In such cases, the safe harbor provision is unavailable. After all, even “predictions and opinions contain ‘at least three implicit factual assertions: (1) that the statement is genuinely believed, (2) that there is a reasonable basis for that belief, and (3) that the speaker is not aware of any undisclosed facts tending to seriously undermine the accuracy of the statement.'” Helwig, 251 F.3d at 557 (quoting In re Apple Comput. Sec. Litig., 886 F.2d 1109, 1113 (9th Cir. 1989)). Where the speaker is alleged to have been aware of undisclosed facts undermining the accuracy of the statement, the premise behind protecting predictions from liability falls apart-as does the meaningfulness of cautionary language. See Weiner v. Tivity Health, 365 F.Supp.3d 900, 911 (M.D. Tenn. 2019) (“Cautionary language cannot be meaningful if it is misleading in light of historical facts that were established at the time the statement was made....By analogy, the safe harbor would not protect from liability a person who warns his hiking companion to walk slowly because there might be a ditch ahead when he knows with near certainty that the Grand Canyon lies one foot away.” (cleaned up) (quoting In re Harman Int'l Indus., Inc. Sec. Litig., 791 F.3d 90, 102-03 (D.C. Cir. 2015))). Because Plaintiffs have plausibly alleged that statements about the temporary nature of the loan problem were made with knowledge of inconsistent historical facts, they are not subject to the PSLRA safe harbor. (See Compl. ¶¶ 197, 213, ECF No. 45).

c. Item 303 of Regulation S-K

Relatedly, Plaintiffs also allege that the Upstart Defendant's failure to disclose this trend of increasing volumes of loans retained on Upstart's balance sheet trend is a violation of Item 303 of Regulation S-K, and therefore exposes them to Section 10(b) liability. (See generally Compl. ¶¶ 179-91, ECF No. 45). Item 303 requires companies to disclose “any known trends or uncertainties . . . that are reasonably likely to have a material favorable or unfavorable impact on net sales or revenues or income from continuing operations” in SEC filings, such as Form 10-Q and 10-K reports. 17 C.F.R. 229.303(b)(2)(ii). Whether the duty to disclose can serve as a basis for a securities fraud claim under Section 10(b) is unclear, as the Sixth Circuit has never ruled on the issue. The Second, Third, and Eleventh Circuits think it can under certain circumstances; the Ninth Circuit disagrees. See Stratte-McClure v. Morgan Stanley, 776 F.3d 94 (2d Cir. 2015); In re NVIDIA Corp. Sec. Litig., 768 F.3d 1046 (9th Cir. 2014); Carvelli v. Ocwen Fin. Corp., 934 F.3d 1307 (11th Cir. 2019).

Section 10(b) liability typically does not apply to omissions, except where the corporation has a duty to disclose. See Basic, 485 U.S. at 239 n.17. Such a duty might derive from a statute or regulation requiring disclosure. Glazer v. Formica Corp., 964 F.2d 149, 157 (2d Cir. 1992). And Item 303 is a regulation that imposes a duty to disclose, a violation of which the Sixth Circuit has acknowledged can give rise to Section 11 liability. See J & R Mktg., SEP v. Gen. Motors Corp., 549 F.3d 384 (6th Cir. 2008). But, as then-Judge Alito observed, “the materiality standards for Rule 10b-5 and SK-303 differ significantly.” Oran, 226 F.3d at 288. Item 303 imposes a far broader duty to disclose, covering all known trends “reasonably likely to occur,” see Management's Discussion and Analysis of Financial Condition and Results of Operations, Exchange Act Release No. 34-26831, 54 Fed.Reg. 22427, 22430 (May 24, 1989), whereas Rule 10b-5 disclosures are predicated “upon a balancing of both the indicated probability that the event will occur and the anticipated magnitude of the event in light of the totality of the company activity.” Basic, 485 U.S. at 238 (quoting SEC v. Tex. Gulf Sulphur Co., 401 F.2d 833, 849 (2d Cir. 1968) (en banc)). In short, Item 303 looks only to probability (i.e., the likelihood that a known trend will occur), whereas Section 10(b)/Rule 10b-5 looks both to probability and magnitude.

Accordingly, the Third Circuit concluded in Oran that “a violation of SK-303's reporting requirements does not automatically give rise to a material omission under Rule 10b-5.” 226 F.3d at 288. The Ninth Circuit understood this language to preclude Section 10(b) liability based on an Item 303 violation in all circumstances. In re NVIDIA, 768 F.3d at 1054-56. This Court, however, is more persuaded by the Second and Eleventh Circuits' interpretations of the same language. Giving full effect to Oran's use of the word “automatically,” those courts concluded that a violation of Item 303 can give rise to 10b-5 liability in certain, limited circumstances-but not always. Stratte-McClure, 776 F.3d at 103-04; Carvelli, 934 F.3d at 1331 (“On its face, Item 303 imposes a more sweeping disclosure obligation than Rule 10b-5, such that a violation of the former does not ipso facto indicate a violation of the latter.”). Accordingly, this Court adopts the rule set forth in Stratte-McClure: “failure to comply with Item 303 in a Form 10-Q [or Form 10-K] can give rise to liability under Rule 10b-5 so long as the omission is material under Basic, and the other elements of Rule 10b-5 have been established.” 776 F.3d at 103-04.

Having determined that Item 303 can serve as the basis for Rule 10b-5 liability, this Court next turns to the substance of Plaintiffs' claim. They allege that the Upstart Defendants “fail[e]d to disclose the impact to its revenue about the material amounts of [inaccurately priced and riskier] loans accumulating on its balance sheet” in SEC filings. (Compl. ¶ 183, ECF No. 45). They further suggest that the Upstart Defendants failed to disclose the limitations of Upstart's platform and deficient external funding in the IPO Registration Statement. (Id. ¶ 191). But the Upstart Defendants did disclose the accumulation of loans on the company's balance sheet on each Form 10-K and Form 10-Q, and the amount of loans retained. (See Upstart MTD at 23, ECF No. 58-1). Plaintiffs now argue, avoiding the disclosures made in the filings, that the real trend at issue was “why more loans were on the Company's balance sheet” and not just the trend of more loans. (Plaintiffs' Opp'n at 34, ECF No. 61). But there is no indication of this in the Complaint. (See Compl. ¶¶ 183-84, 186, 189, ECF No. 45). Thus, while there are cases where Item 303 can form the basis for a Section 10(b) claim, this is not that case.

d. Sarbanes-Oxley Certifications

The final category of challenged statements consists of the Sarbanes-Oxley (“SOX”) certifications signed by Girouard and Datta on the 2020 and 2021 Forms 10-K. (See Compl. ¶ 193, ECF No. 45). Section 302 of the Sarbanes-Oxley Act of 2002, 15 U.S.C. § 7241, requires public companies to adopt internal controls for ensuring the accuracy of financial statements and the CEOs and CFOs to attest to the effectiveness of the disclosure controls and procedures in SEC filings. Pursuant to that provision, Girouard and Datta signed certifications that they reviewed the Forms 10-K and 10-Q, that the reports did not contain untrue statements of material facts (or omissions of material fact), that the financial statements therein were accurate, and that they had established and maintained proper disclosure controls and procedures. (See, e.g., FY 2021 Form 10-K at 64-67, ECF No. 58-12).

The Complaint fails to plead with particularity the falsity of the SOX certifications. It does not contain allegations that Girouard and Datta failed to evaluate Upstart's internal controls, see In re Gentiva Sec. Litig., 932 F.Supp.2d 352, 570 (E.D.N.Y. 2013), that they were aware of some deficiency or other issue in the company's internal controls but failed to report the issues, see id.; Barrett v. PJT Partners Inc., No. 16-cv-02841, 2017 WL 3995606, at *6 (S.D.N.Y. Sept. 8, 2017), or that they signed on to documents that did not comport with a “fair presentation” of the company's financial information. City of Roseville Emps.' Ret. Sys. v. Horizon Lines, Inc., 686 F.Supp.2d 404, 418-20 (D. Del. 2009). Instead, it simply alleges that the certifications falsely represented that the disclosure controls and procedures were effective without any underlying allegations of problems with the internal controls.

Plaintiffs suggest in their opposition brief that the certifications “were false and misleading because . . . the Company's annual and quarterly reports contained material falsehoods and omissions. Furthermore, the Company's SEC filings that were subject to the SOX certifications did not ‘fairly present[]' Upstart's financial condition.” (Plaintiffs' Opp'n at 24, ECF No. 61). But they never pled such allegations in their complaint. (See Compl. ¶¶ 192-98, ECF No. 45) (focusing only on “the materially false and misleading evaluation of Disclosure Controls and Procedures,” Girouard and Datta's “evaluations of the Company's internal controls,” and false representations about “Upstart's disclosure controls and procedures”).

2. Scienter

This Court next turns to the question of whether Plaintiffs have adequately alleged a strong inference of scienter, i.e., that the Upstart Defendants made the allegedly false or misleading statements with a “‘knowing and deliberate intent to manipulate, deceive, or defraud' or ‘recklessness.'” Astec Indus., 29 F.4th at 812 (quoting Doshi v. Gen. Cable Corp., 823 F.3d 1032, 1039 (6th Cir. 2016)). A strong inference, as explained by the Supreme Court in Tellabs, is one that is “more than merely plausible or reasonable-it must be cogent and at least as compelling as any opposing inference of nonfraudulent intent.” Tellabs, 551 U.S. at 314. But this burden should not be overstated. Plaintiffs need not show that the inference of scienter is more likely than a non-fraudulent inference; “where two equally compelling inferences can be drawn, one demonstrating scienter and the other supporting a nonculpable explanation, Tellabs instructs that the complaint should be permitted to move forward.” Frank v. Dana Corp., 547 F.3d 564, 571 (6th Cir. 2011) (quoting Tellabs, 551 U.S. at 324 n.5; citing ACA Fin. Guar. Corp. v. Advest, Inc., 512 F.3d 46, 59 (1st Cir. 2008)).

In the Sixth Circuit, courts apply a three-part test to answer this question. See Astec Indus., 29 F.4th at 812. First, courts “accept all factual allegations in the complaint as true.” Id. (quoting Tellabs, 551 U.S. at 322). Next, courts review the allegations holistically to determine whether the facts alleged give rise to a strong inference of scienter. See Dougherty v. Esperion Therapeutics, Inc., 905 F.3d 971, 979 (6th Cir. 2018). As part of this holistic review, courts in the Sixth Circuit typically consider the non-exhaustive list of so-called Helwig factors. See generally Helwig, 251 F.3d at 552. And finally, courts “take into account plausible opposing inferences and decide whether a reasonable person would deem the inference of scienter cogent and at least as compelling as any opposing inference one could draw from the facts alleged.” Dougherty, 905 F.3d at 979 (internal quotation marks and citation omitted). Complaints that fail to allege a strong inference of scienter “shall” be dismissed. 15 U.S.C. § 78u-4(b)(3)(A).

Accepting the factual allegations as true, this Court first considers several of the Helwig factors, before weighing other holistic considerations. See also Bond v. Clover Health Invs., Corp., 587 F.Supp.3d 641, 676 (M.D. Tenn. 2022) (“The Helwig factors are an analytical tool for courts, not a pleading requirement.”). In doing so, this Court follows the Upstart Defendants' decision to organize their analysis of the factors on a group basis, rather than on an individual basis. (Upstart MD at 25 n.18, ECF No. 58-1). Where relevant, this Court will also discuss specific allegations that are unique to a particular defendant. See Astec Indus., 29 F.4th at 813; see also Bridgestone, 399 F.3d at 689-90 (noting that the Sixth Circuit has not decided whether group pleading has survived the PSLRA). This analysis is conducted with the context of Plaintiffs' overarching case in mind: that the Upstart Defendants “engaged in a scheme to defraud investors in order to conceal the truth about the Company's platform and business operations,” including flaws with the AI model and Upstart's funding streams. (See generally Compl. ¶ 4, ECF No. 45). The alleged scheme also involves Third Point buying Upstart loans to prop up the appearance of demand and the strategic timing of Upstart's IPO and SPO at a time where the effects of stimulus funding made Upstart's business seem artificially strong.

a. Suspicious Insider Trading

The first factor, the presence of suspicious insider trading, favors Plaintiffs, but only modestly. Three of the Upstart Defendants, Girouard, Gu, and Counselman, collectively sold over $384 million in Upstart stock between August and November 2021 (there is no allegation that Datta sold stock during the Class Period). (Id. ¶ 101, ECF No. 45). These sales began promptly after the end of the standard lock-up period associated with Upstart's IPO, and the majority of the sales had concluded before Upstart's announcement of its Q3 2021 financial results on November 9, 2021, in which it revealed that loans on its balance sheet had increased by 57%. (See id. ¶ 10). The timing of these sales-while Upstart began taking on more loans that it apparently could not move of its books and while its stock was still rising but already close to its peak-carries a whiff of suspicion.

The Upstart Defendants raise several valid counterarguments, but while they undermine the inference of scienter, they fail to negate the inference of scienter or compel the conclusion that a nonfraudulent inference is more compelling than an inference of scienter. First, the Upstart Defendants point to the context of the trades; in particular, the fact that Girouard sold only a small portion of his holdings and Gu was a net acquirer during the Class Period. (See Upstart MTD at 26-27, ECF No. 58-1). Courts have routinely rejected, however, the idea that there is a minimum percentage of stock sales that must be made before an inference of scienter can be drawn, just as the volume of sales alone does not indicate scienter. See In re Cardinal Health, 426 F.Supp. at 728; see also Philip Morris, 75 F.3d at 814-15 (finding that the “retention of a large position . . . gives rise to an inference that [the defendant] engaged in insider trading”); Nursing Home Pension Fund, Loc. 144 v. Oracle Corp., 380 F.3d 1226, 1232 (9th Cir. 2004) (“[W]here, as here, stock sales result in a truly astronomical figure, less weight should be given to the fact that they may represent a small portion of the defendant's holdings.”). Similarly, courts have rejected the idea that stock purchases negate an inference of scienter without information about the context of the purchases, like “whether the officers purchasing the stock were required to exercise options at that time.” In re Green Tree Fin. Corp. Stock Litig., 61 F.Supp.2d 860, 868 n.8 (D. Minn. 1999), rev'd sub nom., Fla. St. Bd. of Admin. v. Green Tree Fin. Corp., 270 F.3d 645 (8th Cir. 2001). Thus, the fact that Gu was a net acquirer does little to negate the inference of scienter.

While securities fraud plaintiffs are typically required to provide “meaningful trading history” as context to demonstrate the unusualness of allegedly suspicious trades, see Konkol v. Diebold, Inc., 590 F.3d 390, 399 (6th Cir. 2009), that would be impossible where the trades occurred immediately after an IPO. There is no trading history to be provided in such cases.

This need for context is applicable broadly when it comes to drawing inferences from stock sales. For example, while some courts find insider stock sales not suspicious if the shares were sold below the peak of the share price, see, e.g., Greebel v. FTP Software, Inc., 194 F.3d 185, 206 (1st Cir. 1999) (concluding that stock sales were not “very suspicious” because they were not made at the peak of the share price); In re Hertz Glob. Holdings Inc., 905 F.3d 106, 120 (3d Cir. 2018) (same), other courts have observed that the alleged insider trading need not be perfectly optimized to be problematic. In re Am. Int'l Grp., Inc., 965 A.2d 763, 801 (Del. Ch. 2009) (“But it is not a defense that [insiders] could have committed an even larger breach of their fiduciary duties.”). Similarly, insiders will sometimes “hold on to a portion of [their] shares to hedge against the unforeseen or to obscure the insider trading from the SEC.” In re Worlds of Wonder Sec. Litig., 35 F.3d 1407, 1427 (9th Cir. 1994). Although Girouard, Gu, and Counselman did not maximize their profits by selling all of their stock at once or at the peak of Upstart's stock price, the context necessary to understand fully their trading patterns is missing. As such, this Court cannot conclude, based only on the small percentage of stocks sold, the net acquisition of stocks, or the sales at less-than-peak share price, that the nonfraudulent inference is more compelling than the inference of scienter.

The same reasons lead this Court to reject the Upstart Defendants' Rule 10b-5 trading plan argument. (See Upstart MTD at 27-28, ECF No. 58-1). Trading plans involve corporate insiders setting up predetermined sales of shares they own far in advance of the actual sale, and are intended to avoid allegations of insider trading. The Upstart Defendants suggest that all of the challenged transactions made by Girouard, Gu, and Counselman were executed pursuant to Rule 10b-5 trading plans that they entered into in May 2021. Several courts have concluded that trading plans do “not completely immunize [a corporate insider] from suspicion” but do “mitigate any inference of improper motives surrounding his sales.” Yates v. Mun. Mortg. & Equity, LLC, 744 F.3d 874, 891 (4th Cir. 2014); see also Elam v. Neidorff, 544 F.3d 921, 928 (8th Cir. 2008); Metzler Inv. GmbH v. Corinthian Colls., Inc., 540 F.3d 1049, 1067 n.11 (9th Cir. 2008). Such a conclusion would be hasty at this juncture, however, where this Court has not been presented with the trading plans and has not had an opportunity to review their contents. See Azar v. Yelp, Inc., No. 18-cv-00400, 2018 WL 6182756, at *18-19 (N.D. Cal. Nov. 27, 2018). This is, again, an issue of context. A trading plan provides an affirmative defense only where “the plan was ‘entered in good faith'”-a determination that this Court cannot make on a motion to dismiss. Stocke v. Shuffle Master, Inc., 615 F.Supp.2d 1180, 1193 (D. Nev. 2009) (quoting 17 C.F.R. § 240.10b5-1(c)(1)(i)); cf. Ind. Pub. Ret. Sys. v. Pluralsight, 45 F.4th 1236, 1266 (10th Cir. 2022) (“The text and history of Rule 10b5-1 shows that such plans can be manipulated easily for personal financial gain.”). Accordingly, the existence of Rule 10b5-1 plans that have not been introduced into the record does not negate the inference of scienter at this early stage of the litigation.

b. Divergence Between Internal Reports and External Statements

The second factor asks whether there was a “divergence between internal reports and external statements on the same subject.” Helwig, 251 F.3d at 552. This factor has been deemed the “key factor,” Bridgestone, 399 F.3d at 688, because “it cannot be said that [corporate insiders] perpetrated a fraud . . . in not disclosing these facts . . . if [they] did not know of the inaccurate reporting.” Albert Fadem Tr. v. Am. Elec. Power Co., Inc., 334 F.Supp.2d 985, 1006 (S.D. Ohio 2004). Typically, security fraud plaintiffs point to “internal reports, memoranda, or the like” that corporate insiders possessed at the time they made the challenged statements, In re Bos. Tech., Inc. Sec. Litig., 8 F.Supp.2d 43, 57 (D. Mass. 1998), though courts should not rely on an overly “formalistic definition” of what constitutes “internal reports.” Dougherty, 905 F.3d at 981.

Consideration of this factor is premature at this stage, given that “Plaintiffs would need discovery to discern the contents of internal reports as compared to external statements.” In re FirstEnergy Corp., 2022 WL 681320, at *20. It certainly does not favor Plaintiffs: they have identified no particular documents that contradict the challenged statements. Instead, they simply allege that the Upstart Defendants had real-time information about changes to the company's AI underwriting model and to negative information about the loans on the balance sheet, including the soaring amount of loans-but identify no documents (let alone internal reports) with such information. See Bridgestone, 399 F.3d at 688-69 (identifying internal tests, data reports, and customer lawsuits that defendants had in their possession); Dougherty, 905 F.3d at 981 (FDA minutes and meeting notes); Astec Indus., 29 F.4th at 813 (inspection reports and conference call notes). In short, their argument relies entirely on speculation. It assumes that Plaintiffs must have had access to some sort of information about the AI model and about the accumulation of loans. That may well be true, but it is clearly lacking in specificity.

c. Closeness in Time Between Challenged Statements and Later Corrective Disclosures

The third factor cuts in favor of Plaintiffs, but only narrowly and only with respect to statements by Girouard on September 9, 2021, March 9, 2022, March 21, 2022, and May 10, 2022. (Compl. ¶¶ 197, 200-01, 212-13, ECF No. 45). The alleged corrective disclosures identified by Plaintiffs include: Upstart's Q3 2021 Form 10-Q, which was filed on November 9, 2021, and disclosed a substantial increase in the amount of loans on its balance sheet; a press release on May 9, 2022, disclosing another increase in loans on its business sheet; and an announcement on July 7, 2022, that it had converted loans on its balance sheet to cash because of the increasing rate environment. (Compl. ¶¶ 103, 112, 252, ECF No. 45; see Plaintiffs' Opp'n at 40-41, ECF No. 61). The statements by Girouard are all within two months of one of these three disclosures. Two months is at the outer edge of what has been considered “close” for the third Helwig factor: the Sixth Circuit has “found that a one-week span between an allegedly fraudulent statement and a subsequent inconsistent disclosure was probative of scienter, but [] rejected the same inference when confronted with a four-month gap.” Dougherty, 905 F.3d at 981 (citation omitted); see also id. (finding that a “six-week gap” weighed in favor of the plaintiff). This suggests that the twomonth gap found in this case favors the plaintiff, though only narrowly.

d. Disregard of Most Current Factual Information

The sixth factor favors Plaintiffs, who have alleged that Girouard and Datta knew that demand for loans was not “there,” that the accumulation of loans was not purely for R&D purposes, and the loan problem was not “temporary,” even while making statements to the contrary. This is corroborated by, among other things, the fact that loans continued to accumulate rapidly during the time periods when Girouard and Datta were reassuring investors that there was no issue or that the issue was temporary.

e. Self-Interested Motivation of Defendants to Save their Salaries or Jobs

The ninth Helwig is implicated only where the defendants have specific motives to commit fraud, and not just “motives common to [all] corporations and executives generally” like “a desire for their companies to appear successful.” PR Diamonds, Inc. v. Chandler, 364 F.3d 671, 690 (6th Cir. 2004). Thus, “an executive's desire to protect his position within a company or increase his compensation” does not count as a specific motive for fraud, id. (citing Kalnit v. Eichler, 264 F.3d 131, 140 (2d Cir. 2001); In re Criimi Mae, Inc. Sec. Litig., 94 F.Supp.2d 652, 660 (D. Md. 2000)), whereas compensation incentives tied directly to share price like performance-based pay do count. See In re Cardinal Health, 426 F.Supp. at 751; In re First Energy, 2022 WL 681320, at *21.

Here, however, Plaintiffs have not provided any specific to commit fraud. They allege only that Girouard, Counselman, and Gu's compensation was tied to Upstart's success, but this is exactly the kind of “generalized motive . . . [that] could be imputed to [corporate officers at] any publicly-owned, for-profit endeavor,” which does not meet the ninth Helwig factor. PR Diamonds, 364 F.3d at 690 (quoting Chill v. Gen. Elec. Co., 101 F.3d 263, 268 (2d Cir. 1996)); Boykin v. K12, Inc., 54 F.4th 175, 186 (4th Cir. 2022) (same). Plaintiffs have not, on the other hand, alleged that Girouard, Counselman, or Gu's compensation was tied to performance-based incentives, or that there were other extraordinary circumstances suggestive of a specific motive to commit fraud. See In re FirstEnergy, 2022 WL 681320, at *21.

f. Remaining Helwig Factors

The fourth, fifth, and eighth factors-evidence of bribery, the quick settlement of an ancillary lawsuit alleging fraud, and attempts to hide stock sales from disinterested directors, respectively-are inapplicable. Plaintiffs have not presented allegations suggesting that any of the above are present. Because the Helwig factors are non-exhaustive, the absence of these factors “does not meaningfully detract from a strong inference of scienter.” Strougo v. Tivity Health, 551 F.Supp.3d 839, 851 (M.D. Tenn. 2021). The seventh factor, on the other hand, which considers whether Plaintiffs have alleged that the Upstart Defendants disclosed accounting information in such a way that only the sophisticated could understand, cuts against Plaintiffs. Here, the Upstart Defendants disclosed the accumulation of loans on its balance sheet on a quarterly basis (including breakdowns of the number of loans held for sale or for investment). The transparency of these disclosures undermines the inference of scienter.

g. Other Holistic Considerations

Finally, this Court notes some additional considerations relevant to the holistic analysis. The challenged statements cover topics that the Upstart Defendants voluntarily and repeatedly discussed with the public, which “demonstrates [their] sensitivity” towards the issues. Gauquie v. Albany Molecular Rsch., Inc., No. 14 CV 6637, 2016 WL 4007591, at *2 (E.D.N.Y. July 26, 2016) (citing Reese v. Malone, 747 F.3d 557, 576 (9th Cir. 2014)). The statements were often made in answer to questions from investors, which implies an intent to “emphasize and induce reliance upon” the statements. Casella, 883 F.2d at 808. And the statements were made about core components of Upstart's business like the AI model and the nature of the business model that the Upstart Defendants would have been intimately familiar with, which also favors a finding of scienter. In re Huffy Corp. Sec. Litig., 577 F.Supp.2d 968, 1000 (S.D. Ohio 2008) (“[T]he more central a fact is to a company's core operations the more likely its executive acted with scienter.” (citation omitted)); In re Cardinal Health, 426 F.Supp. at 724 (same); see also In re First Energy, 2022 WL 681320, at *20 (noting that the core operations doctrine “aid[s], but not alone determin[es], an inference of scienter”). This is especially applicable here, given Gu and Datta's oversight of the teams charged with refining the AI model and monitoring the company's balance sheet, respectively, and the fact that investors and analysts were fixating on these issues. See In re First Energy, 2022 WL 681320, at *20 (noting that investors and analysts' focus on certain issues “supports scienter for the executives charged with overseeing said operations”).

On the other hand, this Court is not particularly persuaded by the Upstart Defendants' argument that the lack of “multiple, obvious red flags,” or “confidential witnesses, [] accounts of conversations or meetings, and [] internal documents or reports” shows decisively that Plaintiffs have failed to allege a strong inference of scienter. (Upstart MTD at 35, ECF No. 58-1). As a threshold matter, Plaintiffs are not required to plead “an egregious refusal to see the obvious, or to investigate the doubtful,” which are signs of recklessness, where they have pled actual knowledge, as they do in this case. PR Diamonds, 364 F.3d at 693. The inclusion of internal reports or the recounting of conversations requires a degree of omniscience that is unrealistic at the preliminary stages of the litigation process. And there are plenty of “red flags” in Plaintiffs' pleadings, such as the fact that Datta and Girouard continued to claim that demand for loans was robust even as Upstart was forced to hold on to more and more loans on its own balance sheet-before eventually acknowledging defeat in converting loans into cash at a loss.

Viewing Plaintiffs' allegations collectively, this Court finds that a “reasonable person” would “deem the inference of scienter at least as strong as any opposing inference.” Tellabs, 551 U.S. at 326. The suspiciously timed insider sales, the disregard for the most current factual information, and the fact that the challenged statements go to the core of Upstart's business model and were intended to reassure investors all point in favor of a strong inference of actual knowledge. The opposing inference is compelling too, given the lack of divergence between the challenged statements and internal reports. But it is not stronger than the inference of scienter, at least on the preliminary record currently before this Court.

For certain statements made by the Upstart Defendants, Plaintiffs have made a sufficient threshold showing to warrant discovery and should be permitted to test their inferences against a more developed factual record. As such, the following statements by the Upstart Defendants are actionable: (1) specific, material, and verifiable statements about the superiority of Upstart's AI underwriting model over traditional, FICO-based models; (2) statements about the AI model's ability to adapt quickly to changing macroeconomic conditions and its rate-insensitivity; (3) the November 9, 2021, and February 15, 2022, statements that demand for Upstart loans is “there”; (4) the November 9, 2021, and February 15, 2022, statements that loans were only being retained for R&D purposes; and (5) statements that the loan problem was a temporary concern. (Compl. ¶¶ 96, 125-26, para. 1 of statement within 132, 141, 145, 161, 197, 200-01, 203, 212-14, 225, 241, ECF No. 45). Therefore, the Motion to Dismiss as to the preceding statements is Denied. The Motion to Dismiss as to claims alleging that any other statements were false or misleading in violation of federal securities laws is Granted and those claims are Dismissed.

C. Scheme Liability (Count I)

Plaintiffs also allege that both the Upstart Defendants and the Third Point Defendants violated the scheme liability provision of Rule 10b-5. A scheme liability claim requires the plaintiff to show: “(1) that the defendant committed a deceptive or manipulative act, (2) in furtherance of the alleged scheme to defraud, (3) with scienter, and (4) reliance.” Plumber & Steamfitters Loc. 773 Pension Fund v. Danske Bank A/S, 11 F.4th 90, 105 (2d Cir. 2021) (citation omitted). The alleged “scheme” “may rest in part on the same statements or omissions that trigger misstatement liability, or it may embrace separate statements or conduct.” In re First Energy, 2022 WL 681320, at *7 (citing Lorenzo v. SEC, 139 S.Ct. 1094, 1102 (2019)). Here, Plaintiffs claim that Defendants engaged in a scheme to defraud investors involving the promulgation of false and misleading statements about Upstart's business in conjunction with other strategic actions, which allowed Upstart to sell shares at artificially inflated prices during its IPO and SPO (and the Individual Defendants and the Third Point Defendants to do the same post-IPO). (See Compl. ¶ 56, ECF No. 45; TP Opp'n at 9, ECF No. 62).

The Upstart Defendants argue that the scheme liability claim should be dismissed for the reasons as the underlying misstatement claim. (See Upstart MTD at 36-37, ECF No. 58-1). But this Court has concluded that the underlying misstatement claim should not be dismissed in its entirety. See supra Part IV.B. As such, the Upstart Defendants' argument is without merit.

The scheme liability claim against the Third Point Defendants requires a bit more explanation. In their complaint, Plaintiffs suggest that the scheme consisted of “disseminating materially false and misleading statements and/or concealing material adverse facts.” (Compl. ¶ 56, ECF No. 45). Plaintiffs now explain that the scheme involved not just misstatements, but also “a variety of devices” to prop up demand for Upstart loans and paint a false perception of the nimbleness of the AI model. (See TP Opp'n at 6-7, ECF No. 62). These “devices” apparently include inter alia the Third Point Defendants' participation on Upstart's board, the purchase by Third Point Credit (an affiliate of Third Point) of hundreds of millions of dollars' worth of Upstart loan products from 2017 to 2021, and the Third Point Defendants' efforts to help Upstart navigate capital markets. (See id. at 7-8). But this version of the alleged scheme goes far beyond the simple “dissemination of misstatements” scheme that is alleged in the complaint. See United States ex rel. Marlar v. BWXT Y-12, LLC, 525 F.3d 439, 444-45 (6th Cir. 2008) (“Where a complaint alleges a complex and far-reaching fraudulent scheme, then that scheme must be pleaded with particularity and the complaint must also provide examples of specific fraudulent conduct that are representative samples of the scheme.” (cleaned up)). Plaintiffs' attempt to explain that they simply pled a “representative sample” of the scheme beggars belief. It is difficult to see how “disseminating misstatements” is “representative” of “purchasing loans,” for example, or of “giving advice about capital markets.”

Even if this Court read the Complaint as broadly as Plaintiffs would like, it still does not allege a viable scheme liability claim against the Third Point Defendants. The Third Point Defendants, as explained earlier, did not make any of the challenged statements, nor are they alleged to have disseminated the statements. See supra Part IV.B. As to the loan purchases and the capital markets advice, these are alleged to have been “acts in furtherance of the scheme” but were not themselves deceptive or manipulative. There is no allegation that the loan purchases were fake, for example, that the advice was a sham, or that the Third Point Defendants made false or misleading misrepresentations in connection with their actions. See In re Parmalat Sec. Litig., 376 F.Supp.2d 472, 505 (S.D.N.Y. 2005) (“These transactions were not shams. Nor did they depend on fictions....The arrangements therefore were not inventions, projects, or schemes with the tendency to deceive.”); In re Enron Corp. Sec., Derivative & ERISA Litig., 439 F.Supp.2d 692, 720 (S.D. Tex. 2006). Plaintiffs effectively argue that, by purchasing loans, the Third Point Defendants aided and abetted the Upstart Defendants' misrepresentations about the robustness of demand for the loans.

For example, there is nothing deceptive about strategically timing an IPO. Companies do so all the time. Any deception lies in what statements are made in conjunction with the IPO, but not with choosing an optimal time to go public. In fact, failing to pick an opportune time to go public, i.e., selling greater control of the corporation for less capital raised, would be far more worrisome.

But aiding and abetting does not provide a basis for private liability under Section 10(b)/Rule 10b-5. See Central Bank of Denver, N.A. v. First Interstate Bank of Denver, N.A., 511 U.S. 164 (1994); cf. Stoneridge Inv. Partners, LLC v. Sci.-Atlanta, Inc., 552 U.S. 148, 162 (2008) (noting that “[a]iding and abetting liability is authorized in actions brought by the SEC but not by private parties”). This is true even if the Third Point Defendants purchased the loans while “knowing or even intending that [Upstart] would misrepresent the nature of the arrangements” and assert that loan demand was “there.” In re Parmalat, 376 F.Supp.2d at 505; see also SEC v. Rio Tinto plc, 41 F.4th 47, 55 (2d Cir. 2022) (rejecting “a widened scope of scheme liability” against aiders and abettors because to do so “would defeat the congressional limitation on the enforcement of secondary liability, multiply the number of defendants subject to private securities actions, and render the statutory provision for secondary liability superfluous” (citing 15 U.S.C. § 78t(e))). As such, Plaintiffs' reconstructed allegations, suggesting only that the Third Point Defendants aided and abetted the deceptive scheme and not that their actions were themselves deceptive, fail to state a claim under Rule 10b-5(a) and (c). The scheme liability claim is therefore Dismissed as to the Third Point Defendants.

D. Control Liability (Count II)

Section 20(a) of the Exchange Act allows plaintiffs to hold liable “[e]very person who, directly or indirectly, controls any person liable under any provision of this chapter or of any rule or regulation thereunder.” 15 U.S.C. § 78t(a). A plaintiff asserting a claim pursuant to Section 20(a) must show that the controlled individual committed “an underlying violation of the securities laws or the rules and regulations promulgated thereunder” and that “the ‘controlling person' defendant . . . directly or indirectly controlled the person liable for the securities law violation.” PR Diamonds, Inc., 364 F.3d at 696 (citation omitted). “Control” is defined by regulation as “the possession, direct or indirect, of the power to direct or cause the direction of the management and policies of a person, whether through the ownership of voting securities, by contract, or otherwise.” 17 C.F.R. § 230.405. Courts have read this provision as requiring plaintiffs to plead that the defendant had “the practical ability to direct the actions of the people who committed the violation.” In re Nat'l Century Fin. Enters., Inc., 504 F.Supp.2d 287, 300 (S.D. Ohio 2007) (internal quotation marks and citation omitted). Some courts further require that plaintiffs prove the actual exercise of control; the Sixth Circuit has not spoken on whether this is required. See also In re FirstEnergy, 2022 WL 681320, at *30.

In Sanders Confectionary Prods., Inc. v. Heller Fin., Inc., the Sixth Circuit cited approvingly the Eighth Circuit's test, which requires a plaintiff show both the authority to control and actual control. 973 F.2d 474, 486 (6th Cir. 1992) (citing Metge v. Baehler, 762 F.2d 621, 631 (8th Cir. 1985)). As this Court once noted:

In Sanders, the Sixth Circuit described the standard applied in Metge as the “least rigorous” one available. Since Sanders was decided, courts in the Fifth and Ninth Circuits have applied a less rigorous standard. Though other district courts in this circuit view Sanders as approving the Metge standard, the Sixth Circuit expressly declined to adopt a standard for controlling person liability. This Court does not view Sanders as adopting the Metge standard. Rather, the Sixth Circuit applied it because it was the least rigorous standard used at that time.

In re Nat'l Century, 504 F.Supp.2d at 303..

In seeking to dismiss this claim, the Upstart Defendants argue that Plaintiffs have failed to plead adequately a predicate violation of federal securities laws and regulations. The Third Point Defendants adopt that argument, and further suggest that Plaintiffs' allegations of control are inadequate. (See Upstart MTD at 37, ECF No. 58-1; TP MTD at 13-17, ECF No. 60-1). The first argument can be dealt with easily: as discussed above, Plaintiffs have adequately pled that the Upstart Defendants violated Section 10(b) of the Exchange Act, which satisfies the “underlying violation” requirement of a control liability claim.

As to the second argument, the Third Point Defendants challenge both the premise that they possessed the power to control Upstart's actions and the proposition that they exercised that power. The following facts, according to Plaintiffs, show that the Third Point Defendants had the power to control Upstart: First, that Third Point held almost 20% of Upstart's stock; Second, that Schwartz sat on Upstart's board; and Third, that Third Point was actively involved in Upstart's organizational development, staffing and talent acquisition, and go-to-market and capital market strategies. (See Compl. ¶¶ 44, 67-73, ECF No. 45). But it is difficult to see how these allegations give rise to a “reasonable inference” of control. Ownership of a minority of stock interest in a company has rarely been deemed sufficient for control liability, especially when the stake is as low as 20%. See In re Alstom SA Sec. Litig., 406 F.Supp.2d 433, 492 (S.D.N.Y. 2005) (“Minority stock ownership and the ability to appoint a minority of the board do not create power to direct management and policies, and thus do not constitute sufficient control under Section 20(a).” (citations omitted)); see, e.g., In re Flag Telecom Holdings, Ltd. Sec. Litig., 352 F.Supp.2d 429, 458-59 (S.D.N.Y. 2005) (finding 30% of voting shares insufficient for control liability); In re Nat'l Century Fin. Enters., Inc. Fin. Invest. Litig., 553 F.Supp.2d 902, 912-13 (S.D. Ohio 2008) (same for 20.76% ownership); Andrew v. Primus Telecomm. Grp., Inc., 107 Fed.Appx. 301, 306 (4th Cir. 2004) (20%). Nor does Third Point's right to appoint a single director to Upstart's board of eight directors move the needle. See In re Flag Telecom, 352 F.Supp.2d at 458-59 (no control liability against defendant who appointed three of nine board members).

Finally, the Third Point Defendants' involvement in Upstart's managerial affairs, while substantial, cannot be classified as “functional intertwinement” or “direct and supervisory involvement in [] day-to-day operations.” See In re FirstEnergy, 2022 WL 681320, at *31; In re FirstEnergy, 316 F.Supp.2d at 600-01. Plaintiffs' allegations suggest only that Third Point “delivered value” on Upstart's strategy and operations, purchased loans, and had “intimate knowledge of their management teams.” (Compl. ¶¶ 69-70, ECF No. 45; see also id. ¶ 73). Vague corporate phrases like “delivering value” do not demonstrate “heavy involvement,” let alone so close a degree of involvement that Third Point could exercise control, direct or indirect, over Upstart's operations or strategic decisions. Cf. In re First Energy, 2022 WL 681320, at *32 (sustaining control liability claim based on “direct oversight and responsibility”).

Because Plaintiffs have failed to plead allegations giving rise to a “reasonable inference” that the Third Point Defendants possessed “the power to direct or cause the direction of the management and policies of” the Upstart Defendants, Iqbal, 556 U.S. at 678; 17 C.F.R. § 230.405, the control liability claim is Dismissed as to the Third Point Defendants.

E. Contemporaneous Trading Liability (Count III)

This Court turns at last to Plaintiffs' claim for contemporaneous trading liability pursuant to Section 20A of the Exchange Act, which “provides a private right of action to buyers of securities who buy ‘contemporaneously' with anyone who purchases or sells a security while in possession of material, nonpublic information.” Beach v. Healthways, Inc., No. 2:08-0569, 2009 WL 650408, at *5 (M.D. Tenn. Mar. 9, 2009) (citing 15 U.S.C. § 78t-1(a)). A plaintiff alleging a violation of § 20A must: “(1) allege a requisite independent, predicate violation of the Exchange Act (or its rules and regulations), e.g., § 10(b), and (2) show that he has standing to sue under § 20A because he ‘contemporaneously with the purchase or sale of securities that is the subject of such violation has purchased . . . or sold . . . securities of the same class' as the insider defendant.” In re Enron Corp. Sec., Derivative & ERISA Litig., 258 F.Supp.2d 576, 599 (S.D. Tex. 2003) (quoting 15 U.S.C. § 78t-1(a)). What is considered a “contemporaneous” trade has been the subject of vigorous debate; the term is not defined by statute. See id. at 600-01 (noting the lack of “clear agreement”). “Different courts have found that ‘contemporaneity' requires the insider and the investor/plaintiff to have traded anywhere from on the same day, to less than a week, to within a month, to ‘the entire period while relevant and nonpublic information remained undisclosed.'” Id. (collecting cases).

This Court need not wade into that debate today. The Upstart Defendants challenge the contemporaneous trading claim not for a lack of contemporaneous trading, but on the same grounds that they make against the Section 10(b) claim. (See Upstart MTD at 37, ECF No. 58-1) (incorporating by reference their arguments that Plaintiffs failed to allege a predicate violation of the Exchange Act). But this Court has not dismissed Plaintiffs' Section 10(b) claim (at least not entirely). Put differently, this Court has not dismissed the alleged predicate violation of the Exchange Act underlying the Section 20A claim against the Upstart Defendants. The Upstart Defendants' arguments that the Section 20A claim fails for lack of a predicate violation is not well taken. See Ross v. Abercrombie & Fitch Co., 501 F.Supp.2d 1102, 1120 (S.D. Ohio 2007).

On the other hand, this Court has concluded that Plaintiffs failed to plead adequately a securities fraud claim under § 10(b) and Rule 10b-5 against the Third Point Defendants (or any other claims under the securities laws or the rules and regulations thereunder). See supra Part IV.B-D. Without a predicate violation, Plaintiffs' Section 10A claim against the Third Point Defendants fails.

V. CONCLUSION

For the reasons set forth herein, this Court Grants in part and denies in part plaintiffs' motion to strike (ecf no. 65). the forms 13f are therefore stricken. this court also grants in part and denies in part the upstart Defendants' Motion to Dismiss (ECF No. 58) and Grants the Third Point Defendants' Motion to Dismiss (ECF No. 60). All claims against the Third Point Defendants are Dismissed, as are claims of securities fraud pursuant to § 10(b) and Rule 10b-5 against the Upstart Defendants except those based on the following statements: (1) statements about the advantages of Upstart's AI underwriting model over traditional, FICO-based models; (2) statements about the ability of the model to “quickly adjust” to changing macroeconomic conditions; (3) the November 9, 2021, and February 15, 2022, statements that demand for Upstart loans is “there”; (4) the November 9, 2021, and February 15, 2022, statements that loans were only being retained for R&D purposes; and (5) statements that the loan problem was a temporary concern.

IT IS SO ORDERED.

In re Upstart Holdings Sec. Litig.

Summary

Opinion

In re Upstart Holdings Sec. Litig.

In re Upstart Holdings Sec. Litig.

Case Details

Citations

Citing Cases

In re Upstart Holdings Sec. Litig.

Summary

Opinion

In re Upstart Holdings Sec. Litig.

In re Upstart Holdings Sec. Litig.

Case Details

CitationsCopy Citation

Citing Cases

Citations