IN THE UNITED STATES DISTRICT COURT FOR THE DISTRICT OF MARYLAND UNITED STATES OF AMERICA, v. Civil Action No. CCB-22-1603

BOOZ ALLEN HAMILTON INC., et al.

MEMORANDUM Booz Allen Hamilton1 and EverWatch2 were the primary competitors for OPTIMAL

DECISION—a contract where the winner would provide modeling simulation and signals intelligence services to the National Security Agency (“NSA”). Booz Allen and EverWatch agreed to merge before the NSA opened the OPTIMAL DECISION contract for bidding. The United States of America (“the Government”) sued to stop the companies from consummating their merger agreement, alleging the acquisition violated antitrust laws by reducing each company’s incentive to compete for the OPTIMAL DECISION contract. Pending before the court is the Government’s Motion for a Preliminary Injunction. (ECF 29, Pl.’s Mot. for Prelim. Inj.) The motion is fully briefed (ECF 29; ECF 93, Defs.’ Resp. to Pl.’s Mot. for Prelim. Inj.; ECF 100, Pl.’s Reply Supp. Mot. for Prelim. Inj.), and the parties presented evidence during a two-day long hearing (ECF 207, Sept. 15, 2022, Min. Entry; ECF 208, Sept. 16,

2022, Min. Entry.) For the reasons below, the court will Deny the Government’s Motion for a

1 “Booz Allen Hamilton” refers collectively to the corporate entities known as “Booz Allen Hamilton Holding Corp.” and “Booz Allen Hamilton Inc.” 2 “EverWatch” refers collectively to the corporate entities known as “Everwatch Corp.,” “EC Defense Holdings, LLC,” and “Analysis, Computing & Engineering Solutions, Inc.” Preliminary Injunction.3 Based on the somewhat abbreviated record, however, a final judgment will not be entered at this time. I. BACKGROUND A. Signals Intelligence Modeling and Simulation Services

The National Security Agency is responsible for providing policymakers and military officials with the information they need to respond to global events. The NSA’s “signals intelligence”4 mission aims to do just that. Signals intelligence involves collecting foreign intelligence from communications and information systems. (ECF 29-3, Dunshee Decl. ¶ 3.) But the NSA does not simply dump raw data on decisionmakers; organization and modeling is key to the efficient and effective use of the data. To that end, the NSA uses “modeling and simulation services” to create mathematical models based on signals intelligence data. (Prelim. Inj. Hr’g Tr. Sept. 15 a.m. (Dunshee) 36:10–17.) And by helping the NSA analyze real-life systems, these models drive engineering, design, and trade-off decisions. (Id.; see also ECF 163-3, Dunshee Dep. 24:10–16.) Other government actors, in turn, call upon the NSA for their signals intelligence

modeling and simulation needs. (ECF 29-3 ¶ 5.) When the NSA obliges, those civil and military officials become the “mission customer” of the NSA. (Prelim. Inj. Hr’g Tr. Sept. 15 a.m. (Dunshee) 48:25–49:3, 62:5–13, 88:5–6.) The NSA is a customer too. Private companies support the NSA’s signals intelligence mission by providing modeling and simulation services to the agency. (Id. 37:11–17.) Put simply, these companies build, operate, and sustain the NSA’s models. (Id.) Booz Allen has fulfilled this important role for more than two decades. (ECF 29-3 ¶ 4.) The first modeling and simulation

3 This Memorandum constitutes the court’s findings of fact and conclusions of law. See Fed. R. Civ. P. 52(a)(2). 4 Certain documents in the record refer to “signals intelligence” as “SIGINT.” contracts were the “MASON” projects, starting with the 2002 MASON I contract. No matter the year, no matter the competitor, the result was the same: The NSA selected Booz Allen over competitors for its modeling and simulation needs. Take the 2007 MASON II and 2014 MASON III contracts. Each contract had two bidders. Booz Allen won both. (Prelim. Inj. Hr’g Tr. Sept. 15

a.m. (Dunshee) 39:19–40:5.) The NSA’s current modeling and simulation contract with Booz Allen, MASON III, is set to expire in March 2023. (Id. 41:15–16.) But MASON III has been extended multiple times, so it is unclear whether the March deadline will stick.5 The NSA and Booz Allen have agreed to two types of extensions over the years: “negotiated” extensions and “no-cost” extensions. “Negotiated” extensions led to increased labor rates, while “no-cost” extensions, as the name suggests, did not increase the labor rates. (Id. 43:18–45:5.) B. The OPTIMAL DECISION Procurement OPTIMAL DECISION is set to replace the MASON projects in hosting the NSA’s modeling and simulation tasks. (Id. at 41:19–21.) OPTIMAL DECISION is a “cost-plus-award-

fee” contract. (Id. at 46:14-22.) That means the NSA will reimburse the prime contractor for their costs in executing the contract, while also paying an “award fee” to the contractor. (Id. at 47:4– 10.)6 A percentage of the prime contractor’s costs are put in an “award fee pool.” OPTIMAL DECISION will use a 12 percent award fee pool. (Id. at 66:1-5; ECF 163-5 Kevin Y. Dep. 150:21– 25.)7 Every six months the NSA will evaluate the contractor’s performance against the “award fee

5 The NSA originally intended for MASON III to be a five-year contract. (Prelim. Inj. Hr’g Tr. Sept. 15 a.m. 41:9- 14.) In the last three years, the NSA changed the OPTIMAL DECISION RFP release date at least eight times. (ECF 163-15, Pl.’s First Am. Obj. and Resp. to Defs.’ Second Interog., at 4–6.) Despite the RFP’s release, whether the NSA accepts and finalizes a bid by March 2023 remains an open question. 6 Labor is the largest cost component for a contract like OPTIMAL DECISION. (Prelim. Inj. Hr’g Tr. Sept. 15 a.m. (Dunshee) 47:13–48:10.) 7 The surnames of some NSA witnesses are identified only by their first letter due to NSA policy. criteria”—a grading rubric set by the contract. (Prelim. Inj. Hr’g Tr. Sept. 15 a.m. (Dunshee) 47:4– 10.) The score is a number between zero and 100 and translates to the percentage given to the contractor from the award fee pool. The math is simple: the higher the costs, the larger the pool; the higher the score, the larger the award.

Despite the name change, OPTIMAL DECISION will continue the same scope of work set out by its MASON I, II, and III predecessors. (Id. at 54:3–4.) Yet differences remain between OPTIMAL DECISION and the MASON trilogy. The NSA tinkered with OPTIMAL DECISION to reduce the number of key personnel required for the contract. The NSA also removed “past performance on a similar contract” as a factor in evaluating prospective bidders. Both changes, according to the NSA, were made to stir up competitive interest in the contract. (Id. at 54:13– 55:4.)8 The NSA cast a wide net for OPTIMAL DECISION’s prospective suitors. In October 2020, the NSA identified 178 companies as potential contractors for OPTIMAL DECISION. (ECF 29- 3 at ¶ 6.) When the NSA surveyed those companies, 14 noted their interest in becoming the prime

contractor. (Id.) One year later, in October 2021, the NSA sought Letters of Intent to Bid from those 14 companies. Only two companies responded: Booz Allen and EverWatch. (Prelim. Inj. Hr’g Tr. Sept. 15 a.m. (Dunshee) 52:11–15.) After long delays, the NSA finally released the OPTIMAL DECISION Request for Proposals (“RFP”) on September 14, 2022, the day before the evidentiary hearing on this motion for a preliminary injunction. (Id. 42:3–7.) Final bids are due October 28, 2022. (Id. 42:8–9.)

8 The Government argues these changes reflect the NSA’s intent to signal that it was open to companies other than Booz Allen. C. Booz Allen’s Proposed Acquisition of EverWatch Booz Allen is a consulting firm that offers management, technology, and engineering services to public and private sector organizations. The company has more than 29,000 employees and totaled $8.4 billion in revenue in fiscal year 2022.9 EverWatch is a newer and smaller company specializing in defense, intelligence, and mission support services and technology.10

In March 2022, Booz Allen agreed to acquire EverWatch for $440 million (“Proposed Acquisition”). Booz Allen contends the transaction was motivated by its desire to compete against larger systems integrators like Boeing, Raytheon, and Lockheed Martin. And from EverWatch’s perspective, the deal offered the company an escape from a Goldilocks purgatory: too big to qualify for small business opportunities, yet too small to credibly challenge larger companies for lucrative contracts. But the Proposed Acquisition has not yet closed. The companies must meet several closing requirements to seal the deal. The Government argues Booz Allen acquired EverWatch to avoid competition for OPTIMAL DECISION. Booz Allen viewed EverWatch as its main competitor for OPTIMAL

DECISION. The Proposed Acquisition, under the Government’s theory, ensures that Booz Allen will ultimately reap the profits no matter which company “wins” the contract. With Booz Allen guaranteed the profit, neither company has the incentive to compete for OPTIMAL DECISION, potentially resulting in higher prices or lower quality services for the NSA. So the Government sued under Section 1 of the Sherman Act and Section 7 of the Clayton Act to stop the transaction.11

9 Booz Allen Hamilton, Fact Sheet: About Booz Allen, https://www.boozallen.com/content/dam/boozallen_site/esg /pdf/slick_sheet/booz-allen-hamilton-fact-sheet.pdf (last visited Sept. 20, 2022). 10 EverWatch, What We Do, https://everwatchsolutions.com/what-we-do/ (last visited Sept. 20, 2022). 11 The Government’s modified request for relief would still let the defendants merge so long as they do not finalize the deal until 90 days after the deadline to submit bids for OPTIMAL DECISION, along with EverWatch retaining the right to walk away from the merger. (ECF 200-1, Pl.’s Revised Proposed Order.) II. LEGAL STANDARD A party seeking a preliminary injunction must establish: (1) the likelihood of success on the merits; (2) the likelihood of irreparable harm in the absence of preliminary relief; (3) that the balance of equities favors issuing the preliminary injunction; and (4) that the injunction is in the

public interest. See Winter v. Nat’l Res. Def. Council, Inc., 555 U.S. 7, 20 (2008). III. DISCUSSION A. Likelihood of Success on the Merits Section 1 of the Sherman Act prohibits “[e]very contract, combination in the form of trust or otherwise, or conspiracy, in restraint of trade or commerce . . . .” 15 U.S.C. § 1. Because nearly every contract restrains trade to some extent, the Supreme Court has made clear that “Congress intended to outlaw only unreasonable restraints” on trade. State Oil Co. v. Khan, 522 U.S. 3, 10 (1997) (emphasis added); see also NCAA v. Alston, 141 S. Ct. 2141, 2151 (2021) (“[T]he phrase ‘restraint of trade’ is best read to mean ‘undue restraint.’”) (quoting Ohio v. American Express Co., 138 S. Ct. 2274, 2283 (2018)). To prove a Section 1 violation, a plaintiff must show: “(1) a

contract, combination, or conspiracy;” (2) “imposed an unreasonable restraint of trade.” See Robertson v. Sea Pines Real Est. Cos., 679 F.3d 278, 284 (4th Cir. 2012) (quoting Dickson v. Microsoft Corp., 309 F.3d 193, 202 (4th Cir. 2002)). The parties agree Booz Allen’s agreement to buy EverWatch is a contract under the first element. The parties diverge, however, on whether the merger agreement unreasonably restrains trade. The court may answer that question with a “quick look” or a “rule of reason” analysis.12

12 A Section 1 plaintiff may also use a per se theory, but the Government has elected not to pursue that here. 1. “Quick Look” Analysis Under a “quick look” approach, the court can declare a practice illegal if an observer with “even a rudimentary understanding of economics could conclude that the arrangements in question would have an anticompetitive effect on customers and markets.” California Dental Ass’n v. FTC,

526 U.S. 756, 770 (1999). When an agreement poses an “intuitively obvious” restraint on trade, a court relying on the “quick look” method need analyze the transaction with the “detailed treatment” required by the rule of reason. Id. at 781. Few situations justify a “quick look” review; none like this case. A “quick look” might be proper, for example, where an agreement expressly limits output and fixes a minimum price. Id. at 770 (citing NCAA v. Bd. of Regents of Univ. of Okla., 468 U.S. 85, 104 (1984)). An “absolute ban on competitive bidding,” Nat’l Soc. of Pro. Eng’rs v. United States, 435 U.S. 679, 692 (1978), or a horizontal agreement to withhold services from customers, FTC v. Ind. Fed’n of Dentists, 476 U.S. 447, 459 (1986), may also warrant a “quick look” approach. The Proposed Acquisition bears little resemblance to these extreme restraints on trade.

Several factors mitigate the risk of the agreement causing anticompetitive harm. See infra Section III.A.2.ii; see also Cal. Dental Ass’n, 526 U.S. at 771 (refusing to apply a “quick look” analysis to an agreement without “comparably obvious” anticompetitive effects). And a “quick look” approach is proper only where courts have “amassed ‘considerable experience with the type of restraint at issue’ and ‘can predict with confidence that it would be invalidated in all or almost all instances.’” See Alston, 141 S. Ct. at 2156 (quoting Leegin Creative Leather Products, Inc. v. PSKS, Inc., 551 U.S. 877, 886–887 (2007)). Here, the Government seeks to invalidate the Proposed Acquisition because it may harm one customer on a single contract. Few courts have analyzed antitrust challenges under such narrow circumstances. Accordingly, a “quick look” cannot resolve the novel questions raised here.13 2. “Rule of Reason” Analysis The court will analyze the Proposed Acquisition under the “rule of reason.” See Texaco

Inc. v. Dagher, 547 U.S. 1, 5 (2006) (“[T]his Court presumptively applies rule of reason analysis, under which antitrust plaintiffs must demonstrate that a particular contract or combination is in fact unreasonable and anticompetitive before it will be found unlawful.”). A rule of reason analysis demands a “fact-specific assessment of market power and market structure” aimed at evaluating a “challenged restraint’s actual effect on competition—especially its capacity to reduce output and increase price.” Alston, 141 S. Ct. at 2155 (quoting American Express, 138 S. Ct. at 2284) (internal quotation marks omitted). Context is key. The court may, for example, explore “specific information about the relevant business and the restraint’s history, nature, and effect.” Leegin Creative Leather Prods., Inc., 551 U.S. at 885 (quoting State Oil Co. v. Khan, 522 U.S. 3, 10 (1997)) (internal quotation marks omitted). “Whether the businesses involved have market power

is a further, significant consideration.” Id. at 885–86. A three-step, burden shifting framework often guides a rule of reason analysis.14 Under this framework, “the plaintiff has the initial burden to prove that the challenged restraint has a substantial anticompetitive effect.” Alston, 141 S. Ct. at 2160 (quoting American Express Co., 138

13 The court does not discuss the potential procompetitive effects of the Proposed Acquisition in detail. However, the Proposed Acquisition could arguably increase competition in some areas. That is, EverWatch and Booz Allen may have a stronger position to challenge entrenched incumbents like Raytheon or Lockheed Martin for lucrative government contracts. (Prelim. Inj. Hr’g Tr. Sept. 16 am. (Cooper) 42:23–43:13.) The existence of this argument makes a quick look approach inappropriate. Because the Proposed Acquisition might “have a net procompetitive effect, or possibly no effect at all on competition, more than a ‘quick look’ is required.” See 1-800 Contacts, Inc. v. FTC, 1 F.4th 102, 115 (2d Cir. 2021) (quoting Major League Baseball Props., Inc. v. Salvino, Inc., 542 F.3d 290, 318 (2d Cir. 2008)) (internal punctuation omitted). 14 The three-step framework does not “represent a rote checklist.” Alston, 141 S. Ct. at 2160. The court uses the burden- shifting framework as a loose organizational tool, not “an inflexible substitute for careful analysis.” Id. S. Ct. at 2284.) If the plaintiff carries that burden, the burden then “shifts to the defendant to show a procompetitive rationale for the restraint.” Id. (quoting American Express Co., 138 S. Ct. at 2284). Should the defendant make that showing, “the burden shifts back to the plaintiff to demonstrate that the procompetitive efficiencies could be reasonably achieved through less

anticompetitive means.” Id. (quoting American Express Co., 138 S. Ct. at 2284). Proving a “substantial anticompetitive effect” under the first step is “no slight burden.” Alston, 141 S. Ct. at 2160–61 (noting “courts have disposed of nearly all rule of reason cases in the last 45 years on the ground that the plaintiff failed to show a substantial anticompetitive effect”). Plaintiffs may “make this showing directly or indirectly.” American Express Co., 138 S. Ct. at 2284. “Direct evidence of anticompetitive effects would be proof of actual detrimental effects on competition, such as reduced output, increased prices, or decreased quality in the relevant market.” Id. (internal citations, brackets, and quotation marks omitted). “Indirect evidence would be proof of market power plus some evidence that the challenged restraint harms competition.” Id. (citations omitted).

The court need not go past the initial step here for three reasons. First, no direct evidence suggests the Proposed Acquisition has “actual detrimental effects” on competition. Second, the Proposed Acquisition does not pose a likely or significant risk of anticompetitive harm because several countervailing incentives prevent unwarranted price hikes. See Dickson, 309 F.3d 193 at (requiring plaintiff to show “that harm is not only possible but likely and significant”) (emphasis added). Third, the Government has not sufficiently defined a relevant market so allegations about Booz Allen’s “market power” are analytically incomplete. The court will address these three problems for the Government’s case in turn. i. No Direct Evidence of Anticompetitive Effects “Direct evidence” must show actual detrimental effects on competition such as an expressed intent to reduce output, increase prices, or decrease quality. See American Express Co., 138 S. Ct. at 2284. No such evidence exists here. The Government instead relies on

uncontextualized statements from the defendants’ employees to show the companies have a reduced incentive to compete for OPTIMAL DECISION.15 But these comments simply represent excitement and uncertainty among lower-level employees when the Proposed Acquisition was announced. Most statements were made the same day the news broke, others no more than a few weeks later. These remarks are unsurprising. The Proposed Acquisition was a surprising update for these employees. Banter with coworkers is a natural and expected response to big news. The court cannot predict a company-wide shift in bidding strategy based on these off-handed and speculative comments.16 At best, these contemporaneous statements reflect the individual perception of specific employees, not broad corporate strategy. Where the speakers sit in the chain of command matters.

Lower-level employees were not told about the transaction in advance, so their confusion and excitement about the announcement is unsurprising. Each comment comes from a lower-level

15 See, e.g., ECF 217-17, Pl.’s Ex. 26; ECF 217-15, Pl.’s Ex. 22; ECF 217-11, Pl.’s Ex. 18; ECF 217-18, Pl.’s Ex. 27. 16 To be sure, our evidentiary system holds sacrosanct “excited utterances” like these. See, e.g., Fed. R. Evid. 803(2). But even crediting their truth, most comments simply acknowledge that EverWatch and Booz Allen are the main competitors for OPTIMAL DECISION. No message explicitly expresses an intent to submit a less competitive bid. Take what may be the Government’s strongest evidence: a Booz Allen employee asking, “how hard we need to try to win [OPTIMAL DECISION].” (ECF 217-18, at BAH_DOJ_00047426.) The Government’s post-hearing brief quotes the response as follows: I don’t think we have competition at all” and “I just don’t see any scenario where EverWatch spends B&P [bid and proposal] on this . . . even though they can’t act like they’ve been acquired yet.” (ECF 217, Pl.’s Post-Hearing Br., at 12.) The Government, however, omits a critical phrase in the employee’s response: “I don’t think we have competition at all but I suspect Becky will have thoughts on it and will make those thoughts clear early on.” (ECF 217-18, at BAH_DOJ_00047426.) When Rebecca (“Becky”) Robertson weighed in her message was clear: Booz Allen must compete aggressively for OPTIMAL DECISION. (ECF 214-12, Gosnell Dep. 114:13–17, 118:7– 17.) employee, making it difficult to extrapolate their remarks to structural shifts in the companies’ intent to compete.17 When these employees did speak with their higher-ups, the message from executives was clear: the race for OPTIMAL DECISION was “full steam ahead.” (Prelim. Inj. Hr’g Tr. Sept. 16

a.m. (Cooper) 53:4–9; see also ECF 214-12, Gosnell Dep, 114:12–16.) Leadership instructed employees to continue “business as usual.” (ECF 214-12, Gosnell Dep, 118:7–17.) Although one EverWatch employee first told his team to “stand down” on OPTIMAL DECISION work, that sentiment changed quickly. Once his superiors told him the Proposed Acquisition would not change the competition for OPTIMAL DECISION, the employee and his team adjusted accordingly. (Prelim. Inj. Hr’g Tr. Sept. 16 a.m. (Catrambone) 111:4–18.) Of course, executives can say one thing and do the opposite behind closed doors. But even armed with targeted discovery,18 the Government has found nothing beyond informal chats and emails from lower-level employees surprised to learn about the deal. Two weeks after the announcement, the dust settled, and employees and leadership alike seemed ready to compete. (See

id. (Dotson) 20:18–21:9, (Robertson) 87:6–25, (Catrambone) 113:2–114:3.) Little evidence suggests the companies, or their capture-team employees, intend to give the NSA anything less than their best proposal. Competition exists now. Formal reviews are ongoing. Teams are practicing for oral presentations. (See id. (Catrambone) 113:8–10.) Employees have been running this race for months with the finish line just weeks away. Nobody seems willing to give up in the final stretch.

17 The Government cites United States v. Aetna Inc., 240 F. Supp. 3d 1, 70 (D.D.C. 2017), in arguing that “contemporaneous email exchanges” are more persuasive than “in-court attempts to explain or disavow those documented exchanges.” But that case involved contemporaneous emails sent by executives of the company. Id. at 67. 18 In addition to traditional discovery, the Government used Civil Investigative Demands for pre-litigation information gathering. See 15 U.S.C. § 1312. ii. No Likely or Significant Competitive Harm Without direct evidence of anticompetitive harm, the Government must provide indirect proof of anticompetitive effects, which requires “some evidence that the challenged restraint harms competition” and a properly defined market. Am. Express, 138 S. Ct. at 2284–85. The Government

has shown neither. The Government’s theory of anticompetitive harm is straightforward but ultimately unpersuasive. Before the Proposed Acquisition’s announcement, Booz Allen and EverWatch were competing for OPTIMAL DECISION. Then, according to the Government, the companies decided to combine rather than compete. Now, neither company has an incentive to offer the NSA higher quality or lower cost services. After all, a price cut from either company’s bid would simply take profit from the soon-to-be-merged company. The simplicity is alluring, but illusory. The bottom- line is that the Government cannot show “that harm is not only possible but likely and significant.” See Dickson, 309 F.3d at 206 (emphasis added). To start, the Government’s expert witness, Dr. Mark Chicu, ignores countervailing

incentives that protect and sustain competition. Dr. Chicu argues the Proposed Acquisition reduced the defendants’ incentives to compete because no matter which company wins OPTIMAL DECISION, Booz Allen will ultimately gain the contract after acquiring EverWatch. (Prelim. Inj. Hr’g Tr. Sept. 15 p.m. (Chicu) 60:1–12.) But his analysis suffers from tunnel vision. By focusing only on Booz Allen’s decreased incentive to compete, Dr. Chicu ignores other incentives encouraging Booz Allen’s vigorous competition. (Prelim. Inj. Hr’g Tr. Sept. 16 p.m. (Bailey) 7:23–8:10.) Dr. Chicu spends little time evaluating whether these counterincentives—such as Booz Allen’s reputation with the NSA, its desire to win other contracts, and regulatory constraints— might outweigh potential profit increases on OPTIMAL DECISION. (Id. 8:17–24.) Neither people nor companies act for one reason and one reason alone. Even the most mundane decisions involve weighing competing incentives. After all, “an incentive is just the first step along the way to evaluating whether or not there’s an effect.” (Prelim. Inj. Hr’g Tr. Sept. 15 p.m. (Chicu) 85:17– 18.) Analyzing one changed incentive is only part of the picture, so the court must consider the

broader context surrounding the Proposed Acquisition. (Prelim. Inj. Hr’g Tr. Sept. 16 p.m. (Bailey) 9:2–8.) Booz Allen has strong countervailing incentives to maintain a competitive bid. OPTIMAL DECISION is just a fraction of the NSA’s contracting work; there are billions of dollars at stake in the industry. (Prelim. Inj. Hr’g Tr. Sept. 16 a.m. (Robertson) 81:3–4.) The Proposed Acquisition could allow EverWatch and Booz Allen to more effectively compete with Raytheon, Lockheed Martin, and other entrenched incumbents for more lucrative contracts. (Id. (Cooper) 42:23–43:13.) Booz Allen needs a sterling reputation to have a shot at these other opportunities. (Id. (Dotson) 13:1–3; see also id. (Cooper) 51:18–19 (“In our market, you know, credibility and reputation is everything.”)) The NSA often looks to “past performance” when evaluating proposals, so any

unjustified price increase on OPTIMAL DECISION would risk losing future procurement opportunities. (Id. (Dotson) 13:1–3) The NSA would know if Booz Allen over-charged or under- performed on OPTIMAL DECISION. After all, Booz Allen has provided modeling and simulation services to the NSA for more than two decades. (Id. (Dotson) 13:11–17.) If the NSA spies a bad deal, Booz Allen could lose billions in future contract opportunities. That threat should deter Booz Allen from submitting anything but their best on OPTIMAL DECISION. See generally Gen. Dynamics Corp. v. United States, 563 U.S. 478, 491 (2011) (“Government contractors—especially cutting-edge defense contractors . . . are repeat players” with a “strong incentive to behave rather than risk missing out on the next multibillion-dollar defense contract.”). Booz Allen’s employees also have individual reputational interests constraining price increases. Team leaders have developed reputations with their NSA colleagues for decades. (Prelim. Inj. Hr’g Tr. Sept. 16 a.m. (Robertson) 79:10–24.) For people like Vice President Rebecca Robertson, failing to deliver a quality product for the right price on OPTIMAL DECISION would

hurt her credibility and professional relationships. (Id.) Others have an innate desire to win that is not so easily dissolved by a different logo on their paystub. (See id. (Catrambone) 114:6–7 (“My personal compensation doesn’t change whether I win. It’s more pride.”)) Tangible financial incentives, too, encourage ongoing competition for OPTIMAL DECISION. EverWatch employees will lose their bonuses if they fail to win the contract. (Id. (Cooper) 52:2–11.) And Booz Allen has committed to enticing competition even after the Proposed Acquisition closes. (See ECF 214-6, Proposed Final Judgment and Order.)19 For example, the company guaranteed bonuses to either EverWatch’s or Booz Allen’s team—not both—if the NSA picks their respective bid. The Government’s theory of human behavior is far too reductive. It is hard to believe the defendants’ employees were vigorously competing earlier

this year, but stopped competing once the Proposed Acquisition was announced, only to continue competing once the Government brought this litigation, and will again stop competing if the court denies the Government’s motion. Competition is not a light switch; it’s a sliding scale. Some competitive incentives stick even when one incentive fades. Booz Allen has enough reasons to pause before increasing its price on OPTIMAL DECISION. The Government suggests the companies would risk it all for a better profit margin on OPTIMAL DECISION. But OPTIMAL DECSISION is a “relatively small”

19 Booz Allen and EverWatch have committed to submit separate bids, stick with those bids, implement various firewalls between the two bidding teams, and create financial incentives for the winning team. The defendants committed to implementing these procedures even if the court declines to mandate the defendants complete these actions. (Closing Arg. Tr. 41:2–25.) contract by all accounts, so Booz Allen likely would not jeopardize its long-term financial interests for a trivial payout on a single project. (Prelim. Inj. Hr’g Tr. Sept. 15 p.m. (Kevin Y.) 22:11–14; Prelim. Inj. Hr’g Tr. Sept. 16 a.m. (Robertson) 69:22–70:2.) Recent history shows price increases are not inevitable even when Booz Allen has no other

bidding competition. MASON III’s first extension in April 2019 was sole-sourced to Booz Allen. (Prelim. Inj. Hr’g Tr. Sept. 16 p.m. (Bailey) 23:10–16.)20 When the NSA extends existing contracts, no bidding occurs. Put bluntly, no competition exists. Booz Allen could have renegotiated the contract for a higher payout. But Booz Allen’s rates did not increase. (ECF 163- 27, Dr. Bailey Expert Report, Ex. 1, at 12.) The parties agreed to continue the contract on the same terms, for the same price, as a “no cost” extension. (Prelim. Inj. Hr’g Tr. Sept. 16 a.m. (Robertson) 73:1–7.) As noted by the defendants’ expert witness, Dr. Elizabeth Bailey, the price of Booz Allen’s services decreased from September 2018 to April 2019 after adjusting for inflation. (ECF 163-27, Dr. Bailey Expert Report, Ex. 1, at 12.) Again, Booz Allen faced no competition at that time.21 If the Government were correct about Booz Allen’s incentives to increase prices, the

company would have seized this opportunity to raise the cost or decrease the quality of their services.22 Yet neither event occurred.23 Only the defendants’ theory explains the price stability

20 “Sole source” negotiations are bilateral price negotiations that don't include other bidders. The negotiations include only the NSA and the prime contractor, not subcontractors. (Prelim. Inj. Hr’g Tr. Sept. 15 p.m. (Kevin Y.) 33:24– 34:5.) 21 Curiously, Dr. Chicu considers the April 2019 “no cost” extension as an example of a competitive market, despite the extension being “sole sourced” to Booz Allen. 22 The Government asserts that Booz Allen’s bidding strategy in the JUNGLE PRINCE contract supports its theory. But the record concerning JUNGLE PRINCE is insufficiently developed, and information about the contract may have been withheld from discovery. Accordingly, the court will not draw conclusions from the Government’s passing references to JUNGLE PRINCE. 23 See, e.g., Prelim. Inj. Hr’g Tr. Sept. 15 a.m. (Dr. S.) 119:19–21 (“Q: You would agree [Booz Allen] ha[s] been an effective mission partner with you throughout the whole time? A: They have been an effective mission partner, yes.”); Prelim. Inj. Hr’g Tr. Sept. 15 p.m. (Kevin Y.) 38:3–7 (“Q: You can't recall a single situation with Booz Allen where you were negotiating with them on a contract and the end result was anything other than a fair and reasonable price, can you, sir? A: No.”) between September 2018 to April 2019. Certain considerations—reputational or otherwise—must have made an artificial price hike not worth it. Eventually, Booz Allen did increase its rates when the parties extended MASON III in May 2021 and April 2022. On the one hand, the Government submits these rate increases as evidence

of Booz Allen’s intent and capability to upcharge the NSA when no other company bids against it. On the other hand, those rate increases came more than six years after the parties signed the original contract.24 The defendants attribute those recent cost increases to rising labor rates. (See Prelim. Inj. Hr’g Tr. Sept. 16 p.m. (Bailey) 25:1–10.) Indeed, the NSA was not surprised by Booz Allen’s request for higher labor rates. (See Prelim. Inj. Hr’g Tr. Sept. 15 a.m. (Dr. S.) 123:16–17.) But Dr. Chicu deflated Booz Allen’s rates to account for the change in labor costs over time, so simply noting that labor costs have generally increased does not necessarily resolve the issue. (See ECF 217-25, Chicu Am. Decl., at USDOJ-019-00000393-001; see also Closing Arg. Tr. 18:15– 23.) Still, recently increased demand for Booz Allen’s talent sufficiently explains the May 2021

and April 2022 costs increases. The defendants have faced an employee exodus lately. Many employees now seek work outside the defense industry. (See Prelim. Inj. Hr’g Tr. Sept. 16 a.m. (Dotson) 23:6–11; id. (Robertson) 74:24–25, 75:4–9.) Indeed, companies like Microsoft, Amazon, and Salesforce recruit heavily from the defendants’ rosters. (Id. (Cooper) 56:9–17.) With those businesses offering big payouts for professionals with security clearances, the defendants have needed to fight to retain top-notch talent. (Id. (Doston) 23:8–11.) Dr. Chicu’s standardized rates reflect only general adjustments for inflation, and ignore the increased demand for IT professionals

24 The parties dispute whether the adjustments for inflation by the expert witnesses account for rising labor costs. The court need not resolve that dispute because other evidence in the record, such as the increased demand for talent from other industries, sufficiently explains those cost increases for now. in the defense industry.25 In fact, the NSA set OPTIMAL DECISION’s ICE at

Accordingly, the May 2021 and April 2022 rate increases do not necessarily imply Booz Allen will impose increased rates without a

competing bidder. Even if Booz Allen artificially increased the cost or decreased the quality of its work, the NSA still has significant control over Booz Allen’s profit on OPTIMAL DECISION.26 To start, the NSA cannot award OPTIMAL DECISION to any company unless it determines the bid is “fair and reasonable.” (See Prelim. Inj. Hr’g Tr. Sept. 15 a.m. (Dunshee) 81:11–15.) That is, the NSA could send the final bidders back to the drawing board or negotiate certain parts of the bid. (See Prelim. Inj. Hr’g Tr. Sept. 16 a.m. (Cooper) 50:16–51:7.) The NSA has complete control over every detail in the now-released OPTIMAL DECISION RFP. (See Prelim. Inj. Hr’g Tr. Sept. 15 a.m. (Dunshee) 81:8–10.) OPTIMAL DECISION is a “cost plus award fee” which means the NSA will reimburse the prime contractor for their costs in executing the contract, while also paying an

“award fee” to the contractor. (Id. 47:4–10.) The size and range of award fees remain within the NSA’s discretion. In fact, the NSA could even award zero dollars if it so chooses. (Id. 81:19– 82:4.) And this power never expires during the contract. The NSA continually reviews

25 Dr. Chicu adjusts for inflation by relying on the BLS Employment Cost Index for “private industry workers in professional, scientific, and technical services.” See ECF 217-25, Chicu Amend. Decl., at USDOJ-019-00000393- 001 n.1.) The “professional, scientific, and technical services” sector includes activities such as “legal advice and representation; accounting, bookkeeping, and payroll services; architectural, engineering, and specialized design services; computer services; consulting services; research services; advertising services; photographic services; translation and interpretation services; veterinary services; and other professional, scientific, and technical services.” See U.S. Bureau of Labor Statistics, Professional, Scientific, and Technical Services, https://www.bls. gov/iag/tgs/iag54htm. 26 To be clear, the NSA’s power to tinker with a contractor’s profit does not immunize defense contractors from antitrust scrutiny. An agency’s regulatory power to mitigate anticompetitive effects is simply another consideration when evaluating whether a firm is likely to cause a significant anticompetitive harm. performance every six months to determine what award, if any, the contractor will receive. (Id. 81:23–82:1.) Put simply, the NSA can mitigate the Government’s worst case scenario, so significant competitive harm is unlikely to materialize. iii. Insufficient Market Definition

Even if there were “some evidence” the Proposed Acquisition would likely cause anticompetitive effects, see Am. Express, 138 S. Ct. at 2284, the Government has not sufficiently defined the relevant economic market at this stage. Properly defining the market is a “necessary predicate” to examining the competitive effects of a horizontal merger. Brown Shoe Co. v. United States, 370 U.S. 294, 335 (1962).27 Indeed, without a market definition “there is no way to measure [the Proposed Acquisition’s] ability to lessen or destroy competition.” See Am. Express, 138 S. Ct. at 2285. A “relevant market” is “the area of effective competition,” id., and includes products “reasonably interchangeable by consumers for the same purpose[],” United States v. E.I. du Pont de Nemours & Co., 351 U.S. 377, 395 (1956). The issue disputed here is whether signals intelligence modeling and simulation services under OPTIMAL DECISION constitutes a relevant product market.28 The answer on the current record is no.

The Government’s proposed market depends on the application (OPTIMAL DECISION) of a broad tool (modeling and simulation) to one network system (signals intelligence) for one

27 “An economic arrangement between companies performing similar functions in the production or sale of comparable goods or services is characterized as ‘horizontal.’ The effect on competition of such an arrangement depends, of course, upon its character and scope. Thus, its validity in the face of the antitrust laws will depend upon such factors as: the relative size and number of the parties to the arrangement; whether it allocates shares of the market among the parties; whether it fixes prices at which the parties will sell their product; or whether it absorbs or insulates competitors.” Brown Shoe, 370 U.S. at 334–35. 28 A relevant antitrust market has two distinct but related elements—a geographic market and product market. Brown Shoe Co. v. United States, 370 U.S. 294, 324 (1962). Here, the geographic component is the United States. The NSA is the only customer in the Government’s proposed market, and the NSA must procure OPTIMAL DECISION services from domestic companies. (See Prelim. Inj. Hr’g Tr. Sept. 15 pm. (Chicu) 61:11–18.) The parties dispute the product component. customer (the NSA). By defining the market so narrowly, the Government attempts to “gerrymander its way to victory without due regard for market realities.” It’s My Party, Inc. v. LiveNation, Inc., 811 F.3d 676, 683 (4th Cir. 2016). The Government folds “signals intelligence” into “modeling and simulation services” to give its proposed market the gloss of exclusivity, but

that synergy cannot establish an antitrust market. Modeling and simulation services are not unique to the NSA. Nor do such services have “peculiar characteristics.” See Brown Shoe, 370 U.S. at 325. In fact, “computer networks all over the United States” use modeling and simulation. (Prelim. Inj. Hr’g Tr. Sept. 15 a.m. (Dunshee) 99:3–11.) The federal government alone has several modeling and simulation contracts across multiple industries. (Prelim. Inj. Hr’g Tr. Sept. 16 p.m. (Bailey) 35:2–7.) With nearly a dozen different federal agencies awarding modeling and simulation contracts to 32 companies since 2020, (ECF 163-27, Dr. Bailey Expert Report, at ¶ 21), these services do not appear to have particularly “distinct customers” or “specialized vendors.” See Brown Shoe, 370 U.S. at 325. “Modeling and simulation” is simply a “systems engineering technique.” (Prelim. Inj. Hr’g

Tr. Sept. 15 a.m. (Dunshee) 36:10–17.) Users wield this technique to develop mathematical or computer science-based models of systems, which may assist with decision making processes. (Id.) At first glance, the definition seems vague. Any number of “techniques” can process information and help people make decisions about complex systems. Dr. S., the NSA’s Chief of Enterprise Modeling and Simulation, sought to illuminate the process by describing it as “a tool to answer problems that are difficult to . . . answer easily.” (Id. (Dr. S.) 106:2–3.) That definition, too, feels too broad. But no further specificity lurks beneath the lexicon. Distilled to its core, “modeling and simulation” is just a “broad tool” that uses mathematical models to “come up with more insightful solutions.” (Id. 106:3–5.) The Government seeks to put OPTIMAL DECISION under a microscope by defining the relevant market as “signals intelligence modeling and simulation.” Modeling and simulation services do not stand alone—they use and apply data from underlying systems. Here, the NSA models and simulates “signals intelligence” data as its underlying system. OPTIMAL DECISION will perform modeling and simulation on a large scale, or at an “enterprise level.”29 (Prelim. Inj.

Hr’g Tr. Sept. 15 a.m. (Dunshee) 36:21–37:2.) While the federal government has more than a dozen contracts concerning “signals intelligence,” (Prelim. Inj. Hr’g Tr. Sept. 16 p.m. (Bailey) 35:8–10), OPTIMAL DECISION is the only contract applying “modeling and simulation services” to “signals intelligence.” (Id. 70:18–22.) Put plainly, the Government’s position is that OPTIMAL DECISION is the only way the NSA can receive these modeling and simulation services. Because the NSA needs its contractors to have “domain knowledge” in signals intelligence, the agency cannot look to other providers. And security clearance requirements impose high startup costs, so newcomers are dissuaded from joining the market.

Adding qualifiers like “signals intelligence” to whittle a relevant market down to a single contract gives too much weight to one customer’s preference. A type of service, such as modeling and simulation, is distinct from an application of that service to different contexts like “signals intelligence” data.30 The Government argues that OPTIMAL DECISION requires “domain knowledge” in signals intelligence, so the services performed under the contract are not interchangeable with other modeling and simulation services. Dr. S. explained that “domain

29 Modeling and simulation at the “enterprise level” means the models look “across the entire NSA cryptologic enterprise.” (Prelim. Inj. Hr’g Tr. Sept. 15 am. (Dunshee) 40:16–24.) 30 Consider the largest technology companies for example. These companies likely use modeling and simulation services to analyze data on their own networks. No two networks are the same; data collected by Amazon is likely different than the data gathered by Apple and Tesla. If the Government were correct, then each company’s use of modeling and simulation services would constitute a separate market. knowledge” essentially means detailed “background knowledge.” (Prelim. Inj. Hr’g Tr. Sept. 15 a.m. (Dr. S.) 108:17.) But the specific background knowledge “needed” for OPTIMAL DECISION is just an idealized list of qualities the NSA hopes its mission partner will have. Some qualities have nothing to do with signals intelligence at all. Notably, Dr. S. would like the winner to have

familiarity with NSA missions, (id. 108:23–25), an ability to “relate to the customer’s problems,” (id. 108:15), and experience with NSA software, (id. 114:7–15). In other words, Dr. S. wants the winner to have “knowledge about everything that goes on at NSA, infinite knowledge on the tools available, [and] knowledge all of our customers have.” (Id. 129:12–21.) The court need not help the NSA find its perfect match. Services that are “reasonably interchangeable” will do just fine. Queen City Pizza, Inc. v. Domino’s Pizza, Inc., 124 F.3d 430, 438 (3d Cir. 1997) (“The test for a relevant market is not commodities reasonably interchangeable by a particular plaintiff, but ‘commodities reasonably interchangeable by consumers for the same purposes.’”) (quoting E.I. du Pont de Nemours & Co., 351 U.S. at 438). Although Booz Allen or EverWatch likely have the best shot at checking these aspirational boxes, the record reflects others

could do the job. In fact, the NSA identified over a hundred companies as potential contractors for OPTIMAL DECISION and fourteen expressed an interest in being the prime contractor. (ECF 29- 3, Dunshee Decl. ¶ 6.) Why these companies dropped from the race is irrelevant. What matters is the NSA, at some point, recognized that the market for potential contractors is much larger than Booz Allen and EverWatch.31 Relevant signals intelligence domain knowledge may make a company better at modeling and simulation in certain contexts. But that is not a reason to require domain knowledge in defining

31 The Government’s curates its market based on a moment suspended in time. With only a few weeks left until the bidding deadline, it is likely only Booz Allen and EverWatch will submit a bid. But a snapshot at the finish line says nothing about the competitors running the race initially. the relevant market here. See Spectrofuge Corp. v. Beckman Instruments, Inc., 575 F.2d 256, 282 (5th Cir. 1978) (“The fact that a company limits its competitive activity to a single firm’s products (and at only one competitive level) cannot control the definition of the relevant market.”). After all, applying modeling and simulation to a signals intelligence system is “fairly easy” according to

EverWatch’s technical lead. (Prelim. Inj. Hr’g Tr. Sept. 16 a.m. (Catrambone) 104:14–16.) And if domain knowledge about signals intelligence was so inextricably tied to effective modeling and simulation, the NSA would not have cut signals intelligence specialists from the “key personnel” list on OPTIMAL DECISION. (Prelim. Inj. Hr’g Tr. Sept. 15 a.m. (Dr. S.) 127:25–128:16.) In fact, the NSA intends to hire Booz Allen to train the company selected for OPTIMAL DECISION were Booz Allen not awarded the contract. (Id. 131:10–24.) Accordingly, “domain knowledge” is not some immutable characteristic, but a transferrable trait—a trait that Booz Allen may bestow upon any number of potential contractors. Antitrust law generally does not cater to the preferences of a single consumer. Courts across the country have held that “the preferences of a single purchaser cannot define a product market.”

City of New York v. Grp. Health Inc., No. 06-cv-13122, 2010 WL 2132246, at *4 (S.D.N.Y. May 11, 2010);32 see also Global Disc. Travel Servs., LLC v. Trans World Airlines, Inc., 960 F. Supp. 701, 705 (S.D.N.Y. 1997) (Sotomayor, J.) (explaining that any given consumer may choose “Pepsi because she prefers the taste, or NBC because she prefers ‘Friends,’” but for antitrust purposes “Pepsi is one of many sodas, and NBC is just another television network”). Although the NSA prefers certain qualities in their mission partners, “preference[s] alone” cannot justify excluding other reasonably interchangeable companies from the relevant market definition. See Oksanen v. Page Mem’l Hosp., 945 F.2d 696, 709 (4th Cir. 1991) (en banc) (citations omitted).

32 Unpublished cases are cited for the soundness of their reasoning, not for their precedential value. Courts have questioned relevant market definitions based on a single contract. A single contract may carry the signs of a product market. Contracts frequently involve products created by “specialized vendors” for “distinct customers” under specific contractual requirements. See Brown Shoe, 370 U.S. at 325. But multiple courts have found that “a loss by the plaintiff of a single

contract with a single purchaser is simply not equivalent to a deleterious effect on the market.” Havoco of Am., Ltd. v. Shell Oil Co., 626 F.2d 549, 558 (7th Cir. 1980). “[T]here must be some allegation of a harmful effect on a more generalized market than [a single contracting party].” Id. If the Government’s theory were correct, then whoever wins OPTIMAL DECISION could hypothetically have a monopoly over the market for signals intelligence modeling and simulation. Indeed, if the court found that OPTIMAL DECISION constituted a market, “the mere fact that one party bid successfully against another party for [the] contract would be equivalent to an anticompetitive effect.” See id. The “specter of an antitrust action” should not be raised so easily during the ordinary course of competition. Id.; see also United States ex rel. Blaum v. Triad Isotopes, Inc., 104 F. Supp. 3d 901, 925 (N.D. Ill. 2015) (“While Defendants’ actions clearly had

an anticompetitive effect, that effect extended only to a single (albeit substantial) contract.”); Smalley & Co. v. Emerson & Cuming, Inc., 808 F. Supp. 1503, 1512 (D. Colo. 1992), aff’d, 13 F.3d 366 (10th Cir. 1993) (“A relevant product market defined as one product sold to one customer does not make sense under the antitrust laws.”). To be clear, the court does not suggest a single contract can never be a relevant market. If “a single purchaser represented the entire market for a commodity,” establishing a Section 1 violation becomes more likely. See Havoco, 626 F.2d at 559 n.6 (emphasis added). But the “commodity” here—modeling and simulation services—has a much broader market than the Government suggests. A single purchaser cannot, as the Government argues here, claim their particular use or application of a commodity is a market separate from the commodity itself. The Government relies on cases involving military technology and weaponry purchased only by the United States military or foreign sovereigns. See, e.g., Northrop Corp. v. McDonnell

Douglas Corp., 705 F.2d 1030, 1037, 1055 (9th Cir. 1983), cert. denied, 464 U.S. 849 (1983) (F- 18 fighter jets); Grumman Corp. v. LTV Corp., 527 F. Supp. 86, 89–90 (E.D.N.Y. 1981) (carrier- based aircraft for the United States Navy), aff’d, 665 F.2d 10 (2d Cir. 1981); FTC v. Alliant Techsystems Inc., 808 F. Supp. 9, 15 (D.D.C. 1992) (120 millimeter ammunition for M1 Abrams tanks). Because the NSA exclusively uses signals intelligence modeling and simulation services, the Government argues this case is no different. But modeling and simulation services exist in a substantial market beyond the NSA. (See Prelim. Inj. Hr’g Tr. Sept. 15 a.m. (Dunshee) 99:3–11, noting “computer networks all over the United States” use modeling and simulation.) The same cannot be said for F-18s, aircraft carriers, and tank ammunition.33 And while there remains a broader market for ammunition generally, the weapons and vehicles at issue in those cases are

much more individually distinctive than modeling and simulation services. Take the Alliant case for instance. The 120 millimeter tank rounds were not interchangeable with other types of ammunition. The contractors made those munitions specifically for the United States Army’s M1A1 and M1A2 Abrams battle tanks. See Alliant, 808 F. Supp. at 14. That the NSA uses modeling and simulation services in a specific environment (signals intelligence) does not mean the NSA’s application of these services constitutes a unique market.

33 Other courts have distinguished these military weaponry cases in a similar manner. See Int’l Logistics Grp., Ltd. v. Chrysler Corp., No. 85-CV-70047-DT, 1988 WL 106905, at *8 (E.D. Mich. Apr. 8, 1988), aff’d, 884 F.2d 904 (6th Cir. 1989) (“In Northrop, it is true that there were no other customers for the F–18 weapons system; but neither were there any other customers for anything that could be regarded as a reasonable substitute for this weapons system.”). Dr. Chicu’s attempt to define the relevant market with a “hypothetical monopolist test” falls short. Under the test, “if a hypothetical monopolist could impose a small but significant non- transitory increase in price (‘SSNIP’) in the proposed market, the market is properly defined.” FTC v. Penn. State Hershey Med. Ctr., 838 F.3d 327, 338 (3d Cir. 2016) (citing Department of Justice

and Federal Trade Commission Merger Guidelines, § 4, at 7–8.2) (footnote omitted). But if consumers respond to the SSNIP in a way that makes the SSNIP unprofitable, the proposed market definition is too narrow. Id. Here, Dr. Chicu never analyzed the profitability of any rate increase by a hypothetical monopolist in the proposed market, so his test has limited use. (See Prelim. Inj. Hr’g Tr. Sept. 15 p.m. (Chicu) 106:4–8.) Further, Dr. Chicu looked solely at rate increases on an actual contract, not those of a hypothetical monopolist.34 The court need not define the relevant market. That burden was on the Government, and it failed to carry it. And because the Government has not shown direct or indirect evidence of competitive harm, the court need not continue past the first step in the rule of reason. On this record, the Government has not shown it is likely to prevail on the merits.

B. Irreparable Harm As to irreparable harm, the Government must show the harm is “neither remote nor speculative, but actual and imminent.” Direx Israel Ltd. v. Breakthrough Med. Corp., 952 F.2d 802, 812 (4th Cir. 1991) (citations omitted). “Issuing a preliminary injunction based only on a possibility of irreparable harm is inconsistent with [the Supreme Court’s] characterization of injunctive relief as an extraordinary remedy that may only be awarded upon a clear showing that

34 And even in looking at that contract, Dr. Chicu identified a non-competitive extension of MASON III as an extension with competition. See infra Section III.A.2.ii. the plaintiff is entitled to such relief.” See Winter, 555 U.S. at 22 (citing Mazurek v. Armstrong, 520 U.S. 968, 972 (1997) (per curiam)).35 Here, the Government failed to show likely or significant competitive harm. See supra Section III.A.2.ii. And the fact the NSA may award OPTIMAL DECISION whenever it chooses

means the harm is not necessarily imminent. C. Balance of Equities and the Public Interest “A preliminary injunction is an extraordinary remedy never awarded as of right.” See Winter, 555 U.S. at 24 (citing Munaf v. Geren. 553 U.S. 674, 689–90 (2008)); see also Micro Strategy, Inc. v. Motorola, Inc., 245 F.3d 335, 339 (4th Cir. 2001) (citations omitted) (noting a preliminary injunction is a remedy “granted only sparingly and in limited circumstances”). The balance of equities and consideration of the overall public interest in this case support denying the Government’s motion. The Government’s alleged window of effective competition is rapidly closing. OPTIMAL DECISION bids are due in just a few weeks. The uncertainty caused by this litigation has already delayed the defendants’ ability to merge. EverWatch employees have

been left in limbo wondering what the future holds; some opted to leave rather than waiting to see the merger through. (Prelim. Inj. Hr’g Tr. Sept. 16 a.m. (Cooper) 56:9–57:7.) To some extent, the very existence of this litigation gave the Government what it was looking for—competition fueled by uncertainty. The Government’s modified request for relief would still let the defendants merge

35 The Government argues that it does not need to show irreparable harm because such harm is presumed in cases brought by the United States under statutes permitting injunctive relief. The court disagrees. The primary case for the Government’s proposition is In re Sanctuary Belize Litig., 409 F. Supp. 3d 380, 396, 419 (D. Md. 2019). But that case concerned a specific statute stating a different standard for preliminary relief for the FTC, not the DOJ. Generally, a plaintiff must satisfy each of the four factors to prevail on a motion for a preliminary injunction. See eBay Inc. v. MercExchange, L.L.C., 547 U.S. 388, 391 (2006) (noting preliminary injunction requirements based on “well- established principles of equity” in the context of the Patent Act); see also Salinger v. Colting, 607 F.3d 68, 77–78 (2d Cir. 2010) (“[N]othing in the text or the logic of eBay suggests that its rule is limited to patent cases. On the contrary, eBay strongly indicates that the traditional principles of equity it employed are the presumptive standard for injunctions in any context.”). so long as they do not finalize the deal until 90 days after they submit their OPTIMAL DECISION bids, along with EverWatch retaining the right to walk away from the merger. (ECF 200-1, Pl.’s Revised Proposed Order.) But the Government has not responded to the defendants’ claim that pushing the closing deadline any further would kill the deal. Denying the preliminary injunction

gives the defendants the ability to merge on their own terms, if they so choose. D. Final Judgment The court reserved the right to enter a final order on dispositive issues after the preliminary injunction hearing. (ECF 73, Order.) Although the court sides with the defendants at this procedural stage, the court is inclined to let the Government pursue its claim under Section 7 of the Clayton Act. To be sure, the market definition inquiry under Section 1 of the Sherman Act overlaps with the Government’s Section 7 claim to some extent. But given the expedited nature of these proceedings, the abbreviated record, and the condensed procedural format of the hearing, the court declines to enter a final judgment in the defendants’ favor for now. IV. CONCLUSION

For the reasons stated here, the Government’s Motion for a Preliminary Injunction is Denied. A separate Order follows.

10/11/2022 /s/ Date Catherine C. Blake United States District Judge