Musk's Lawyers Repeatedly Accused Twitter of Failure to Disclose Data in Failed Attempt to Back Out

Feature Image: HackerNoon’s Midjourney AI, Prompt “to hell with data privacy”

FACTUAL ALLEGATIONS

VI. Musk grasps for an out

B. Defendants’ lawyer letters

82. Even as Musk was violating his own contractual obligations, Twitter continued to respond cooperatively to his representatives’ increasingly unreasonable inquiries. Between May 16 and May 20, the company provided detailed written responses to several information requests.

83. On May 20, 2022, Musk’s team sent a request for Twitter’s “firehose” data — which is essentially a live-feed of data concerning activity (Tweeting, Retweeting, and “liking” Tweets, for example) associated with the public accounts on Twitter’s platform. Again, no explanation was offered for how this request furthered a “reasonable business purpose related to the consummation of the transactions contemplated by” the merger agreement, as required by Section 6.4. Nor can the firehose data even be used to accurately estimate the prevalence of spam or false accounts. As Agrawal had explained in his May 16 Tweets, that estimate depends in part on private data not available in the firehose. Conversely, the firehose includes Tweets that Twitter’s systems and processes catch and do not count within mDAU for that day.

84. On May 21, 2022, Twitter hosted a third diligence session with Musk’s team and yet again discussed Twitter’s processes for calculating mDAU and estimates of spam or false accounts. Twitter also provided a detailed summary document describing the process the company uses to estimate spam as a percentage of mDAU.

85. Defendants responded with increasingly invasive and unreasonable requests. And rather than use “reasonable best efforts to minimize any disruption to the respective business of the Company and its Subsidiaries that may result from requests for access,” Ex. 1 § 6.4, defendants repeatedly demanded immediate responses to their access requests. The scope of the requests and the deadlines defendants imposed on their satisfaction were unreasonable, disruptive to the business, and far outside the bounds of Section 6.4.

86. Twitter nonetheless continued to work with Musk to try to respond to the requests. It extended an ongoing offer to engage with Musk and his representatives regarding its calculation of mDAU, and held several more diligence sessions through the end of May. It also provided detailed written responses, including custom reporting, to his escalating requests for information.

87. On May 25, 2022, defendants’ counsel sent the first of a series of aggressive letters copying their litigation counsel at Quinn Emmanuel. This one falsely asserted that Twitter had “failed to respond to any” of defendants’ information requests and insisted that defendants be granted access to the firehose data so Musk could “make an independent assessment of the prevalence of fake or spam accounts on Twitter’s platform.” Though the letter called Twitter’s own spam detection methodologies “lax,” it identified no basis for that charge.

88. Nor, again, did defendants explain how fulfillment of the firehose data demand would further consummation of the merger or what basis they had to demand the right to “make an independent assessment” of the prevalence of false or spam accounts on the platform. Even assuming that was a proper purpose, reviewing the full firehose data would not result in an accurate assessment or mimic the rigorous process that Twitter employs by sampling accounts and using public and private data to manually determine whether an account constitutes spam — as Twitter’s representatives had already repeatedly explained to Musk’s team.

89. On May 27, 2022, Twitter responded by noting its weeks-long active engagement with Musk’s team and explaining that some of defendants’ requests sought disclosure of highly sensitive information and data that would be difficult to furnish and would expose Twitter to competitive harm if shared. After all, Musk had said he would do one of three things with Twitter: sit on its board, buy it, or build a competitor. He had already accepted and then rejected the first option, and was plotting a pretextual escape from the second. Musk’s third option — building a competitor to Twitter — remained. Still, Twitter again responded constructively and reiterated its commitment to work with Musk’s team to provide reasonable access to requested information.

90. On May 31, 2022, defendants lobbed another missive, again falsely asserting that Twitter had “refused” to provide requested data and that the company’s spam or false account detection methods were “inadequate.” The letter claimed Musk was willing to implement protocols to protect against “damage or competitive harm to the company.”

91. On June 1, 2022, Twitter responded by refuting that it had “refused” provision of data, demonstrating that, to the contrary, it had been working with Musk’s team to honor their requests within the bounds of the contract. To help set the protocols Musk had said he was willing to honor, Twitter asked a series of questions directed at how the data would be used and by whom, and how it would be protected.

92. Defendants’ response on June 6, 2022 made no effort to answer those questions or identify data-protection protocols; instead, it accused Twitter of breach and advanced a false narrative that Twitter had been stonewalling Musk’s requests. Musk publicly filed the letter, which repeated his baseless and damaging charge that Twitter had “lax” detection methods. He included none of Twitter’s correspondence in that filing and omitted all details about the information Twitter had provided. He thus continued to present the public with a misleadingly incomplete narrative about his communications with Twitter, with equally misleading implications about the likelihood that the merger would be completed and about Twitter’s operations.

93. Steadfast in its commitment to consummate the merger, Twitter continued to try to get Musk’s team what it demanded while safeguarding its customers’ data and harboring very real concerns about how Musk might use the data if he succeeded in escaping the deal. On or about June 9, 2022, Musk’s counsel indicated that granting access to 30 days’ worth of historical firehose data would satisfy Musk’s request for the firehose data. So, on June 15, the company gave Musk’s team secure access to that raw data — about 49 tebibytes’ worth. It did so even though the merger agreement did not require the sharing of this information.

94. Musk’s next lawyer letter, dated June 17, 2022, skimmed over this massive data production. Like the earlier correspondence, the June 17 letter described an alternative reality in which Twitter had failed to cooperate in supplying Musk with information, entirely contrary to the facts, apparently in the belief that repeating a falsehood enough can make it true. The letter also continued to move the goal posts by adding a new request for “the sample set” and “calculations” Twitter used to estimate that fewer than 5% of its mDAUs are false or spam accounts over the past eight quarters. Thus, with no basis, defendants sought to audit information Twitter consistently had caveated as an “estimate” requiring “significant judgment” to prepare.

95. The June 17 letter further contained a litigation-style discovery demand for information Musk asserted was needed to investigate “the truthfulness of Twitter’s representations to date regarding its active user base, and the veracity of its methodologies for determining that user base.” It broadly demanded board materials relating to mDAU and spam, as well as emails, text messages, and other communications about those topics — highly unusual requests in the context of good faith efforts toward completion of any merger transaction, and absurd in the context of this one, which has no diligence condition. Musk propounded these unreasonable requests and touted his contrived narrative about Twitter’s methodologies, all without ever identifying a basis for questioning the veracity of Twitter’s methodologies or the accuracy of its SEC disclosures.

96. On June 20, 2022, Twitter set the record straight in a detailed response letter. It noted that the two sides had been working collaboratively to clear regulatory hurdles and “address voluminous data requests” from defendants, that Twitter had “dedicated significant resources” to providing defendants with the data requested, and that Twitter had already provided a wealth of data sweeping far beyond the bounds of what might conceivably be deemed reasonably necessary to consummate the transaction. Twitter noted that Musk, while continuing to accuse Twitter of misrepresenting its spam or false account estimate, had offered not a single fact to support the accusation. And Twitter observed that defendants’ “increasingly irrelevant, unsupportable, and voluminous information requests” appeared directed not at consummating the merger but rather the opposite: trying to avoid the merger.

97. Nonetheless, in a continuing effort at cooperation, Twitter agreed to provide Musk everything he now demanded regarding the firehose, including access to “100% of Tweets and favoriting activity.” Twitter cautioned, as it had so many times before, that this data would not allow Musk to accurately assess the number of spam or false accounts. But on June 21, 2022, it gave defendants’ counsel the demanded access.

98. Meanwhile, Agrawal and Twitter CFO Ned Segal had been trying to set up a meeting with Musk to discuss the company’s process in estimating the prevalence of spam or false accounts. On June 17, 2022, Segal proposed a discussion with Musk and his team to “cover spam as a % of DAU.” Musk responded that he had a conflict at the proposed time. When Agrawal sought to reengage on the matter, Musk agreed to a time on June 21, but then bowed out and asked Agrawal and Segal to speak with his team not about the spam estimation process but “the pro forma financials for the debt.”

99. On June 29, 2022, Musk complained through counsel that Twitter purportedly had “placed an artificial cap on the number of searches” Musk’s experts could run on the firehose data, and had failed to respond to certain of the new requests made on June 17. (False again, as explained below.) The June 29 letter notably did not take issue with Twitter’s refusal to provide responses to the discovery-like requests for emails, text messages, and other communications in the June 17 letter. But it contained a slew of new demands — several asking Twitter to create more custom reporting.

100. On July 1, 2022, Twitter pointed out just how far beyond the scope of Section 6.4 defendants’ requests had strayed. Nonetheless, Twitter noted that it was providing yet more information in response to recent requests and would continue to devote the “time and considerable resources” necessary to respond to outstanding requests. Twitter also explained that it had placed “no artificial throttling of rate limits.” In follow-up correspondence, it became clear that the “limit” Musk had bumped up against was not the result of throttling but a default 100,000-per-month limit on the number of queries that could be conducted. With his undisclosed team of data reviewers working behind the scenes, Musk had hit that limit within about two weeks. Twitter immediately agreed to, and did, raise the monthly search query limit one hundred-fold, to 10 million — more than 100 times what most paying Twitter customers would get.

101. From the outset of this extraordinary post-signing information exchange process, Musk accused Twitter of “lax” methodologies for calculating spam or false accounts. Knowing that his actions risked harm to Twitter and its stockholders, wreaked havoc on the trading price of Twitter’s stock, and could have serious consequences for the deal, Musk leveled serious charges, both publicly and through lawyer letters, that Twitter had misled its investors and customers. But Musk exhibited little interest in understanding Twitter’s process for estimating spam accounts that went into the company’s disclosures. Indeed, in a June 30 conversation with Segal, Musk acknowledged he had not read the detailed summary of Twitter’s sampling process provided back in May. Once again, Segal offered to spend time with Musk and review the detailed summary of Twitter’s sampling process as the Twitter team had done with Musk’s advisors. That meeting never occurred despite multiple attempts by Twitter.

102. From the outset, defendants’ information requests were designed to try to tank the deal. Musk’s increasingly outlandish requests reflect not a genuine examination of Twitter’s processes but a litigation-driven campaign to try to create a record of non-cooperation on Twitter’s part. When Twitter nonetheless bent over backwards to address the increasingly burdensome requests, Musk resorted to false assertions that it had not.