Shareholder Representative Services, LLC v. Alexion Pharmaceuticals Inc.

Opinion

IN THE COURT OF CHANCERY OF THE STATE OF DELAWARE

SHAREHOLDER REPRESENTATIVE ) SERVICES LLC solely in its capacity ) as representative of the Securityholders, ) ) Plaintiff, ) ) v. ) C.A. No. 2020-1069-MTZ ) ALEXION PHARMACEUTICALS, ) INC., ) ) Defendant. )

MEMORANDUM OPINION Date Submitted: March 4, 2025 Date Decided: June 11, 2025

Michael A. Barlow, QUINN EMANUEL URQUHART & SULLIVAN, LLP, Wilmington, Delaware; Andrew M. Berdon, Angus Chen, Alexandria Deep Conroy, Courtney C. Whang, QUINN EMANUEL URQUHART & SULLIVAN, LLP, New York, New York; Joseph M. Paunovich, David M. Elihu, James Bieber, Andrew Brayton, QUINN EMANUEL URQUHART & SULLIVAN, LLP, Los Angeles, California, Attorneys for Plaintiff and Counterclaim Defendant Shareholder Representative Services LLC.

David E. Wilks, Scott B. Czerwonka, WILKS LAW, LLC, Wilmington, Delaware; Deborah E. Fishman, Carson D. Anderson, ARNOLD & PORTER KAYE SCHOLER LLP, Palo Alto, California; Daniel L. Reisner, Jeffrey A. Fuisz, Angela R. Vicari, Matthew M. Wilk, ARNOLD & PORTER KAYE SCHOLER LLP, New York, New York; Howard Sklamberg, Jeremy Cobb, ARNOLD & PORTER KAYE SCHOLER LLP, Washington, DC, Attorneys for Defendant and Counterclaim Plaintiff Alexion Pharmaceuticals, Inc.

ZURN, Vice Chancellor. Defendant Alexion Pharmaceuticals, Inc. purchased nonparty Syntimmune,

Inc. to develop a drug to treat rare diseases. The merger agreement promised discrete

lump sum payments to Syntimmune’s former stockholders upon the achievement of

development milestones, and obligated Alexion to use commercially reasonable

efforts to achieve those milestones. The agreement designated plaintiff Shareholder

Representative Services, LLC (“SRS”) as the former Syntimmune stockholders’

representative.

Alexion terminated the drug development program. SRS sued for breach of

the efforts obligation. After trial, I concluded Alexion had breached its efforts

obligation by terminating the drug development program.

With the benefit of supplemental briefing, this opinion addresses the

expectation damages Alexion owes for that breach. Because the earnout provision

provides for lump sum payments for contingent events, this decision employs an

expected value approach. It calculates damages by weighting each milestone’s

earnout payment by its probability of success, discounted to present value at the time

of breach. It calculates that SRS is entitled to $180,944,915.32 in damages for

Alexion’s breach of its efforts obligation, plus pre- and post-judgment interest.

1 I. BACKGROUND

This decision relies on the factual findings set forth in the post-trial opinion

on liability (the “September Opinion”) and the trial record. 1 The facts set forth

herein were proven by a preponderance of the evidence at trial.

A. The Syntimmune Merger And Earnout Agreement

In September of 2018, Alexion acquired Syntimmune to develop and

commercialize a monoclonal antibody that became known as ALXN1830. 2 The

purchase price included $400 million up front and $800 million in earnout payments

tied to eight development milestones. 3 Milestone 1 provided for a $130 million

payment upon the completion of a successful Phase 1 Clinical Trial, as defined by

the Merger Agreement.4 The September Opinion concluded Milestone 1 had been

1 S’holder Representative Servs. LLC v. Alexion Pharms., Inc., 2024 WL 4052343 (Del. Ch. Sept. 5, 2024) [hereinafter “Sept. Op.”]. This opinion assumes familiarity with the September Opinion and uses its defined terms and citation formats. Citations in the form “SRS Op. Suppl. Br. —” refer to SRS’s post-trial opening supplemental damages brief, available at docket item (“D.I.”) 384. Citations in the form “ALXN Ans. Suppl. Br. —” refer to Alexion’s post-trial answering supplemental damages brief, available at D.I. 392. Citations in the form “SRS Reply Suppl. Br. —” refer to SRS’s post-trial supplemental damages reply brief, available at D.I. 398. A leading treatise on damages observes, “Only so much judicial time can be used to investigate the precise losses suffered or the gains received from a contract breach.” 3 Dan B. Dobbs, Law of Remedies: Damages—Equity—Restitution § 12.1(2), at 18 (2d ed. 1993) [hereinafter “Dobbs”]. This opinion has probably exceeded whatever that amount of time should be. 2 Merger Agr. §§ 1.1, 3.8(b). 3 Id. § 3.8(b). 4 Id. § 3.8(a)(i). 2 achieved, held Alexion breached its contractual obligation to pay SRS $130 million

upon achievement of that milestone, and awarded damages in that amount. 5

Under the Merger Agreement, Alexion promised earnout payments for the

successful completion of Milestones 2 through 8, as follows 6:

Earnout Provision Summary for Milestones 2 Through 8

Milestone Milestone Amount Triggering Event First dosing of the first patient in a Pivotal Clinical 2 $ 120,000,000.00 Trial for any first Indication. First dosing of the first patient in a Pivotal Clinical 3 $ 120,000,000.00 Trial for a second Indication. Receipt of Regulatory Approval from the FDA for 4 $ 150,000,000.00 any first Indication. Receipt of Regulatory Approval from the FDA for 5 $ 150,000,000.00 a second Indication. Receipt of Regulatory Approval from the EMA for 6 $ 25,000,000.00 any first Indication. Receipt of Regulatory Approval from the EMA for 7 $ 25,000,000.00 a second Indication. The determination at the end of Alexion’s fiscal year that the Net Sales for such fiscal year across 8 $ 80,000,000.00 all Indications equals or exceeds One Billion Dollars ($1,000,000,000).

The Merger Agreement provides that Milestones 6 and 7 “shall be achieved upon

receipt of the applicable reimbursement and/or pricing approval from the applicable

Governmental Entity in three (3) out of the following five (5) countries: United

5 Sept. Op. at *48. 6 Merger Agr. § 3.8(a)(ii)–(viii). 3 Kingdom, France, Italy, Germany or Spain.” 7 Each milestone payment is due forty-

five days after the milestone’s achievement.8

To propel ALXN1830 toward those milestones, the Merger Agreement

required Alexion to use Commercially Reasonable Efforts (“CREs” and the “CRE

Obligation”), as defined by the agreement for seven years. 9

This opinion defines a “Milestone Event” as the achievement of each

milestone, noted in the form 𝑀𝑀𝑖𝑖 , where i represents a given Milestone Event number.

This opinion notates the probability of a Milestone Event (each a “Milestone

Probability”) as 𝑃𝑃(𝑀𝑀𝑖𝑖 ). This opinion also deals with conditional probabilities, that

is, the probability that an event will occur given that some other event occurred. It

notates the probability that a later Milestone Event will occur given that an earlier

Milestone Event occurred as 𝑃𝑃�𝑀𝑀𝑖𝑖 �𝑀𝑀𝑗𝑗 �. For example, the conditional probability of

𝑀𝑀3 given 𝑀𝑀2 is notated 𝑃𝑃(𝑀𝑀3 |𝑀𝑀2 ).

B. Syntimmune’s Largest Former Stockholder Values Its Right To Milestone Payments. Shortly after Alexion acquired Syntimmune, Syntimmune’s largest former

stockholder Apple Tree Partners (“ATP”) valued its right to future distributions from

Milestones 2 through 8 based on the milestone amounts and probabilities of

7 Id. § 3.8(c). 8 Id. § 3.8(e). 9 Id. § 3.8(f). 4 achievement.10 ATP estimated the Milestone Probabilities “[b]ased on discussions

with management and considering the current status of clinical trials” as well as

“observed clinical trial success rates.”11

ATP Milestone Probability Estimates Milestone Event Probability 𝑀𝑀2 0.80 𝑀𝑀3 0.75 𝑀𝑀4 0.53 𝑀𝑀5 0.45 𝑀𝑀6 0.53 𝑀𝑀7 0.45 𝑀𝑀8 0.10 12

10 JX 2962; Hall Tr. 107, 109–16. 11 JX 2962 at 22–26. ATP discounted to present value based on estimates of when the milestones would be achieved. ATP’s probabilities for Milestones 4 through 7 assume the probability of EMA approval is the same as the probability of FDA approval for a given indication. See id. at 23. 12 ATP’s valuation report does not directly estimate 𝑃𝑃(𝑀𝑀8 ) as 0.10. ATP valued its Milestone 8 distribution using a repeated random sampling technique called a Monte Carlo simulation, using 100,000 simulation paths. Id. at 24–25, 30 n.2, 32, 34. The 0.10 probability can be inferred from ATP’s report. ATP calculated that if Milestone 8 were achieved, it would receive $57,886,923. ATP assumed that if the milestone were achieved, it would be achieved in 2026. Id. at 24. In Monte Carlo “simulation paths where the total simulated sales across the three indications was shown to be equal to or greater than $1 billion,” ATP’s projected distribution “was discounted back to present value using a discount rate” of 4.49% over 8.16 years. Id. at 25, 32. Based on those figures, the present value of ATP’s distributions in successful simulations was $40,451,131.57. ATP’s report states the “average present value across 100,000 different simulation paths” was $4,047,941. Id. at 25. Because unsuccessful simulation paths have a present value of $0 (and because ATP always discounted to present value using the same discount rate and timing for successful paths), the average present value of ATP’s distributions is

5 C. The ALXN1830 Program

Alexion initially focused the ALXN1830 program on the PV, gMG, and

WAIHA indications. 13 But ALXN1830 faced significant development obstacles

after the merger, including a contaminated drug supply and adverse patient reactions

that forced it to pause several clinical trials. 14 The emergence of COVID-19 halted

all of Alexion’s trials, while its competitors were able to push ahead.15 In April

2020, Alexion shifted funding away from ALXN1830, further delaying its

development.16 The next month, Alexion decided the PV indication was not worth

pursuing. 17

But the program regained some momentum. In March 2021, Alexion began

dosing in a Phase 1 trial in healthy volunteers called HV-108. 18 Alexion also

equal to the success rate multiplied by the present value of a successful simulation path, as follows: $4,047,941 = 𝑅𝑅 ∙ $40,431,131.57, where 𝑅𝑅 is the success rate. See Joseph K. Blitzstein & Jessica Hwang, Introduction to Probability 149–50 (2015). Solving that equation yields a success rate of about 10%, which corresponds to the probability of success for Milestone 8. 13 JX 1229 at 1–2; JX 1424 at 1–2; JX 609 at 2, 12; JX 697 at 1. 14 Ledwith Tr. 1029–31, 1038; see JX 923 at 39; JX 1139.04 at 25. 15 See Ledwith Tr. 1063, 1070; JX 1333 at 3; JX 1337; JX 1659; JX 1994; JX 2272; JX 2296; JX 2302; JX 2349; JX 2359; JX 2388; JX 2415; JX 2451; JX 2486; JX 2569; JX 2570; JX 2583; JX 2745; JX 2791. 16 JX 1451 at 2–4; Orloff Tr. 862. 17 JX 1477. 18 JX 2367 at 1; Pirozzi Tr. 1460. 6 planned Phase 2 studies in gMG and WAIHA, even though it was clear ALXN1830

would be later to market than originally anticipated relative to its competitors. 19

In July 2021, Alexion was acquired by AstraZeneca plc. 20 AstraZeneca had

promised its shareholders $500 million in recurring synergies from the acquisition,

so Alexion launched a full portfolio review of its drug programs. 21 Soon after the

merger, Alexion deprioritized the ALXN1830 gMG and WAIHA programs in favor

of TED and cAMR.22 None of Alexion’s competitors were pursuing TED or cAMR

treatments, so Alexion believed it could be the first to market in these indications.23

Alexion used an internal metric for probability of technical and regulatory

success (“PTRS”) “as a guide” to assess its programs.24 PTRS has two components:

probability of technical success, and probability of regulatory success given

technical success.25 Overall PTRS is found by multiplying the two components and

19 JX 2298 at 1–2, 8; JX 2299 at 1–2, 8; JX 1699 at 7; JX 609 at 12. 20 JX 1865 at 3. 21 See JX 1946 at 3, 7–9; JX 1933 at 1; JX 1997; Washburn Tr. 638. 22 See JX 1928; JX 1933 at 1; JX 2021 at 1; JX 2042 at 1; JX 2065 at 1; Lee Tr. 445, 457; see also JX 1948; Russell Tr. 732. 23 See JX 2226 at 2; JX 1955 at 7; Russell Tr. 772. 24 Lee Tr. 476. 25 See Borboroglu Tr. 1403. The technical success component is further broken into the conditional probabilities of success at each stage of preclinical and clinical testing. See JX 1863 at 53. 7 maps onto the probability of FDA approval from the outset. 26 Shortly before

receiving HV-108 data, Alexion estimated its cAMR program had a 50% chance of

a successful Phase 2 study. 27 Alexion set the cAMR program’s overall PTRS at

34%.28 As for the TED program, Alexion estimated a 43% chance of a successful

Phase 2 study 29 and an overall PTRS of 30%. 30

26 Borboroglu Tr. 1403 (“Technical success assigned by our clinical teams; regulatory success assigned by our regulatory teams. You multiply the two. That’s PTRS.”); Washburn Tr. 645 (“[PTRS] is a way to assess at different stages of a development program what the probability is that you will be successful in getting a regulatory approval as well as a compound . . . that can be manufactured and delivered.” (emphasis added)). The preponderance of the evidence at trial showed an industry understanding of near-universal overlap in approval by the FDA and EMA, supporting the assumption that FDA and EMA approval go hand in hand. SRS’s antibody development expert assumed EMA approval was guaranteed upon FDA approval, and Alexion assumed the same in its internal projections. Kinch Tr. 353–54, 413; JX 2498 at 73; see JX 690 at 276; JX 2962 at 23. So overall PTRS may be thought of as the probability of FDA approval, the probability of EMA approval, or the probability of FDA and EMA approval. Neither the record nor the parties suggest that “regulatory success” requires obtaining the “reimbursement and/or pricing approval[s]” for Milestones 6 and 7. Merger Agr. § 3.8(c). 27 JX 1863 at 53. The PTRS data show a 100% chance of a successful Phase 1 study and a 50% chance of a successful Phase 2 study given a successful Phase 1 study. Therefore, the data show a 50% chance of a successful Phase 2 study from the outset. See id. 28 Id. 29 See JX 2608 at 4; JX 2225 at 5. The PTRS data show a 100% chance of a successful Phase 1 study and a 43% chance of a successful Phase 2 study given a successful Phase 1 study. Therefore, the data show a 43% chance of a successful Phase 2 study from the outset. See JX 2608 at 4. 30 See JX 2608 at 4; JX 2225 at 5. 8 In August 2021, HV-108 was paused due to a COVID-19 outbreak. 31 Alexion

received preliminary data from the paused study in September. The data showed a

high immunogenicity rate, 32 but that was not news to Alexion. 33 A preliminary

assessment from the day the data was received showed there was no impact on an

important efficacy indicator, and the “safety profile remain[ed] unchanged.” 34 But

by the next day, “the current view [at Alexion was] that development of 1830 [was]

going to be stopped.”35

Alexion received additional HV-108 data in November. An outside

consultant Alexion hired to analyze the data concluded that the “detected ADA

response d[id] not appear to compromise overall benefit vs. risk,” and there was “an

adequate weight of evidence to resume study HV-108.”36 But Alexion had made up

its mind about ALXN1830. In response to the HV-108 data, Alexion reduced TED’s

probability of a successful Phase 2 study from 43% to 20%,37 and reduced TED’s

31 JX 1972.02. 32 Immunogenicity rates were determined based on the presence of antidrug antibodies in study subjects. See Kinch Tr. 248–49. 33 JX 1987 at 3; JX 2006 at 1. 34 JX 1990 at 2. 35 Id. at 1. 36 JX 2189 at 26. 37 JX 2608 at 4. The 20% figure was the conditional likelihood of a successful Phase 2 study given a successful Phase 1 study. But the probability of a successful Phase 1 study continued to be 100%. Id. 9 overall PTRS from 30% to 10%. 38 Alexion also reduced cAMR’s overall PTRS to

10%.39

Alexion decided to terminate the program on December 14, 2021. 40 The

September Opinion held that breached Alexion’s CRE Obligation. 41 The September

Opinion determined “[t]he preponderance of the evidence supports the conclusion

that the decision was influenced, motivated by, or driven by AstraZeneca’s pursuit

of merger synergies.”42

D. Dr. Michael Kinch’s Testimony

At trial, SRS’s antibody development expert Dr. Michael Kinch opined about

the probability of achieving Milestones 2 through 5 had Alexion used CREs. Kinch

offered calculations based on his open-source database “of experimental medicines

and their likelihood of being approved” called the Clinical Drug Experience

Knowledgebase (“CDEK”). 43 Kinch built CDEK to provide a database for

academics who could not afford the high subscription costs of the private databases

pharmaceutical companies use to evaluate the probability of success for drug

38 JX 2225 at 13–14. 39 Id. at 4, 13–14. The record does not indicate how much Alexion reduced cAMR’s probability of a successful Phase 2 study. 40 JX 2226 at 2. 41 Sept. Op. at *41. 42 Id. at *48. 43 E.g., Kinch Tr. 207–09, 218–19, 341–42, 349–54, 361–65; see JX 2498 at 66–76. 10 development programs. 44 CDEK only includes publicly available information based

on the reported progress of drug molecules. 45 Kinch acknowledged failed clinical

trials are often underreported and that including that missing data “would increase,

potentially, the likelihood of failure” predicted using CDEK. 46

Based on his comparisons to similar molecules in CDEK, Kinch estimated the

probability of 𝑀𝑀2 was between 0.582 and 1. 47 He opined that the probability of 𝑀𝑀4

given that 𝑀𝑀2 occurred was 0.684. 48

44 Kinch Tr. 208–10, 385–86. 45 Id. at 376. 46 Id. at 377–79; see also Jagannathan Tr. 1336–37. 47 Kinch Tr. 348; see JX 2498 at 68–69. Kinch understood that estimating 𝑃𝑃(𝑀𝑀2 ) required him to estimate ALXN1830’s likelihood of reaching a PCT in any first indication, as opposed to estimating a given indication’s likelihood of advancing to a PCT. Kinch Tr. 220 (“Q. So taking a look at milestone No. 2, Dr. Kinch, what do you understand that to require in order to trigger the payment of milestone No. 2? A. Well, as indicated here, it’s the first dosing of the first patient in a pivotal clinical trial for any first indication.” (emphasis added)). 48 Kinch Tr. 348; JX 2498 at 68–69 (estimating the probability of “receiving FDA approval for a first indication, and therefore, reaching Milestone 4” by multiplying the probability of reaching Milestone 2 by 0.684). This is an estimate of ALXN1830’s likelihood of receiving FDA approval in any first indication given that it reached a PCT in some first indication. Built into this probability is the possibility that ALXN1830 might achieve Milestone 4 with the following sequence: (1) reach a PCT in indication X; (2) reach a PCT in indication Y; (3) receive FDA approval for indication Y. 11 Kinch opined the probability of 𝑀𝑀3 given that 𝑀𝑀2 occurred was equal to the

initial probability of 𝑀𝑀2 . 49 Similarly, he opined the probability of 𝑀𝑀5 given that 𝑀𝑀4

occurred was equal to the probability of 𝑀𝑀4 . 50

Kinch’s relevant opinions are summarized below:

Kinch’s Opinions Based on CDEK Opinion Notation The probability of 𝑀𝑀2 is between 0.582 and 0.582 ≤ 𝑃𝑃(𝑀𝑀2 ) ≤ 1. 1. The probability of 𝑀𝑀4 given that 𝑀𝑀2 𝑃𝑃(𝑀𝑀4 |𝑀𝑀2 ) = 0.684. occurred is 0.684. The probability of 𝑀𝑀3 given that 𝑀𝑀2 𝑃𝑃(𝑀𝑀3 |𝑀𝑀2 ) = 𝑃𝑃(𝑀𝑀2 ). occurred is equal to the probability of 𝑀𝑀2 . The probability of 𝑀𝑀5 given that 𝑀𝑀4 𝑃𝑃(𝑀𝑀5 |𝑀𝑀4 ) = 𝑃𝑃(𝑀𝑀4 ). occurred is equal to the probability of 𝑀𝑀4 .

Regarding Milestones 6 and 7, Kinch’s opinion was “limited to an assessment

of the likelihood of regulatory approval from the EMA and d[id] not consider the

likelihood of reimbursement and/or pricing approval from the EMA.” 51 Kinch

49 See Kinch Tr. 349–50 (indicating he assumed that for all relevant metrics, ALXN1830’s probability of success in a second indication given success in a first indication was equal to the molecule’s probability of success in a first indication); JX 2498 at 76 n.3. Importantly, this is very different than suggesting 𝑀𝑀2 was just as likely as 𝑀𝑀3 from the outset. If 𝑀𝑀2 and 𝑀𝑀3 had the same probability, that would mean ALXN1830’s chances of reaching a PCT in at least two indications were the same as its chances of reaching a PCT in at least one indication. In other words, it would imply that a second indication was guaranteed to reach a PCT upon a first indication doing so. That would have been great news for monoclonal antibody research. 50 See Kinch Tr. at 349–50; JX 2498 at 76 n.5. 51 JX 2498 at 10 n.22; Kinch Tr. 413. 12 testified to a near-universal overlap between FDA and EMA approval.52 He made

partial calculations for 𝑃𝑃(𝑀𝑀6 ) and 𝑃𝑃(𝑀𝑀7 ) based on the assumption that EMA

approval was guaranteed for any indication that received FDA approval. 53 Kinch

noted Alexion made the same assumption in its internal projections. 54 Kinch’s

calculations did not account for the requisite country-specific approvals for

Milestones 6 and 7.

E. John Russell’s Testimony

John Russell was SRS’s expert on the evaluation of the pharmaceutical

competitive landscape, market opportunities, and commercial potential and pricing

and reimbursement.55 Russell testified that if ALXN1830 received EMA approval,

it was likely to receive the country-specific approvals to satisfy Milestones 6 and

7. 56

Russell also testified, based on Alexion’s 2021 global revenue projections,

that ALXN1830 had the potential to achieve Milestone 8 if it obtained approval.57

52 Kinch Tr. 353–54. 53 See id. at 413; JX 2498 at 76. 54 Kinch Tr. 352–54; see JX 690 at 276. 55 Russell Tr. 669. 56 Id. at 672. 57 Id. at 744–45 (“Q. So did you form an opinion as to whether ALXN1830 still had the potential to achieve up to a billion dollars in net sales? A. Yes, it did, based on looking at this data and analyzing it, yes. Q. And was Alexion’s internal forecasting consistent with that? A. Yes, it was. Yes.”). 13 The Alexion model he relied on forecasted a peak of $1.35 billion in combined

annual revenues for TED and cAMR, and revenues over $1 billion for the same two

indications in four more years. 58

II. ANALYSIS

SRS pursues two alternative paths to damages regarding Milestones 2 through

8. SRS seeks damages for Alexion’s breach of its CRE Obligation.59 It also seeks

damages under another breach of contract theory: SRS contends Alexion breached

its obligation under Section 3.8(f) of the Merger Agreement not to take any action

the primary purpose of which is to avoid the achievement of any milestone (the

“Non-Avoidance Obligation”).60 I begin with SRS’s first theory.

A. Alexion Owes Expectation Damages For Breach Of Its CRE Obligation. SRS proved Alexion breached its CRE Obligation, as the September Opinion

explained. “Under Delaware law, the standard remedy for breach of contract is

based on the reasonable expectations of the parties that existed before or at the time

of the breach.”61 “This principle of expectation damages is measured by the amount

of money that would put the promisee in the same position as if the promisor had

58 JX 1953. 59 SRS Op. Suppl. Br. 1–5, 9–31; see D.I. 155 ¶¶ 252–65 [hereinafter “Compl.”]. 60 SRS Op. Suppl. Br. 5–6, 32–39; Merger Agr. § 3.8(f); see Compl. ¶¶ 272–78. PharmAthene, Inc. v. Siga Techs., Inc. (“Siga I”), 2014 WL 3974167, at *7 (Del. Ch. 61

Aug. 8, 2014), aff’d, 132 A.3d 1108 (Del. 2015). 14 performed the contract.” 62 To recover expectation damages, the plaintiff must prove

the fact of damage—that the breach “caused it injury”—to a reasonable certainty,

and an estimate of the amount of damage.63 The measure of expectation damages is

informed by the nature of the contractual right. 64 My analysis breaks SRS’s burden

into three parts: injury, causation, and an estimate of damage.

62 Duncan v. Theratx, Inc., 775 A.2d 1019, 1022 (Del. 2001). 63 Fortis Advisors LLC v. Johnson & Johnson, 2024 WL 4048060, at *35 (Del. Ch. Sept. 4, 2024); Siga Techs., Inc. v. PharmAthene, Inc. (“Siga II”), 132 A.3d 1108, 1111 (Del. 2015) (“[W]hen a contract is breached, expectation damages can be established as long as the plaintiff can prove the fact of damages with reasonable certainty.” (emphasis omitted)); see also Cura Fin. Servs. N.V. v. Elec. Payment Exch., Inc., 2001 WL 1334188, at *19–20 (Del. Ch. Oct. 22, 2001) (“[R]easonable certainty is not equivalent to absolute certainty; rather, the requirement that plaintiff show defendant’s breach to be the cause of his injury with ‘reasonable certainty’ merely means that the fact of damages must be taken out of the realm of speculation.” (quoting Tanner v. Exxon Corp., 1981 WL 191389 (Del. Super. July 23, 1981))); cf. Holland Loader Co. v. FLSmidth A/S, 769 F. App’x 40, 42 (2d Cir. 2019) (upholding ruling that the fact of damages based on lost earnout payments calculated as a percentage of gross sales during the earnout period—as opposed to a fixed value based on the achievement of milestones—had not been proven because the evidence “failed to indicate with reasonable certainty that, but for [defendant’s] breach, any sales would have been made during the five-year earnout period” (emphasis added)). 64 See Dobbs § 12.2(1), at 25 (“In some cases expectancy is measured by the market value of the performance promised at the date performance was due. In others, the plaintiff’s actual cost of getting a substitute performance, is the measure. In still others, expectancy may be protected only by the award of consequential or special damages such as lost profits or collateral expenses incurred because of the breach.” (footnotes omitted)). Compare Fortis, 2024 WL 4048060, at *35, *50–53 (measuring expectation damages for breaches that interfered with promisee’s ability to achieve contingent earnout payments by the expected value of the earnout payments at the time of breach), with Duncan, 775 A.2d at 1020–22 (measuring expectation damages where an issuer’s breach prevented stock trading “by calculating the difference between (1) the highest intermediate price of the shares during a reasonable time at the beginning of the restricted period, which functions as an estimate of the price that the stockholders would have received if they had been able to sell their shares, and (2) the average market price of the shares during a reasonable period after the restrictions were lifted”). 15 1. Injury

“If a breach is of a promise conditioned on a fortuitous event and it is uncertain

whether the event would have occurred had there been no breach, the injured party

may recover damages based on the value of the conditional right at the time of

breach.” 65 As recently observed in Fortis Advisors LLC v. Johnson & Johnson,

earnout provisions coupled to an efforts clause are contingent in nature: they require

the buyer to pay additional consideration if the buyer or target achieves specified

goals, buttressed by a standard to which the buyer must perform. 66 This design

allocates risk between the buyer and seller. The buyer reduces its risk of overpaying

for a business with uncertain prospects, and the seller takes on the risk that the

earnout payment may not be owed.67 Because an earnout payment is contingent, it

is uncertain what the injured party would have received absent the defendant’s

breach.

65 Restatement (Second) of Contracts § 348(3); see Fortis, 2024 WL 4048060, at *35, *50– 53; Maverick Therapeutics, Inc. v. Harpoon Therapeutics, Inc., 2021 WL 1592473, at *2, *10–11 (Del. Ch. Apr. 23, 2021) (measuring expectation damages for breach that interfered with a risky investment based on the diminution of value of the “chance of winning” on the investment); Kansas City, M. & O. Ry. Co. v. Bell, 197 S.W. 322, 323 (Tex. Civ. App. 1917) (holding the plaintiff was entitled to the value of the chance that his hogs would have won a competition if delivered on time and that “the probability that the plaintiff would be successful in the competition would be admissible” evidence). 66 Fortis, 2024 WL 4048060, at *21–23. 67 Brian JM Quinn, Putting Your Money Where Your Mouth Is: The Performance of Earnouts in Corporate Acquisitions, 81 U. Cin. L. Rev. 127, 140–41 (2012); Fortis, 2024 WL 4048060, at *21. 16 In Fortis, the buyer’s breach of its efforts obligation to pursue earnout

milestones made it impossible to achieve certain milestones.68 Fortis identified a

cognizable, and compensable, injury in the decreased expected value of the right to

earnout payments, reasoning damages based on that injury would “put the promisee

in the same position as if the promisor had performed the contract.” 69 Fortis

calculated expected value at the time of breach by weighting each milestone by the

proven likelihood it would be achieved.70

That logic holds here. SRS’s injury is best understood as the lost expected

value of each milestone as compared before and after Alexion’s breach of its CRE

Obligation.71 As in Fortis, an expected value approach reflects the theory behind

expectation damages, which aim to put “the nonbreaching party in as good a position

as he would have been in had the contract been performed, and no better.” 72

Compensating for lost expected value, rather than with full value whenever earnout

payments are likely and zero value whenever earnout payments are unlikely, strives

68 See Fortis, 2024 WL 4048060, at *35–45, *50–53. 69 Id. at *51 (quoting Comrie v. Enterasys Networks, Inc., 837 A.2d 1, 17 (Del. Ch. 2003)). 70 Id. 71 See id. at *35, *50–53. 72 Dobbs § 12.2(1), at 23 (footnote omitted); see Duncan, 775 A.2d at 1022 n.6. 17 to hit the mark on the parties’ reasonable expectations, rather than award windfalls

for some promisees and goose eggs for others.73

Here, each milestone had an expected value of zero after Alexion’s breach.

The ALXN1830 program was terminated. To show a decrease in expected value for

a given milestone, SRS must prove, to a degree of reasonable certainty, that the

expected value was above zero immediately before the breach.74 To demonstrate an

73 SRS calls this approach the “loss of chance” approach. SRS Op. Suppl. Br. 26–31. The terminology makes no difference. What matters is that the approach avoids overcompensation “on the assumption that the [fortuitous condition] would have occurred” and undercompensation “on the ground of uncertainty.” Restatement (Second) of Contracts § 348 cmt. d. This is done by calculating damages “based on the value of [the] conditional contract right at the time of breach.” Id. Delaware law does not prohibit this approach in this context, contrary to Alexion’s suggestions. Alexion’s pretrial brief cites Sherman v. Ellis for the proposition that “Delaware law does not recognize the ‘Loss of Chance’ theory in breach of contract cases.” D.I. 313 at 58 (citing Sherman v. Ellis, 246 A.3d 1126, 1132 (Del. 2021)). Sherman declined to apply an “increased risk of harm” theory (typically used in medical malpractice cases) in a legal negligence action because causation for that claim “requires proof that, but for the attorney’s negligence, the plaintiff would have obtained a more favorable result.” Sherman, 246 A.3d at 1132–33 (quoting Sherman v. Ellis, 2020 WL 30393, at *13 (Del. Super. Ct. Jan. 2, 2020)). In other words, for a legal negligence claim, the causation standard precludes loss of expected value associated with the pursuit of an unlikely favorable result as a compensable injury. Sherman does not suggest an analogous principle applies in the breach of contract context. See generally id. The medical malpractice context uses the related, but different, “increased risk of harm” and “loss of chance” theories. United States v. Anderson, 669 A.2d 73, 75–76 (Del. 1995). Increased risk allows recovery for the increased risk of a future harm before that harm occurs. Loss of chance allows recovery only after the harm occurs. Id. 74 Restatement (Second) of Contracts § 348 cmt. d (“The value of that right must itself be proved with reasonable certainty, as it may be if there is a market for such rights or if there is a suitable basis for determining the probability of the occurrence of the event.”). 18 injury, it is sufficient to show each milestone had a nonzero probability of being

achieved.

The trial record offers four views of the Milestone Probabilities. First, ATP’s

analysis provides estimates around the time of the Syntimmune merger. 75 Next,

Alexion’s PTRS estimates establish Alexion’s view of the probabilities both before

and after receiving the HV-108 data.76 Finally, Kinch’s opinions based on his CDEK

data, paired with Russell’s opinions, provide estimates at the time of breach.77

ATP, PTRS, and CDEK all show each milestone had a nonzero probability of

being achieved. The probabilities based on PTRS and CDEK require some

adjustments. 78 Those adjustments are based on probabilistic reasoning and must be

explained through mathematical notation. A brief primer follows.

a. Mathematical Primer

The definitions and principles utilized in this opinion apply to arbitrary events,

𝐴𝐴 and 𝐵𝐵. The probability of any event 𝐸𝐸 is denoted 𝑃𝑃(𝐸𝐸). This opinion follows the

75 JX 2962 at 22–26; Hall Tr. 107, 109–16. 76 See JX 1863 at 53; JX 2225 at 4–5, 13–14; JX 2608 at 4. Kinch Tr. 207–09, 218–19, 341–42, 349–54, 361–65, 385–86; see JX 2498 at 72–76; 77

Russell Tr. 672, 744–45. 78 Explaining these adjustments requires noting rounded numbers. All calculations are performed using exact values. 19 standard convention of explaining mathematics using the first-person plural. 79

i. Conceptual Framework

An introductory text on probability provides a framework. 80

The mathematical framework for probability is built around sets. Imagine that an experiment is performed, resulting in one out of a set of possible outcomes. Before the experiment is performed, it is unknown which outcome will be the result; after, the result “crystallizes” into the actual outcome. . . .

The sample space 𝑆𝑆 of an experiment is the set of all possible outcomes of the experiment. An event 𝐴𝐴 is a subset of the sample space 𝑆𝑆, and we say that 𝐴𝐴 occurred if the actual outcome is in 𝐴𝐴. . . .

[T]he complement 𝐴𝐴𝑐𝑐 is the event that occurs if and only if 𝐴𝐴 does not occur. 81

79 Steven G. Krantz, A Primer of Mathematical Writing: Being a Disquisition on Having Your Ideas Recorded, Typeset, Published, Read, and Appreciated 33 (2d. ed. 2016). 80 Blitzstein & Hwang, supra note 12. 81 Id. at 3–4. 20 The probability of an event either occurring or not occurring is equal to one. So the

probability that 𝐴𝐴 does not occur is equal to one minus the probability that it does

occur:

𝑃𝑃(𝐴𝐴𝑐𝑐 ) = 1 − 𝑃𝑃(𝐴𝐴). 82

If another event 𝐵𝐵 is also a subset of the sample space 𝑆𝑆, then “the intersection

𝐴𝐴 ∩ 𝐵𝐵 is the event that occurs if and only if both 𝐴𝐴 and 𝐵𝐵 occur.”83

“[T]he union 𝐴𝐴 ∪ 𝐵𝐵 is the event that occurs if and only if at least one of 𝐴𝐴 [or] 𝐵𝐵

occurs.” 84

82 Id. at 21–23. 83 Id. at 4. 84 Id. 21 The area of the dashed region relative to the area of 𝑆𝑆 may be thought of as 𝑃𝑃(𝐴𝐴 ∪

𝐵𝐵). In the previous diagram for 𝐴𝐴 ∩ 𝐵𝐵, the areas of the circles representing 𝐴𝐴 and 𝐵𝐵

correspond to 𝑃𝑃(𝐴𝐴) and 𝑃𝑃(𝐵𝐵). 85 Notice that adding those areas yields a value

greater than 𝑃𝑃(𝐴𝐴 ∪ 𝐵𝐵) because the overlapping portion is counted twice. We may

find 𝑃𝑃(𝐴𝐴 ∪ 𝐵𝐵) by subtracting one of those instances from that sum:

𝑃𝑃(𝐴𝐴 ∪ 𝐵𝐵) = 𝑃𝑃(𝐴𝐴) + 𝑃𝑃(𝐵𝐵) − 𝑃𝑃(𝐴𝐴 ∩ 𝐵𝐵). 86

Lastly, in some cases, 𝐵𝐵 is a subset of 𝐴𝐴. We denote this 𝐵𝐵 ⊆ 𝐴𝐴.

85 See id. at 24. 86 Id. at 21–23. 22 The area of 𝐵𝐵 relative to the area of 𝑆𝑆 represents 𝑃𝑃(𝐵𝐵). The area of 𝐴𝐴 ∩ 𝐵𝐵 represents

𝑃𝑃(𝐴𝐴 ∩ 𝐵𝐵). But because 𝐵𝐵 and 𝐴𝐴 ∩ 𝐵𝐵 are the same region, their areas are the same.

So, if 𝐵𝐵 ⊆ 𝐴𝐴, then 𝑃𝑃(𝐴𝐴 ∩ 𝐵𝐵) = 𝑃𝑃(𝐵𝐵). 87

ii. Conditional Probability

The milestones also introduce the concept of conditional probability. The

conditional probability of 𝐵𝐵 given 𝐴𝐴 is denoted by 𝑃𝑃(𝐵𝐵|𝐴𝐴). If 𝐴𝐴 and 𝐵𝐵 are events

with 𝑃𝑃(𝐴𝐴) > 0, then the conditional probability of 𝐵𝐵 given 𝐴𝐴 is defined as

𝑃𝑃(𝐴𝐴∩𝐵𝐵) 88 𝑃𝑃(𝐵𝐵|𝐴𝐴) = . 𝑃𝑃(𝐴𝐴)

The definition is illustrated by the diagram for 𝐴𝐴 ∩ 𝐵𝐵.

Zooming in on the outcomes covered by 𝐴𝐴:

87 Id. 88 See id. at 46. 23 𝐵𝐵 occurs only in the doubly shaded region 𝐴𝐴 ∩ 𝐵𝐵. The area of that doubly shaded

region as a proportion of the area of 𝐴𝐴 represents the probability that 𝐵𝐵 occurs given

that 𝐴𝐴 occurs. That is captured in the definition of conditional probability:

𝑃𝑃(𝐴𝐴∩𝐵𝐵) 𝑃𝑃(𝐵𝐵|𝐴𝐴) = , where 𝑃𝑃(𝐴𝐴) > 0. 89 𝑃𝑃(𝐴𝐴)

This expression may be rewritten by multiplying both sides by 𝑃𝑃(𝐴𝐴). Doing

so yields the following rule: for any events 𝐴𝐴 and 𝐵𝐵 with 𝑃𝑃(𝐴𝐴) > 0, 𝑃𝑃(𝐴𝐴 ∩ 𝐵𝐵) =

𝑃𝑃(𝐴𝐴) ∙ 𝑃𝑃(𝐵𝐵|𝐴𝐴).90

iii. Definition Of Independent Events

This opinion also utilizes the definition of independent events. 𝐴𝐴 and 𝐵𝐵 are

independent if learning that 𝐴𝐴 “occurred gives us no information that would change

the likelihood of 𝐵𝐵 occurring (and vice versa).” 91 Events A and B are said to be

independent where 𝑃𝑃(𝐴𝐴 ∩ 𝐵𝐵) = 𝑃𝑃(𝐴𝐴) ∙ 𝑃𝑃(𝐵𝐵). 92

89 𝑃𝑃(𝐴𝐴) is restricted to values greater than zero to avoid a zero in the denominator. 90 Blitzstein & Hwang, supra note 12, at 52. 91 Id. at 63. 92 Id. 24 iv. The Law Of Total Probability

The final concept needed here is the “law of total probability.” The law of

total probability “relates conditional probability to unconditional probability,” 93

allowing us to use conditional probability “to decompose complicated probability

problems into simpler pieces.” 94

Let 𝐴𝐴1 , . . . , 𝐴𝐴𝑛𝑛 be a partition of the sample space 𝑆𝑆, as shown.

Then overlay event 𝐵𝐵.

93 Id. at 54. 94 Id. 25 The shaded region 𝐵𝐵 is equal to (𝐴𝐴1 ∩ 𝐵𝐵) + (𝐴𝐴2 ∩ 𝐵𝐵) + ⋯ + (𝐴𝐴𝑛𝑛 ∩ 𝐵𝐵) . This

illustrates that 𝑃𝑃(𝐵𝐵) = 𝑃𝑃(𝐴𝐴1 ∩ 𝐵𝐵) + 𝑃𝑃(𝐴𝐴2 ∩ 𝐵𝐵) + ⋯ + 𝑃𝑃(𝐴𝐴𝑛𝑛 ∩ 𝐵𝐵). We know that

𝑃𝑃(𝐴𝐴𝑖𝑖 ∩ 𝐵𝐵) = 𝑃𝑃(𝐴𝐴𝑖𝑖 ) ∙ 𝑃𝑃(𝐵𝐵|𝐴𝐴𝑖𝑖 ), where 𝐴𝐴𝑖𝑖 denotes a given component of the partition

𝐴𝐴1 , . . . , 𝐴𝐴𝑛𝑛 . 95 Thus, we may rewrite 𝑃𝑃(𝐵𝐵) as

𝑃𝑃(𝐵𝐵) = 𝑃𝑃(𝐴𝐴1 ) ∙ 𝑃𝑃(𝐵𝐵|𝐴𝐴1 ) + 𝑃𝑃(𝐴𝐴2 ) ∙ 𝑃𝑃(𝐵𝐵|𝐴𝐴2 ) + ⋯ + 𝑃𝑃(𝐴𝐴𝑛𝑛 ) ∙ 𝑃𝑃(𝐵𝐵|𝐴𝐴𝑛𝑛 ).

The following chart summarizes the relevant expressions above.

Relevant Expressions Total of one 𝑃𝑃(𝐴𝐴𝑐𝑐 ) = 1 − 𝑃𝑃(𝐴𝐴). Either A or B 𝑃𝑃(𝐴𝐴 ∪ 𝐵𝐵) = 𝑃𝑃(𝐴𝐴) + 𝑃𝑃(𝐵𝐵) − 𝑃𝑃(𝐴𝐴 ∩ 𝐵𝐵). B as subset of A If 𝐵𝐵 ⊆ 𝐴𝐴, then 𝑃𝑃(𝐴𝐴 ∩ 𝐵𝐵) = 𝑃𝑃(𝐵𝐵). Definition of 𝑃𝑃(𝐴𝐴∩𝐵𝐵) conditional If 𝐴𝐴 and 𝐵𝐵 are events with 𝑃𝑃(𝐴𝐴) > 0, 𝑃𝑃(𝐵𝐵|𝐴𝐴) = . 𝑃𝑃(𝐴𝐴) probability Both A and B (regardless of For any events 𝐴𝐴 and 𝐵𝐵 with 𝑃𝑃(𝐴𝐴) > 0, 𝑃𝑃(𝐴𝐴 ∩ 𝐵𝐵) = 𝑃𝑃(𝐴𝐴) ∙ 𝑃𝑃(𝐵𝐵|𝐴𝐴). independence) Definition of independent 𝐴𝐴 and 𝐵𝐵 are independent if and only if 𝑃𝑃(𝐴𝐴 ∩ 𝐵𝐵) = 𝑃𝑃(𝐴𝐴) ∙ 𝑃𝑃(𝐵𝐵). events

Let 𝐴𝐴1 , . . . , 𝐴𝐴𝑛𝑛 be a partition of the sample space 𝑆𝑆. Then, for any Law of total event 𝐵𝐵, 𝑃𝑃(𝐵𝐵) = 𝑃𝑃(𝐴𝐴1 ) ∙ 𝑃𝑃(𝐵𝐵|𝐴𝐴1 ) + 𝑃𝑃(𝐴𝐴2 ) ∙ 𝑃𝑃(𝐵𝐵|𝐴𝐴2 ) + ⋯ + probability 𝑃𝑃(𝐴𝐴𝑛𝑛 ) ∙ 𝑃𝑃(𝐵𝐵|𝐴𝐴𝑛𝑛 ).

b. Milestone Probabilities Using Pre-HV-108 PTRS Data

Alexion’s PTRS estimates do not themselves offer Milestone Probabilities.

To determine whether Alexion’s pre-HV-108 estimates of ALXN1830’s success

95 For any events 𝐴𝐴 and 𝐵𝐵 with 𝑃𝑃(𝐴𝐴) > 0, 𝑃𝑃(𝐴𝐴 ∩ 𝐵𝐵) = 𝑃𝑃(𝐴𝐴) ∙ 𝑃𝑃(𝐵𝐵|𝐴𝐴). 26 demonstrate an expected value for each milestone, that raw data must be translated

first into the milestone concepts, then into Milestone Probabilities.

I take these calculations in stages: first 𝑀𝑀2 through 𝑀𝑀5 , then 𝑀𝑀6 and 𝑀𝑀7 , and

finally 𝑀𝑀8 .

i. 𝑷𝑷(𝑴𝑴𝟐𝟐 ) Through 𝑷𝑷(𝑴𝑴𝟓𝟓 )

The PTRS data supplies the probability of a successful Phase 2 trial.

Milestones 2 and 3 call for dosing in a PCT. A Phase 3 trial qualifies as a PCT.96

The parties characterize a successful Phase 2 trial as one that leads to dosing in a

Phase 3 trial. 97 So for a given indication, the probability of a successful Phase 2 trial

is equivalent to the probability that a patient will be dosed in a PCT, mapping onto

𝑀𝑀2 and 𝑀𝑀3 . The overall PTRS supplies the probability of FDA approval. 98 This

maps onto 𝑀𝑀4 and 𝑀𝑀5 . 99

Based on these observations, Alexion’s PTRS before receipt of the HV-108

data yields the following probabilities for TED and cAMR100:

96 See Kinch Tr. 246–47; Merger Agr. §§ 1.1, 3.8(a)(ii)–(iii). 97 See SRS Op. Suppl. Br. 17 (citing JX 1863 at 53); ALXN Ans. Suppl. Br. 23. 98 See Borboroglu Tr. 1403; Washburn Tr. 645; Kinch Tr. 353–54, 413; JX 2498 at 73; JX 690 at 276. 99 Merger Agr. § 3.8(a)(iv)–(v). 100 JX 1863 at 53; JX 2608 at 4; JX 2225 at 5. Alexion was not pursuing other indications at the time of breach. See JX 1928; JX 1933 at 1; JX 2021 at 1; JX 2042 at 1; JX 2065 at 1; Lee Tr. 445, 457; see also JX 1948; Russell Tr. 732. 27 Pre-HV-108 Data TED cAMR Probability of a successful Phase 2 Trial 0.43 0.50 (Probability of a first dosing in a PCT) Overall PTRS 0.30 0.34 (Probability of FDA approval)

These data account for the dependencies among different stages of clinical

development and regulatory approval.101

Calculating 𝑃𝑃(𝑀𝑀2 ) through 𝑃𝑃(𝑀𝑀5 ) requires breaking the Milestone Events

down into probability inquiries. Let 𝑃𝑃𝑃𝑃𝑃𝑃TED and 𝐹𝐹𝐹𝐹𝐹𝐹TED be the events that TED

reaches a PCT and receives FDA approval, respectively. Let 𝑃𝑃𝑃𝑃𝑃𝑃cAMR and

𝐹𝐹𝐹𝐹𝐹𝐹cAMR be the events that cAMR reaches a PCT and receives FDA approval,

respectively.

Probability Inquiries for 𝑴𝑴𝟐𝟐 through 𝑴𝑴𝟓𝟓 Event Inquiry Notation What is the probability that either TED or cAMR 𝑀𝑀2 𝑃𝑃(𝑃𝑃𝑃𝑃𝑃𝑃TED ∪ 𝑃𝑃𝑃𝑃𝑃𝑃cAMR ) achieves a first dosing in a PCT? What is the probability that both TED and cAMR 𝑀𝑀3 𝑃𝑃(𝑃𝑃𝑃𝑃𝑃𝑃TED ∩ 𝑃𝑃𝑃𝑃𝑃𝑃cAMR ) achieve a first dosing in a PCT? What is the probability that either TED or cAMR 𝑀𝑀4 𝑃𝑃(𝐹𝐹𝐹𝐹𝐹𝐹 𝑇𝑇𝑇𝑇𝑇𝑇 ∪ 𝐹𝐹𝐹𝐹𝐹𝐹cAMR ) obtains FDA approval? What is the probability that both TED and cAMR 𝑀𝑀5 𝑃𝑃(𝐹𝐹𝐹𝐹𝐹𝐹TED ∩ 𝐹𝐹𝐹𝐹𝐹𝐹cAMR ) obtain FDA approval?

I treat the TED and cAMR programs as entirely independent, meaning the

success of one indication does not affect the probability of success in the other

101 See supra Section I.C; Borboroglu Tr. 1403; JX 1863 at 53; JX 2608 at 4. 28 indication. Kinch’s testimony supports this approach. He indicated that while proof

of concept in one indication could increase another indication’s probability of

success, a “more conservative approach” of treating the indications as independent

was warranted.102 I agree the conservative approach is appropriate. I proceed under

the assumption that 𝑃𝑃𝑃𝑃𝑃𝑃TED and 𝑃𝑃𝑃𝑃𝑃𝑃cAMR are independent events and that 𝐹𝐹𝐹𝐹𝐹𝐹TED

and 𝐹𝐹𝐹𝐹𝐹𝐹cAMR are independent events.

The table below summarizes the calculations for 𝑃𝑃(𝑀𝑀2 ) through 𝑃𝑃(𝑀𝑀5 ).

Pre-HV-108 PTRS Milestone Probability Calculations for 𝑴𝑴𝟐𝟐 Through 𝑴𝑴𝟓𝟓 Probability Inquiry Calculation 𝑃𝑃(𝑀𝑀2 ) 𝑃𝑃(𝑃𝑃𝑃𝑃𝑃𝑃TED ∪ 𝑃𝑃𝑃𝑃𝑃𝑃cAMR ) 0.43 + 0.5 − (0.43 × 0.5) = 𝟎𝟎. 𝟕𝟕𝟕𝟕𝟕𝟕. 𝑃𝑃(𝑀𝑀3 ) 𝑃𝑃(𝑃𝑃𝑃𝑃𝑃𝑃TED ∩ 𝑃𝑃𝑃𝑃𝑃𝑃cAMR ) 0.43 × 0.5 = 𝟎𝟎. 𝟐𝟐𝟐𝟐𝟐𝟐. 𝑃𝑃(𝑀𝑀4 ) 𝑃𝑃(𝐹𝐹𝐹𝐹𝐹𝐹 𝑇𝑇𝑇𝑇𝑇𝑇 ∪ 𝐹𝐹𝐹𝐹𝐹𝐹cAMR ) 0.3 + 0.34 − (0.3 × 0.34) = 𝟎𝟎. 𝟓𝟓𝟓𝟓𝟓𝟓. 𝑃𝑃(𝑀𝑀5 ) 𝑃𝑃(𝐹𝐹𝐹𝐹𝐹𝐹TED ∩ 𝐹𝐹𝐹𝐹𝐹𝐹cAMR ) 0.3 × 0.34 = 𝟎𝟎. 𝟏𝟏𝟏𝟏𝟏𝟏.

ii. 𝑷𝑷(𝑴𝑴𝟔𝟔 ) and 𝑷𝑷(𝑴𝑴𝟕𝟕 )

Milestones 6 and 7 depend on both EMA approval and country-specific

approvals.103 The probabilities of FDA and EMA approval are assumed to be equal

because of the near-universal overlap in approval between the two regulatory

entities. 104 Under that assumption, if one of the entities grants approval, the

probability that the other entity will grant approval is 1. Therefore, an indication’s

102 Kinch Tr. 349. 103 Merger Agr. §§ 3.8(a)(vi)–(vii). 104 Kinch Tr. 353–54; see also JX 2962 at 23; JX 1071 at 5. 29 overall PTRS can be thought of not only as its probability of obtaining FDA

approval, but also as its probability of obtaining EMA approval.

As for the country-specific approvals, Russell testified these approvals were

likely once an indication received EMA approval. 105 SRS equated “likely” with a

50.1% chance; I do the same. 106 Thus, for a given indication, the probability of

achieving both EMA approval and the requisite country-specific approvals may be

calculated by multiplying overall PTRS by 0.501. Applied to Alexion’s pre-HV-

108 overall PTRS figures for TED (probability of 0.3) and cAMR (probability of

0.34), this yields the following 107:

Pre-HV-108 Deduced Probabilities (for Milestones 6 and 7) TED cAMR Probability of EMA and country-specific approvals 0.1503 0.17034

As before, applying these indication-specific probabilities to the milestones requires

breaking the Milestone Events down into probability inquiries. Let 𝐸𝐸𝐸𝐸 𝑇𝑇𝑇𝑇𝑇𝑇 be the

event that TED receives the requisite European approvals. Let 𝐸𝐸𝐸𝐸𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 be the event

that cAMR receives the requisite European approvals.

105 Russell Tr. 755–56. 106 SRS Op. Suppl. Br. 27–28. 107 0.501 × 0.3 = 0.1503; 0.501 × 0.34 = 0.17034. 30 Probability Inquiries for 𝑴𝑴𝟔𝟔 and 𝑴𝑴𝟕𝟕 Event Inquiry Notation What is the probability that either TED or cAMR 𝑀𝑀6 𝑃𝑃(𝐸𝐸𝐸𝐸 𝑇𝑇𝑇𝑇𝑇𝑇 ∪ 𝐸𝐸𝐸𝐸𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 ) obtains the requisite European approvals? What is the probability that both TED and cAMR 𝑀𝑀7 𝑃𝑃(𝐸𝐸𝐸𝐸 𝑇𝑇𝑇𝑇𝑇𝑇 ∩ 𝐸𝐸𝐸𝐸𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 ) obtain the requisite European approvals?

The table below summarizes the calculations for 𝑃𝑃(𝑀𝑀6 ) and 𝑃𝑃(𝑀𝑀7 ).

Pre-HV-108 PTRS Milestone Probability Calculations for 𝑴𝑴𝟔𝟔 And 𝑴𝑴𝟕𝟕 Probability Inquiry Calculation 𝑃𝑃(𝑀𝑀6 ) 𝑃𝑃(𝐸𝐸𝐸𝐸 𝑇𝑇𝑇𝑇𝑇𝑇 ∪ 𝐸𝐸𝐸𝐸𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 ) 0.1503 + 0.17034 − 0.1503 × 0.17034 = 𝟎𝟎. 𝟐𝟐𝟐𝟐𝟐𝟐. 𝑃𝑃(𝑀𝑀7 ) 𝑃𝑃(𝐸𝐸𝐸𝐸 𝑇𝑇𝑇𝑇𝑇𝑇 ∩ 𝐸𝐸𝐸𝐸𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 ) 0.1503 × 0.17034 = 𝟎𝟎. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎.

iii. 𝑷𝑷(𝑴𝑴𝟖𝟖 )

Milestone 8, based on $1 billion in net sales in a year across all indications,

requires a different approach. When ATP performed its 2018 valuation, it needed a

Monte Carlo simulation to value this milestone because at the time, Alexion was

pursuing three indications, and there were various combinations of success that

could realistically lead to $1 billion in net sales in a single year. 108 That approach is

not needed here. At the time of breach, Alexion was only pursuing treatments for

TED and cAMR, and SRS has not demonstrated that either could have reached $1

108 See JX 2962 at 24–25. 31 billion in net sales on its own.109 Alexion could have achieved Milestone 8 only if

it obtained regulatory approval for both TED and cAMR.

Russell testified that ALXN1830 had the potential to achieve the milestone

based on Alexion’s internal revenue projections for TED and cAMR.110 Alexion’s

2021 model indicates it expected the indications’ combined revenues to surpass $1

billion in five years, including a $1.35 billion peak. 111 Put differently, Alexion’s

peak projection exceeds the $1 billion threshold by 35%. By that margin, the

preponderance of the evidence shows it is likely ALXN1830 would have crossed

that threshold as long as both TED and cAMR received FDA and EMA approval

(even if the indications did not receive all three of the country-specific approvals

needed to satisfy Milestones 6 and 7).

Because EMA approval is assumed to be guaranteed upon FDA approval, the

probability that both indications would have obtained FDA and EMA approval is

equal to the probability that both indications would have obtained FDA approval—

i.e., the probability of 𝑀𝑀5 . From there, 𝑀𝑀8 is likely. So, 𝑀𝑀8 ’s probability is

calculated by multiplying 𝑀𝑀5 ’s probability of 0.102 by 0.501.

See JX 1953; see also JX 1928; JX 1933 at 1; JX 2021 at 1; JX 2042 at 1; JX 2065 at 1; 109

Lee Tr. 445, 457; JX 1948; Russell Tr. 732. 110 Russell Tr. 744–45; see JX 1953. 111 JX 1953. 32 The pre-HV-108 PTRS probability for 𝑀𝑀8 is 0.0511.112

c. The Post-HV-108 PTRS Data

The record offers another set of PTRS numbers: those Alexion adjusted after

receiving the HV-108 data, just before terminating the ALXN1830 program.

Alexion reduced TED’s chances of a successful Phase 2 trial from 43% to 20%, and

it reduced TED’s overall PTRS from 30% to 10%. 113 Alexion also reduced cAMR’s

overall PTRS from 34% to 10%. 114 Before the reduction, cAMR’s probability of a

successful Phase 2 trial was 50%. 115 But the record does not indicate how much that

probability was reduced in response to the HV-108 data.

The post-HV-108 PTRS data paints the following incomplete picture:

Post-HV-108 Raw Data TED cAMR Probability of a successful Phase 2 Trial 0.2 ? Overall PTRS 0.1 0.1

Applying the same inferences and observations as in the previous section

yields the following:

112 0.102 × 0.501 = 0.0511. 113 JX 2608 at 4. 114 JX 2225 at 13–14. 115 JX 1863 at 53. 33 Post-HV-108 Deduced Probabilities TED cAMR Probability of a first dosing in a PCT 0.2 ? Probability of FDA approval 0.1 0.1 Probability of EMA and country-specific approvals 0.0501 116 0.0501 117

Without the missing cAMR information, 𝑃𝑃(𝑀𝑀2 ) and 𝑃𝑃(𝑀𝑀3 ) cannot be

precisely calculated. But because cAMR has a nonzero probability of FDA

approval, it must also have a nonzero probability of clinical success, including

achieving dosing in a PCT. 𝑃𝑃(𝑀𝑀2 ) and 𝑃𝑃(𝑀𝑀3 ) would still be greater than zero even

under the decreased post-HV-108 PTRS estimates.

𝑃𝑃(𝑀𝑀4 ) through 𝑃𝑃(𝑀𝑀7 ) can be found using the same calculations performed

on the pre-HV-108 data.

Post-HV-108 PTRS Milestone Probability Calculations for 𝑴𝑴𝟒𝟒 Through 𝑴𝑴𝟕𝟕 Probability Inquiry Calculation 𝑃𝑃(𝑀𝑀4 ) 𝑃𝑃(𝐹𝐹𝐹𝐹𝐹𝐹 𝑇𝑇𝑇𝑇𝑇𝑇 ∪ 𝐹𝐹𝐹𝐹𝐹𝐹cAMR ) 0.1 + 0.1 − (0.1 × 0.1) = 𝟎𝟎. 𝟏𝟏𝟏𝟏. 𝑃𝑃(𝑀𝑀5 ) 𝑃𝑃(𝐹𝐹𝐹𝐹𝐹𝐹TED ∩ 𝐹𝐹𝐹𝐹𝐹𝐹cAMR ) 0.1 × 0.1 = 𝟎𝟎. 𝟎𝟎𝟎𝟎. 𝑃𝑃(𝑀𝑀6 ) 𝑃𝑃(𝐸𝐸𝐸𝐸 𝑇𝑇𝑇𝑇𝑇𝑇 ∪ 𝐸𝐸𝐸𝐸𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 ) 0.0501 + 0.0501 − 0.0501 × 0.0501 = 𝟎𝟎. 𝟎𝟎𝟎𝟎𝟎𝟎. 𝑃𝑃(𝑀𝑀7 ) 𝑃𝑃(𝐸𝐸𝐸𝐸 𝑇𝑇𝑇𝑇𝑇𝑇 ∩ 𝐸𝐸𝐸𝐸𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 ) 0.0501 × 0.501 = 𝟎𝟎. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎.

Finally, as in the pre-HV-108 PTRS, 𝑃𝑃(𝑀𝑀8 ) is calculated by multiplying

𝑃𝑃(𝑀𝑀5 ) by 0.501. So the post-HV-108 PTRS probability for 𝑀𝑀8 is about 0.00501.118

116 0.1 × 0.501 = 0.0501. 117 0.1 × 0.501 = 0.0501. 118 0.01 × 0.501 = 0.00501. 34 d. Milestone Probabilities Using CDEK Data

Kinch’s opinions based on his CDEK database provide an alternative starting

point for estimating the Milestone Probabilities. Kinch’s Milestone Probability

calculations do not properly account for the fact that some milestones are contingent

upon others. 119 They must be adjusted to comport with Kinch’s starting

assumptions, reproduced in the following table.

Kinch’s Opinions Based on CDEK Opinion Notation The probability of 𝑀𝑀2 is between 0.582 and 0.582 ≤ 𝑃𝑃(𝑀𝑀2 ) ≤ 1. 1. The probability of 𝑀𝑀4 given that 𝑀𝑀2 𝑃𝑃(𝑀𝑀4 |𝑀𝑀2 ) = 0.684. occurred is 0.684. The probability of 𝑀𝑀3 given that 𝑀𝑀2 𝑃𝑃(𝑀𝑀3 |𝑀𝑀2 ) = 𝑃𝑃(𝑀𝑀2 ). occurred is equal to the probability of 𝑀𝑀2 . The probability of 𝑀𝑀5 given that 𝑀𝑀4 𝑃𝑃(𝑀𝑀5 |𝑀𝑀4 ) = 𝑃𝑃(𝑀𝑀4 ). occurred is equal to the probability of 𝑀𝑀4 .

119 For example, Kinch’s expert report conflates 𝑀𝑀3 ’s conditional probability given 𝑀𝑀2 with 𝑀𝑀3 ’s probability at the outset (i.e., without knowing whether 𝑀𝑀2 occurred). Translated to notation, the report conflates 𝑃𝑃(𝑀𝑀3 |𝑀𝑀2 ) with 𝑃𝑃(𝑀𝑀3 ). As a result, Kinch’s expert report mistakenly suggests 𝑃𝑃(𝑀𝑀3 ) = 𝑃𝑃(𝑀𝑀2 ). See JX 2498 at 68–72, 76. As this section shows, 𝑃𝑃(𝑀𝑀3 ) = 𝑃𝑃(𝑀𝑀2 ) ∙ 𝑃𝑃(𝑀𝑀3 |𝑀𝑀2 ). If 𝑃𝑃(𝑀𝑀3 ) were equal to 𝑃𝑃(𝑀𝑀2 ), that would imply 𝑃𝑃(𝑀𝑀3 |𝑀𝑀2 ) = 1 . In other words, it would imply 𝑀𝑀3 was guaranteed upon the occurrence of 𝑀𝑀2 . That is inconsistent with reality and Kinch’s own testimony. See Kinch Tr. 349 (responding when asked about his calculations for 𝑃𝑃(𝑀𝑀3 ), “You can argue . . . that if you’ve established proof of concept and proof of mechanism, other mechanisms should be more likely . . . but I took the more conservative approach.” (emphasis added)). Kinch’s report also conflates 𝑃𝑃(𝑀𝑀5 |𝑀𝑀4 ) with 𝑃𝑃(𝑀𝑀4 ). See JX 2498 at 68–72, 76. As a result, the report suggests 𝑃𝑃(𝑀𝑀5 ) = 𝑃𝑃(𝑀𝑀4 ) . But as this section shows, 𝑃𝑃(𝑀𝑀5 ) = 𝑃𝑃(𝑀𝑀5 ) ∙ 𝑃𝑃(𝑀𝑀5 |𝑀𝑀4 ). So if 𝑃𝑃(𝑀𝑀5 ) were equal to 𝑃𝑃(𝑀𝑀4 ), that would imply 𝑃𝑃(𝑀𝑀5 |𝑀𝑀4 ) = 1. Again, this is false. My CDEK calculations disregard the errors in Kinch’s expert report. 35 i. 𝑷𝑷(𝑴𝑴𝟐𝟐 )

Kinch estimated a probability range for 𝑀𝑀2 based on his analysis of CDEK.

This figure needs no adjustment. The CDEK probability range for 𝑀𝑀2 is

0.582 ≤ 𝑃𝑃(𝑀𝑀2 ) ≤ 1.

ii. 𝑷𝑷(𝑴𝑴𝟑𝟑 )

ALXN1830 cannot reach a PCT for a second indication unless it has done so

for some first indication. So 𝑀𝑀3 ⊆ 𝑀𝑀2 . It follows that 𝑃𝑃(𝑀𝑀2 ∩ 𝑀𝑀3 ) = 𝑃𝑃(𝑀𝑀3 ).120

From our mathematical primer, we know 𝑃𝑃(𝑀𝑀2 ∩ 𝑀𝑀3 ) = 𝑃𝑃(𝑀𝑀2 ) ∙ 𝑃𝑃(𝑀𝑀3 |𝑀𝑀2 ) . 121

Therefore,

𝑃𝑃(𝑀𝑀3 ) = 𝑃𝑃(𝑀𝑀2 ) ∙ 𝑃𝑃(𝑀𝑀3 |𝑀𝑀2 ).

Recall Kinch opined that 𝑃𝑃(𝑀𝑀3 |𝑀𝑀2 ) = 𝑃𝑃(𝑀𝑀2 ). That substitution yields

𝑃𝑃(𝑀𝑀3 ) = 𝑃𝑃(𝑀𝑀2 )2 .

Plugging in the lower and upper bounds for 𝑃𝑃(𝑀𝑀2 ) yields

0.339 ≤ 𝑃𝑃(𝑀𝑀3 ) ≤ 1. 122

120 If 𝐵𝐵 ⊆ 𝐴𝐴, then 𝑃𝑃(𝐴𝐴 ∩ 𝐵𝐵) = 𝑃𝑃(𝐵𝐵). 121 For any events 𝐴𝐴 and 𝐵𝐵 with 𝑃𝑃(𝐴𝐴) > 0, 𝑃𝑃(𝐴𝐴 ∩ 𝐵𝐵) = 𝑃𝑃(𝐴𝐴) ∙ 𝑃𝑃(𝐵𝐵|𝐴𝐴). 122 0.5822 = 0.339; 12 = 1. 36 iii. 𝑷𝑷(𝑴𝑴𝟒𝟒 )

First, because it is impossible for any indication to receive FDA approval

unless some indication advances to a PCT, 𝑀𝑀4 ⊆ 𝑀𝑀2 . So 𝑃𝑃(𝑀𝑀2 ∩ 𝑀𝑀4 ) = 𝑃𝑃(𝑀𝑀4 ).123

And we know 𝑃𝑃(𝑀𝑀2 ∩ 𝑀𝑀4 ) = 𝑃𝑃(𝑀𝑀2 ) ∙ 𝑃𝑃(𝑀𝑀4 |𝑀𝑀2 ). 124 Therefore,

𝑃𝑃(𝑀𝑀4 ) = 𝑃𝑃(𝑀𝑀2 ) ∙ 𝑃𝑃(𝑀𝑀4 |𝑀𝑀2 ).

Applying Kinch’s opinion that 𝑃𝑃(𝑀𝑀4 |𝑀𝑀2 ) = 0.684 yields

𝑃𝑃(𝑀𝑀4 ) = 𝑃𝑃(𝑀𝑀2 ) ∙ 0.684.

Plugging in the lower and upper bounds for 𝑃𝑃(𝑀𝑀2 ) yields

0.398 ≤ 𝑃𝑃(𝑀𝑀4 ) ≤ 0.684. 125

iv. 𝑷𝑷(𝑴𝑴𝟓𝟓 )

Similarly, because ALXN1830 cannot receive FDA approval for a second

indication unless it has done so in some first indication, 𝑀𝑀5 ⊆ 𝑀𝑀4 . So

𝑃𝑃(𝑀𝑀4 ∩ 𝑀𝑀5 ) = 𝑃𝑃(𝑀𝑀5 ) . 126 Like before, we know 𝑃𝑃(𝑀𝑀4 ∩ 𝑀𝑀5 ) = 𝑃𝑃(𝑀𝑀4 ) ∙

𝑃𝑃(𝑀𝑀5 |𝑀𝑀4 ). 127 Therefore,

𝑃𝑃(𝑀𝑀5 ) = 𝑃𝑃(𝑀𝑀4 ) ∙ 𝑃𝑃(𝑀𝑀5 |𝑀𝑀4 ).

Applying Kinch’s opinion that 𝑃𝑃(𝑀𝑀5 |𝑀𝑀4 ) = 𝑃𝑃(𝑀𝑀4 ) yields

123 If 𝐵𝐵 ⊆ 𝐴𝐴, then 𝑃𝑃(𝐴𝐴 ∩ 𝐵𝐵) = 𝑃𝑃(𝐵𝐵). 124 For any events 𝐴𝐴 and 𝐵𝐵 with 𝑃𝑃(𝐴𝐴) > 0, 𝑃𝑃(𝐴𝐴 ∩ 𝐵𝐵) = 𝑃𝑃(𝐴𝐴) ∙ 𝑃𝑃(𝐵𝐵|𝐴𝐴). 125 0.582 × 0.684 = 0.398; 1 × 0.684 = 0.398. 126 If 𝐵𝐵 ⊆ 𝐴𝐴, then 𝑃𝑃(𝐴𝐴 ∩ 𝐵𝐵) = 𝑃𝑃(𝐵𝐵). 127 For any events 𝐴𝐴 and 𝐵𝐵 with 𝑃𝑃(𝐴𝐴) > 0, 𝑃𝑃(𝐴𝐴 ∩ 𝐵𝐵) = 𝑃𝑃(𝐴𝐴) ∙ 𝑃𝑃(𝐵𝐵|𝐴𝐴). 37 𝑃𝑃(𝑀𝑀5 ) = 𝑃𝑃(𝑀𝑀4 )2 .

Plugging in the above lower and upper bounds of 𝑃𝑃(𝑀𝑀4 ) to calculate the lower and

upper bounds of 𝑃𝑃(𝑀𝑀5 ) yields

0.158 ≤ 𝑃𝑃(𝑀𝑀5 ) ≤ 0.468. 128

v. 𝑷𝑷(𝑴𝑴𝟔𝟔 )

Kinch’s analysis of milestone probabilities was agnostic as to the number of

indications Alexion was pursuing. The CDEK data did not provide information on

the number of indications being pursued for a given molecule. 129 For purposes of

calculating 𝑃𝑃(𝑀𝑀6 ) based on CDEK, I assume Alexion was pursuing exactly two

indications. It is not possible to calculate 𝑃𝑃(𝑀𝑀6 ) with the available information.130

And Alexion was in fact pursuing two indications at the time of breach. And finally,

assuming Alexion was pursuing more than two indications would increase each

remaining Milestone Probability, so my calculations serve as a conservative floor

for 𝑃𝑃(𝑀𝑀6 ).

Given the assumption that Alexion was pursuing two indications, 𝑃𝑃(𝑀𝑀6 )

depends on whether ALXN1830 received FDA approval in zero, one, or two

indications. Let us define three events:

128 0.3982 = 0.158; 0.6842 = 0.468. 129 JX 2498 at 70; cf. Kinch Tr. 230–31. This is because the likelihood of 𝑀𝑀6 depends on the number of indications that receive 130

FDA approval. 38 𝐹𝐹𝐹𝐹𝐹𝐹0 : zero indications receive FDA approval;

𝐹𝐹𝐹𝐹𝐹𝐹1 : one indication receives FDA approval;

𝐹𝐹𝐹𝐹𝐹𝐹2 : two indications receive FDA approval.

Because these events partition the sample space 𝑆𝑆 , the law of total probability

provides that

𝑃𝑃(𝑀𝑀6 ) = 𝑃𝑃(𝐹𝐹𝐹𝐹𝐴𝐴0 ) ∙ 𝑃𝑃(𝑀𝑀6 |𝐹𝐹𝐹𝐹𝐴𝐴0 ) + 𝑃𝑃(𝐹𝐹𝐹𝐹𝐴𝐴1 ) ∙ 𝑃𝑃(𝑀𝑀6 |𝐹𝐹𝐹𝐹𝐴𝐴1 ) + 𝑃𝑃(𝐹𝐹𝐹𝐹𝐴𝐴2 ) ∙ 𝑃𝑃(𝑀𝑀6 |𝐹𝐹𝐹𝐹𝐴𝐴2 ).

We may use this expression, Kinch’s initial opinions, and previous calculations to

calculate lower and upper bounds for 𝑃𝑃(𝑀𝑀6 ).

As to the lower bound, the following table solves for each term on the right-

hand side of the expression for 𝑃𝑃(𝑀𝑀6 ) assuming 𝑃𝑃(𝑀𝑀2 ) = 0.582. The terms are

listed in an order that facilitates calculation.

Lower Bound of Each Term in the Expression for 𝑷𝑷(𝑴𝑴𝟔𝟔 ) Probability Explanation 𝑃𝑃(𝐹𝐹𝐹𝐹𝐴𝐴0 ) = 1 − 𝑃𝑃(𝑀𝑀4 ) , because 𝑀𝑀4 occurs when at least one indication receives FDA approval. 𝑃𝑃(𝐹𝐹𝐹𝐹𝐴𝐴0 ) = 0.602. 𝑃𝑃(𝐹𝐹𝐹𝐹𝐴𝐴0 ) = 1 − 𝑃𝑃(𝑀𝑀4 ) = 1 − 0.398 = 0.602.

𝑃𝑃(𝐹𝐹𝐹𝐹𝐴𝐴2 ) = 𝑃𝑃(𝑀𝑀5 ) because 𝑀𝑀5 occurs when two indications receive FDA approval. 𝑃𝑃(𝐹𝐹𝐹𝐹𝐴𝐴2 ) = 0.158. 𝑃𝑃(𝐹𝐹𝐹𝐹𝐴𝐴2 ) = 𝑃𝑃(𝑀𝑀5 ) = 0.158. 𝐹𝐹𝐹𝐹𝐴𝐴1 occurs if neither 𝐹𝐹𝐹𝐹𝐴𝐴0 nor 𝐹𝐹𝐹𝐹𝐴𝐴2 occur. So 𝑃𝑃(𝐹𝐹𝐹𝐹𝐴𝐴1 ) = 1 − [𝑃𝑃(𝐹𝐹𝐹𝐹𝐴𝐴0 ) + 𝑃𝑃(𝐹𝐹𝐹𝐹𝐴𝐴2 )]. 𝑃𝑃(𝐹𝐹𝐹𝐹𝐴𝐴1 ) = 0.240. 𝑃𝑃(𝐹𝐹𝐹𝐹𝐴𝐴1 ) = 1 − 0.602 − 0.158 = 0.240.

39 𝑃𝑃(𝑀𝑀6 |𝐹𝐹𝐹𝐹𝐴𝐴0 ) = 0. 𝑀𝑀6 is not possible if zero indications receive FDA approval.

Because it is assumed that FDA approval guarantees EMA approval, Russell’s testimony indicates that any indication that receives FDA approval has a 50.1% chance of receiving 𝑃𝑃(𝑀𝑀6 |𝐹𝐹𝐹𝐹𝐴𝐴1 ) = 0.501. the requisite European approvals. 131 So, if exactly one indication receives FDA approval, 𝑀𝑀6 occurs with probability 0.501. So, 𝑃𝑃(𝑀𝑀6 |𝐹𝐹𝐹𝐹𝐴𝐴1 ) = 0.501. If an indication receives FDA approval, its probability of not receiving the requisite country-specific approvals to satisfy 𝑀𝑀6 is (1 − 0.501) . So, if two indications receive FDA approval, the probability that neither receives the requisite 𝑃𝑃(𝑀𝑀6 |𝐹𝐹𝐹𝐹𝐴𝐴2 ) = 0.751. country-specific approvals is (1 − 0.501)2 . Thus, the probability that at least one of the indications receives the requisite country-specific approvals is (1 − (1 − 0.501)2 ).

𝑃𝑃(𝑀𝑀6 |𝐹𝐹𝐹𝐹𝐴𝐴2 ) = 1 − (1 − 0.501)2 = 0.751.

Substituting these values into the expression,

𝑃𝑃(𝑀𝑀6 ) = (0.602) ∙ (0) + (0.240) ∙ (0.501) + (0.158) ∙ (0.751) = 𝟎𝟎. 𝟐𝟐𝟐𝟐𝟐𝟐.

This calculation may also be visualized with a flow chart.

131 As in the PTRS calculations, I assume each indication is entirely independent of the other. 40 As to the upper bound of 𝑃𝑃(𝑀𝑀6 ), the following (abbreviated) table solves for

each term on the right-hand side of the expression for 𝑃𝑃(𝑀𝑀6 ) assuming 𝑃𝑃(𝑀𝑀2 ) = 1.

The terms are again listed in an order that facilitates calculation.

Upper Bound of Each Term in the Expression for 𝑷𝑷(𝑴𝑴𝟔𝟔 ) Probability Explanation 𝑃𝑃(𝐹𝐹𝐹𝐹𝐴𝐴0 ) = 0.316. 𝑃𝑃(𝐹𝐹𝐹𝐹𝐴𝐴0 ) = 1 − 𝑃𝑃(𝑀𝑀4 ) = 1 − 0.684 = 0.316. 𝑃𝑃(𝐹𝐹𝐹𝐹𝐴𝐴2 ) = 0.468. 𝑃𝑃(𝐹𝐹𝐹𝐹𝐴𝐴2 ) = 𝑃𝑃(𝑀𝑀5 ) = 0.468. 𝑃𝑃(𝐹𝐹𝐹𝐹𝐴𝐴1 ) = 0.216. 𝑃𝑃(𝐹𝐹𝐹𝐹𝐴𝐴1 ) = 1 − 0.316 − 0.468 = 0.216. 𝑃𝑃(𝑀𝑀6 |𝐹𝐹𝐹𝐹𝐴𝐴0 ) = 0. The explanation is the same as above. 𝑃𝑃(𝑀𝑀6 |𝐹𝐹𝐹𝐹𝐴𝐴1 ) = 0.501. The explanation is the same as above. 𝑃𝑃(𝑀𝑀6 |𝐹𝐹𝐹𝐹𝐴𝐴2 ) = 0.751. The explanation is the same as above.

41 Substituting these values into the expression for 𝑃𝑃(𝑀𝑀6 ),

𝑃𝑃(𝑀𝑀6 ) = (0.316) ∙ (0) + (0.216) ∙ (0.501) + (0.468) ∙ (0.751) = 𝟎𝟎. 𝟒𝟒𝟒𝟒𝟒𝟒.

0.239 ≤ 𝑃𝑃(𝑀𝑀6 ) ≤ 0.460.

vi. 𝑷𝑷(𝑴𝑴𝟕𝟕 )

The record shows that the requisite European approvals follow only after FDA

approval. So 𝑀𝑀7 ⊆ 𝑀𝑀5 . Thus, 𝑃𝑃(𝑀𝑀5 ∩ 𝑀𝑀7 ) = 𝑃𝑃(𝑀𝑀7 ). Additionally, we know

𝑃𝑃(𝑀𝑀5 ∩ 𝑀𝑀7 ) = 𝑃𝑃(𝑀𝑀5 ) ∙ 𝑃𝑃(𝑀𝑀7 |𝑀𝑀5 ). Therefore,

𝑃𝑃(𝑀𝑀7 ) = 𝑃𝑃(𝑀𝑀5 ) ∙ 𝑃𝑃(𝑀𝑀7 |𝑀𝑀5 ).

Based on Russell’s testimony, (𝑀𝑀7 |𝑀𝑀5 ) = (0.501)2 . 132 So

𝑃𝑃(𝑀𝑀7 ) = 𝑃𝑃(𝑀𝑀5 ) ∙ (0.501)2 .

Plugging in the lower and upper bounds for 𝑃𝑃(𝑀𝑀5 ) yields:

0.0398 ≤ 𝑃𝑃(𝑀𝑀7 ) ≤ 0.117. 133

132 Based on Russell’s testimony, an indication that receives FDA approval has a 0.501 probability of receiving the European approvals. If 𝑀𝑀5 is given, then both indications have received FDA approval. 𝑀𝑀7 occurs only if both of those indications receive the European approvals. Because each indication is assumed to be independent, we can calculate the probability that both obtain EMA approval given that both obtained FDA approval by multiplying each indication’s probability of making that leap on its own. 133 0.158 × (0.501)2 = 0.0398; 0.468 × (0.501)2 = 0.117. 42 vii. 𝑷𝑷(𝑴𝑴𝟖𝟖 )

Milestone 8’s CDEK probability can be calculated in the same manner used

for the PTRS calculations. 134

𝑃𝑃(𝑀𝑀8 ) = 𝑃𝑃(𝑀𝑀5 ) ∙ (0.501).

Using the lower and upper bounds calculated for 𝑃𝑃(𝑀𝑀5 ),

0.0794 ≤ 𝑃𝑃(𝑀𝑀8 ) ≤ 0.234. 135

In sum, the CDEK probabilities are as follows:

CDEK Probabilities Milestone Event Probability 𝑀𝑀2 0.582–1 𝑀𝑀3 0.339–1 𝑀𝑀4 0.398–0.684 𝑀𝑀5 0.158–0.468 𝑀𝑀6 0.239–0.460 𝑀𝑀7 0.0398–0.117 𝑀𝑀8 0.0794–0.234

134 For the PTRS calculations, I reasoned—based on Alexion’s internal projections for TED and cAMR and Russell’s testimony about those projections—that 𝑀𝑀8 was likely to occur if ALXN1830 obtained FDA approval in both TED and cAMR, but that 𝑀𝑀8 otherwise would not occur. To calculate 𝑀𝑀6 ’s CDEK probability, I assumed Kinch’s opinions contemplated two indications. It was not necessary to assume which specific indications were at play. Given that Alexion was in fact pursuing TED and cAMR, I believe it is fair to calculate 𝑀𝑀8 ’s CDEK probability in a manner that mirrors that reality. My 𝑀𝑀8 CDEK calculations thus assume 𝑀𝑀8 was likely to occur if ALXN1830 received FDA approval in the two indications assumed to be at play, but that 𝑀𝑀8 otherwise would not occur. 135 0.158 × 0.501 = 0.079; 0.468 × 0.501 = 0.234. 43 e. Comparison Of Milestone Probability Estimates

I have summarized the probability estimates under the three complete views

from the record evidence, plus the incomplete view based on the post-HV-108 PTRS

data.

Comparison of Milestone Probability Estimates Event ATP Pre-HV-108 PTRS Post-HV-108 PTRS CDEK 𝑀𝑀2 0.80 0.715 >0 0.582–1 𝑀𝑀3 0.75 0.215 >0 0.339–1 𝑀𝑀4 0.53 0.538 0.190 0.398–0.684 𝑀𝑀5 0.45 0.102 0.0100 0.158–0.468 𝑀𝑀6 0.53 0.295 0.0977 0.239–0.460 𝑀𝑀7 0.45 0.0256 0.00251 0.0398–0.117 𝑀𝑀8 0.10 0.0511 0.00501 0.0794–0.234

No dataset indicates the probability of achieving any given milestone was

zero. SRS has proven by a preponderance of the evidence that the probability of

achieving each milestone—and therefore the expected value of each milestone—

was greater than zero at the time of breach. SRS has proven an injury from Alexion’s

breach, in general and as tethered to each milestone.

44 2. Causation

Expectation damages are certainly available when the plaintiff proves the

defendant’s breach was the but-for cause of its injury.136 Here, Alexion’s breach

was the termination of the ALXN1830 program, which eliminated all expected value

in the milestones at the time of breach. That termination was the but-for cause of

SRS’s loss of expected value: there was no time or opportunity for any intervening

cause to contribute to that loss.

Alexion argues SRS must prove ALXN1830 was more likely than not to

achieve a given milestone to establish proximate cause. 137 But Alexion

misunderstands SRS’s injuries as lost milestone payments in toto. As explained, the

injury is the loss of expected value associated with milestone payments triggered by

progress Alexion makes using CREs. SRS must still show causation: it must prove

136 See SIGA Techs., Inc. v. PharmAthene, Inc., 67 A.3d 330, 350–51 (Del. 2013) (“We now hold that where the parties have a Type II preliminary agreement to negotiate in good faith, and the trial judge makes a factual finding, supported by the record, that the parties would have reached an agreement but for the defendant’s bad faith negotiations, the plaintiff is entitled to recover contract expectation damages.”). Expectation damages may also be available when the breach proximately caused the damage. See Great Hill Equity P’rs IV, LLP v. SIG Growth Equity Fund I, LLP, 2020 WL 948513, at *19 (Del. Ch. Feb. 27, 2020). 137 D.I. 313 at 58; D.I. 371 at 53. 45 Alexion’s failure to use CREs caused a loss in expected value with reasonable

certainty. 138 It has done so.

3. Estimate Of Damages

“The amount of damages can be an estimate,” 139 provided “the court has a

basis to make a responsible estimate of damages.”140 Delaware law grants the Court

flexibility to use its “conscience and reason” to determine such an estimate and to

determine whether setting an estimate is appropriate under the circumstances.141 In

establishing that estimate, Delaware courts recognize the wrongdoer rule, which

states,

Doubts [about the extent of damages] are generally resolved against the party in breach. A party who has, by his breach, forced the injured party to seek compensation in damages should not be allowed to profit from his breach where it is established that a significant loss has occurred. A court may take into account all the circumstances of the breach, including willfulness, in deciding whether to require a lesser degree of certainty, giving greater discretion to the trier of the facts. Damages

138 Siga II, 132 A.3d at 1130; cf. Del. Express Shuttle, Inc. v. Older, 2002 WL 31458243, at *15 (Del. Ch. Oct. 23, 2002). 139 Siga II, 132 A.3d at 1111. 140 Del. Express, 2002 WL 31458243, at *15. 141 See Siga II, 132 A.3d at 1130 (quoting Gatz Props., LLC v. Auriga Cap. Corp., 59 A.3d 1206, 1212 (Del. 2012)); Weinberger v. UOP, Inc., 457 A.2d 701, 715 (Del. 1983) (noting the Court of Chancery’s “broad discretion . . . to fashion such relief as the facts of a given case may dictate”). 46 need not be calculable with mathematical accuracy and are often at best approximate. 142

In Fortis, “the parties’ contemporaneous risk-adjusted probabilities of

success” were the best evidence of the milestones’ expected payouts “absent [the

promisor’s] breaches.” 143 Fortis was able to rely on the parties’ own projections for

several reasons. First, little time had passed between the date of those projections

and the time of the promisor’s breaches.144 Second, because the promisor “had deep

knowledge of its own ability to reach the milestones,” and the promisee’s

“independent estimate came after (or during) multiple rounds of due diligence and

was remarkably close to [the promisee’s] predictions,” the probabilities provided “a

credible, responsible basis to calculate” damages.145 And third, the plaintiff’s more

optimistic view of the probabilities was balanced out by the defendant’s more

conservative view.146

142 Cura, 2001 WL 1334188, at *20 (quoting Restatement (Second) of Contracts § 352 cmt. a (1981)); accord Siga I, 2014 WL 3974167, at *8; Beard Rsch., Inc. v. Kates, 8 A.3d 573, 613 (Del. Ch. 2010) (“Public policy has led Delaware courts to show a general willingness to make a wrongdoer ‘bear the risk of uncertainty of a damages calculation where the calculation cannot be mathematically proven.’” (quoting Great Am. Opportunities, Inc. v. Cherrydale Fundraising, LLC, 2010 WL 338219, at *23 (Del. Ch. Jan. 29, 2010)), aff’d sub. nom., ASDI, Inc. v. Beard Rsch., Inc., 11 A.3d 749 (Del. 2010). 143 Fortis, 2024 WL 4048060, at *51. 144 See id. at *26–34, *50–51. 145 Id. at *51. 146 Id. at *52. 47 Similarly, here, the strongest evidence of the Milestone Probabilities is

derived from Alexion’s own pre-HV-108 PTRS estimates made shortly before the

breach. ATP’s November 2018 estimates are not a good indication of SRS’s

expectation damages at the time of breach because those estimates were not

contemporaneous with the breach, and reflect very different circumstances. Most

notably, in 2018, Alexion was pursuing treatments for PV, gMG, and WAIHA; by

the fall of 2021, Alexion’s was solely focused on TED and cAMR.147

The CDEK estimates (based on Kinch’s and Russell’s testimony) are

anchored to the time of breach. But I am skeptical of the reliability of Kinch’s CDEK

database on which those estimates are based. First, because the database only

includes publicly available data, it systematically overestimates probabilities of

success; Kinch acknowledged that failures are underreported. 148 Unlike the private

databases CDEK intends to mimic, CDEK is not a tested tool.149 Pharmaceutical

companies do not use it, and Kinch implied that the only peer-reviewed papers

relying on the database were his own.150

See JX 1477; JX 1928; JX 1933 at 1; JX 2021 at 1; JX 2042 at 1; JX 2065 at 1; Lee Tr. 147

445, 457; see also JX 1948; Russell Tr. 732. 148 Kinch Tr. 377–79. 149 See id. at 208. 150 See id. 48 Further, this opinion has explained the deficiencies in Kinch’s probability

calculations. While I recalculated the CDEK probabilities based on Kinch’s initial

opinions from his database, Kinch’s mathematical errors undermine his overall

credibility.

As for the pre-HV-108 PTRS estimates, those are Alexion’s own internal

predictors of clinical and regulatory success, set by those with detailed knowledge

of the ALXN1830 program just before the breach, and untainted by a desire to

terminate the program. Alexion’s management team was capable of estimating its

own molecule’s probability of success accurately enough to serve as a starting point

for a responsible estimate of damages.

To be sure, the pre-HV-108 PTRS figures do not appear as reliable as those

in Fortis. The September Opinion expressed skepticism of PTRS as subjective and

easy to manipulate.151 It remains unclear how PTRS is calculated and what objective

data it uses.152 The September Opinion afforded Alexion’s reduced post-HV-108

PTRS little weight as an explanation for Alexion’s decision to terminate

ALXN1830, instead concluding Alexion shut down the program to deliver on

151 See Sept. Op. at *25–26, *43. See Lee Tr. 475, 482, 652–53; Pradhan Tr. 898–99; Washburn Tr. 646; Borboroglu Tr. 152

1403; Sept. Op. at *25 n.449. 49 AstraZeneca’s promise of merger synergies. 153 But the record offers no reason to

doubt Alexion’s pre-HV-108 PTRS numbers, despite its subjective inputs. The pre-

HV-108 numbers are a conservative starting point and a responsible one.

a. Calculation Of Expected Milestone Payments

With the pre-HV-108 PTRS probabilities as the starting point for the estimate

of damages, I have calculated each milestone’s expected payment by weighting the

milestone amount by its probability of achievement.154

Expected Milestone Payments Milestone Milestone Amount Probability Expected Payment 2 $ 120,000,000.00 0.715 $ 85,800,000.00 3 $ 120,000,000.00 0.215 $ 25,800,000.00 4 $ 150,000,000.00 0.538 $ 80,700,000.00 5 $ 150,000,000.00 0.102 $ 15,300,000.00 6 $ 25,000,000.00 0.295 $ 7,375,947.45 7 $ 25,000,000.00 0.0256 $ 640,052.55 8 $ 80,000,000.00 0.0511 $ 4,088,160.00 Total $ 219,704,160.00

b. Present Value of Expected Milestone Payments

From there, the expected milestone payments must be discounted to present

value at the time of breach to put SRS in the economic position it would have been

153 See Sept. Op. at *43–44, *47. The post-HV-108 data, even if it provided complete probability estimates, would be an unreliable starting point for a damages estimate. 154 Blitzstein & Hwang, supra note 12, at 149–52. 50 in absent a breach. 155 The present value of a future lump sum (“𝑃𝑃𝑃𝑃”) is a function

of the estimated future value of the lump sum (“𝐹𝐹𝐹𝐹”), a discount rate reflecting risk

(“𝑟𝑟”), and the number of periods between the date the lump sum is received and the

present date (“𝑛𝑛”). 156 The formula for present value can be expressed as 𝑃𝑃𝑃𝑃 =

𝐹𝐹𝐹𝐹/(1 + 𝑟𝑟)𝑛𝑛 .

For each milestone, 𝐹𝐹𝐹𝐹 is the milestone’s expected payment shown in the last

section. I address the variables 𝑛𝑛 and 𝑟𝑟 in turn.

i. Estimating n

The variable n in the present value formula varies based on each milestone’s

expected payment date at the time of breach. For Milestones 2 through 5, the best

evidence of the expected payment dates is a slide deck prepared for the December

14, 2021, meeting in which Alexion decided to terminate ALXN1830.157 The deck

contains draft developmental timelines for TED and cAMR, which evince the

following expected milestone schedule 158:

155 See Duncan, 775 A.2d at 1022; Fortis, 2024 WL 4048060, at *53. 156 See Aswath Damodaran, Investment Valuation: Tools and Techniques for Determining the Value of Any Asset 11–12 (3d ed. 2012). 157 JX 2220 at 14–15. 158 Id. 51 Expected Achievement Schedule for Milestones 2 Through 5 Milestone Month/Quarter of Achievement 2 June 2025 3 August 2025 4 Q2 2028 5 Q2 2029

The draft timelines do not indicate how long after FDA approval EMA

approval would be obtained. In 2018, Alexion predicted EMA approval would occur

two years after FDA approval for a given indication. 159 ATP’s 2018 valuation

assumes FDA approval and EMA approval would occur at approximately the same

time. 160 Applying Alexion’s more conservative timing estimate for EMA approval,

I conclude Milestone 6 was expected to be achieved in Q2 2030, and Milestone 7

was expected to be achieved in Q2 2031.

As for Milestone 8, ATP’s valuation conservatively assumes that if the

milestone were achieved, it would be achieved during the year with the highest

projected sales. 161 ATP’s conservative approach is appropriate for a responsible

estimate of damages. At the time of breach, Alexion projected 2036 as its peak sales

159 See JX 643 at 23. 160 JX 2962 at 30. 161 Id. at 24. 52 year.162 Because Alexion’s fiscal year coincides with the calendar year, I conclude

Milestone 8’s expected achievement date was December 31, 2036.163

The Merger Agreement provides that a given milestone payment is due 45

days after the milestone is achieved. 164 Because Milestones 2 through 7 do not have

an exact expected achievement date, I assume the milestones would be achieved at

the midpoint of the relevant month or quarter. 165

Milestone Expected Achievement Dates and Payment Due Dates Milestone Expected Achievement Date Expected Payment Due Date 2 June 15, 2025 July 30, 2025 3 August 15, 2025 September 29, 2025 4 May 17, 2028 July 1, 2028 5 May 17, 2029 July 1, 2029 6 May 17, 2030 July 1, 2030 7 May 17, 2031 July 1, 2031 8 December 31, 2036 February 14, 2037

162 JX 1953. 163 JX 1708 at 1. 164 Merger Agr. § 3.8(e). 165 Alexion argues SRS cannot be compensated for Milestones 4 through 8 because these timelines extend beyond Alexion’s seven-year CRE Obligation. ALXN Ans. Suppl. Br. 28–29. I disagree. Alexion’s timelines (and its PTRS projections, for that matter) implicitly incorporate information available to Alexion, including Alexion’s seven-year CRE Obligation and its continuing obligation not to take action with the primary purpose of avoiding any milestone. See Merger Agr. § 3.8(f). In theory, the PTRS estimates might have been higher if Alexion’s CRE Obligation extended longer or indefinitely. The corresponding differential in expected value has already been accounted for. The remaining expected value reduction was caused by Alexion’s breach. 53 ii. Estimating r

“A discount rate reflects both the time value of money and risk.” 166 It

represents the total expected rate of return an investor demands.167 SRS proposes

using Alexion’s 9% weighted average cost of capital (“WACC”) from 2018 as the

discount rate. 168 But SRS’s valuation expert was excluded,169 SRS does not explain

why it believes Alexion’s WACC is the appropriate discount rate, and I do not think

it is appropriate except as applied to Milestone 8.

A company’s cost of capital is the expected rate of return the company needs

to attract investors. 170 It can be thought of as an opportunity cost. 171 Riskier

investments have a higher cost of capital because capital is a limited resource: when

investors commit funds to riskier projects, they are foregoing safer alternatives. So

investors demand a risk premium over the risk-free rate: higher expected returns to

compensate for taking on additional risk. 172

Shannon P. Pratt & Roger J. Grabowski, Cost of Capital in Litigation: Applications and 166

Examples 4 (2011). 167 Id. at 5. 168 SRS Op. Suppl. Br. 22; Sarin Tr. 1020. 169 D.I. 312. 170 See Pratt & Grabowski, supra note 166, at 1. 171 Id. 172 See id. at 110. 54 In lost profits cases, the cost of capital may be the proper discount rate.173

“[B]ecause the plaintiff no longer has to bear the investment’s risk, the plaintiff is

not entitled to be compensated for the risk factors.” 174 So the damages award in a

lost profits case must discount not only for the time value of money (using the risk-

free rate), but also for the risk premium. 175 Because the cost of capital incorporates

both, it is often used as the discount rate in such cases. 176

Here, for Milestones 2 through 7, there is no risk premium on the right to

receive milestone payments. To be sure, the probability-weighted expected

payments are not “already risk adjusted” just because they “factor[] in the possibility

of good and bad outcomes.”177 Expected value is a risk-neutral metric. 178 But the

risks underlying the metrics in Milestones 2 through 7 are “diversifiable.” 179 A

173 Id. at 109–10; Siga II, 132 A.3d at 1123. Pratt & Grabowski, supra note 166, at 110 (quoting R.F. Lanzillotti & A.K. Esquibel, 174

Measuring Damages in Commercial Litigation: Present Value of Lost Opportunities, J. Acct., Auditing & Fin., Winter, 1990, at 125–42). 175 Id. 176 See id.; Cede & Co. v. Technicolor, Inc., 884 A.2d 26, 39 (Del. 2005). 177 See Damodaran, supra note 156, at 906. 178 See id.; cf. Pratt & Grabowski, supra note 166, at 21. An investment with a 100% chance of returning $90 has the same expected value as one with a 90% chance of returning $100 and a 10% chance of returning $0. But the second investment involves greater risk and so commands a risk premium. 179 See Appraisal Foundation, Valuations in Financial Reporting Valuation Advisory 4: Valuation of Contingent Consideration 23 (2019) (noting examples of diversifiable risks in the context of contingent consideration, including “a payment contingent upon receiving regulatory approval,” a payment upon the “achievement of technical milestones,” and the

55 diversifiable risk is one “that is peculiar to an individual company.” 180 Such risks

can be “diversified away” through a broad investment portfolio “due to the law of

large numbers.” 181 In the context of contingent consideration, diversifiable risks

include product development milestones (like Milestones 2 and 3) and regulatory

approval milestones (like Milestones 4 through 7).182 Because the risk associated

with Milestones 2 through 7 can be diversified away, that risk commands no risk

premium, and there is no need to discount for a risk premium in awarding damages.

Alexion’s cost of capital is not an appropriate discount rate.

For Milestones 2 through 7, the discount rate is the risk-free rate plus a credit

risk premium.183 For a given milestone, that sum is Alexion’s cost of debt “specific

to the term and seniority of the earnout obligation.” 184 ATP estimated this amount

based on the yield for investment-grade corporate bonds with maturities equal to the

“development of a new product”); id. at 23 n.21 (“While there may be a small degree of systematic risk associated with the achievement of technical or regulatory milestones, in most cases, the non-diversifiable risk is de minimis as compared to the diversifiable risk. Assuming the risk associated with such events to be diversifiable is therefore generally considered reasonable.”). 180 Id. at 23, 23 n.20. 181 Id. at 23 (“A widely accepted valuation principle assumes that rational investors and market participants reduce risk through diversification. As a result, it is assumed that market participants will only require a return premium for those risks that cannot be diversified away.”). 182 Id. at 14, 23, 79. 183 Id. at 39. 184 See id. at 96, 109. 56 expected time to reach a given milestone. 185 So ATP used the bond yield as the

discount rate. 186

ATP’s methodology makes sense because “[t]he interest rates on bonds are

determined by the default risk that investors perceive in the issuer of the bonds,” and

because bond interest rates can be estimated by “the yield on the bond.”187 Without

access to granular data like ATP had, my analysis conservatively uses Moody’s

Seasoned Baa Corporate Bond Yield as of January 2022, which is 3.58%, for

Milestones 2 through 7. 188 This figure is conservative for three reasons. First,

because the breach occurred on December 14, 2021, both the December 2021 yield

and the January 2022 yield were candidates. I chose the higher yield, resulting in

greater discounting. Second, the figure is based on bonds with maturities 20 years

and above. 189 Because default risk increases with time, the figure is some degree

greater than it would be if it were based on corporate bonds with maturities matching

the expected time to reach each milestone. Third, “Baa” is Moody’s lowest

185 See JX 2962 at 22–25, 31 n.2. 186 See id. 187 Damodaran, supra note 156, at 177. 188 Moody’s Seasoned Baa Corporate Bond Yield, retrieved from FRED, Fed. Rsrv. Bank of St. Louis; https://fred.stlouisfed.org/series/BAA (last visited June 11, 2025). 189 Id. 57 investment grade rating.190 Incorporating higher grade bonds would reduce the yield

and thus reduce the discount rate.

Milestone 8 is different because the risk associated with its underlying metric

is “nondiversifiable.” As the name suggests, nondiversifiable (or “systemic”) risk,

“cannot be fully removed through diversification” because it is “correlated with the

market.” 191 Metrics with nondiversifiable risk include financial metrics such as

EBITDA or net sales.192 If an earnout metric carries systemic risk, the discount rate

must include a risk premium “commensurate with the degree” of systemic risk.193

Milestone 8’s net sales metric carries nondiversifiable risk. 194 So the

milestone’s discount rate must therefore include a risk premium. SRS’s suggested

9% discount rate includes a risk premium. Without expert guidance, the

appropriateness of the 9% figure is uncertain. Alexion’s WACC may have changed

since 2018. And it is unclear whether Alexion’s WACC is the best metric to measure

the discount rate in this context. But “[d]oubts [about the amount of damages] are

generally resolved against the party in breach.”195 As the wrongdoer, Alexion cannot

190 R. Glenn Hubbard & Anthony Patrick O’Brien, Economics 259 (7th ed. 2018). 191 Appraisal Foundation, supra note 179, at 23. 192 See id. at 24. 193 Id. at 24, 32–33. 194 See id. at 24. 195 Cura, 2001 WL 1334188, at *20 (quoting Restatement (Second) of Contracts § 352 cmt. a (1981)). 58 be allowed to profit from its breach via an overly conservative discount rate for

Milestone 8, particularly where Alexion offers no input as to the appropriate

discount rate.196 Given Delaware’s recognition of the wrongdoer rule, I believe 9%

is an appropriate discount rate for Milestone 8 taking into account “all the

circumstances of the breach.”197

iii. Pre-Interest Expectation Damages Calculation

I have calculated the present value of expected earnout payments based on the

expected achievement dates noted above, and an annual discount rate of 3.58% for

Milestones 2 through 7 and 9% for Milestone 8. The following table summarizes

my full analysis.

Present Value of Expected Milestone Payments Milestone Milestone Payout Expected Payment Present Value 2 $ 120,000,000.00 $ 85,800,000.00 $ 75,522,351.83 3 $ 120,000,000.00 $ 25,800,000.00 $ 22,576,414.19 4 $ 150,000,000.00 $ 80,700,000.00 $ 64,092,303.86 5 $ 150,000,000.00 $ 15,300,000.00 $ 11,731,346.77 6 $ 25,000,000.00 $ 7,375,947.45 $ 5,460,071.74 7 $ 25,000,000.00 $ 640,052.55 $ 457,425.38 8 $ 80,000,000.00 $ 4,088,160.00 $ 1,105,001.54 Total $ 180,944,915.32

196 See generally Alexion Ans. Suppl. Br. 197 Cura, 2001 WL 1334188, at *20 (quoting Restatement (Second) of Contracts § 352 cmt. a (1981)); Siga I, 2014 WL 3974167, at *8; Beard Rsch., 8 A.3d at 613. 59 SRS’s pre-interest expectation damages for Alexion’s breach of its CRE Obligation

are $180,944,915.32.

B. SRS’s Damages Claim For Breach Of Merger Agreement Section 3.8(f)’s Requirement Not To Take Action To Avoid Milestones

Count IV of SRS’s amended complaint alleges Alexion breached its Non-

Avoidance Obligation by taking actions with the primary purpose of avoiding

milestones. 198 The September Opinion did not address whether Alexion’s

termination of ALXN1830 breached the Non-Avoidance Obligation. The

September Opinion asked the parties to advise if SRS’s Count IV carried with it any

additional potential for damages or practical ramifications.199 Despite SRS’s best

efforts in supplemental briefing,200 Count IV does not carry any additional potential

for damages. 201 SRS is receiving its expectation damages based on Alexion’s

breach of the CRE Obligation. I do not address Count IV.

198 Compl. ¶¶ 272–78. 199 Sept. Op. at *48. 200 SRS Op. Suppl. Br. 35–39; SRS Reply Suppl. Br. 21–22. 201 The September Opinion held SRS had not met its burden to prove Alexion’s decisions to deprioritize the gMG and WAIHA programs were intended to avoid milestones in breach of its Non-Avoidance Obligation. Sept. Op. at *41 n.620 (citing SRS Op. Br. 68–70). I do not consider SRS’s supplemental arguments regarding Alexion’s deprioritization of these indications. SRS Op. Suppl. Br. 36. 60 1. Application Of The Prevention Doctrine Would Not Result In Additional Damages.

SRS contends a breach of Alexion’s Non-Avoidance Obligation would trigger

Delaware’s prevention doctrine. SRS asserts that under that doctrine, unless Alexion

can prove its breach did not materially contribute to the nonoccurrence of a given

milestone, SRS is entitled to full payment on each milestone. For purposes of this

argument, I accept, with some hesitation, SRS’s assertion that the Non-Avoidance

Obligation is a contractual codification of Delaware’s prevention doctrine.

The prevention doctrine “provides that a party may not escape contractual

liability by reliance upon the failure of a condition precedent where the party

wrongfully prevented performance of that condition precedent.”202 “[T]he doctrine

is based on the long-established principle of law that a party should not be able to

take advantage of its own wrongful act.” 203 When the doctrine applies, the

preventing party is “liable for damages caused by the breach.”204

But the “plaintiff is not entitled to take advantage of this situation” to obtain

a windfall.205 “[T]he damages recoverable represent the harm resulting from th[e]

BitGo Hldgs., Inc. v. Galaxy Digit. Hldgs., Ltd., 319 A.3d 310, 333 (Del. 2024) (quoting 202

Mobile Commc’ns Corp. of Am. v. MCI Commc’ns Corp., 1985 WL 11574, at *4 (Del. Ch. 1985)). 203 13 Richard A. Lord, Williston on Contracts § 39:6 (4th ed. 2024). 204 Id. § 39:12. 205 See id. (quoting Siegal v. Haver, 417 P.2d 928, 932 (Ariz. App. 1966)). 61 lack of cooperation.”206 For instance, “[i]f the defendant’s conduct in preventing the

plaintiff’s performance has enabled the plaintiff to avoid expenses, the expenses

saved must be deducted from the damages otherwise recoverable, for otherwise, the

plaintiff would be overcompensated as a result of the defendant’s breach.”207 This

“principle of mitigation” reflects Delaware’s broader policy against awarding

windfalls. 208 Consistent with that policy, the prevention doctrine does not provide a

plaintiff a back door to damages in excess of proven expectation damages. 209

With interest, SRS pegs full milestone payment damages at

$754,877,262.02.210 But as explained, SRS’s expectation damages amount to the

present value of the expected value of those milestones. With interest, this opinion

206 Id. 207 Id. 208 See Paul v. Deloitte & Touche, LLP, 974 A.2d 140, 146 (Del. 2009) (“Contract damages are designed to place the injured party in an action for breach of contract in the same place as he would have been if the contract had been performed. Such damages should not act as a windfall.” (internal quotation marks and citations omitted)); cf. Stayton v. Del. Health Corp., 117 A.3d 521, 534 (Del. 2015) (“In Delaware ‘a plaintiff is entitled to compensation sufficient to make him whole, but no more.’ In other words, the remedy for the tort should put the plaintiff as close as possible to the same position as she was in before the injury.” (quoting Mitchell v. Haldar, 883 A.2d 32, 38 (Del. 2005))). 209 See Murphy Marine Servs. of Del., Inc. v. GT USA Wilm., LLC, 2022 WL 4296495, at *2, *9, *12–14, *16, *21 (Del. Ch. Sept. 19, 2022) (holding that even if obtaining “a final valuation decision was a condition precedent to [defendant’s] performance, the failure of such a condition is excused under the prevention doctrine” because defendant’s contract breach caused the failure, and therefore awarding expectation damages caused by the breach). 210 SRS Op. Suppl. Br. at 38–39. 62 calculates those damages at around $220 million. SRS’s application of the

prevention doctrine would result in a half-a-billion-dollar windfall. Delaware law

does not permit that. Given this opinion’s award of expectation damages, the parties’

expectations have been enforced, and Alexion has not been permitted to take

advantage of its breach. The prevention doctrine cannot offer SRS more than its

expectation damages.

SRS offers no other practical ramifications from Alexion’s alleged breach of

its Non-Avoidance Obligation.211 Count IV is moot.

C. Pre- And Post-Judgment Interest

Alexion does not contest SRS’s entitlement to pre- and post-judgment interest

on its damages. The modern approach calls for compounding interest using a

floating interest rate based on the legal rate.212 The modern approach, compounding

at quarterly intervals, is appropriate for pre- and post-judgment interest here.

D. Attorneys’ Fees And Expenses

SRS’s post-trial brief argues it is entitled to reasonable attorneys’ fees and

expenses under Section 8.2 of the Merger Agreement.213 Section 8.2 provides for

211 See SRS Op. Suppl. Br. 35–39. ITG Brands, LLC v. Reynolds Am., Inc., 2025 WL 670818, at *12–14 (Del. Ch. Mar. 3, 212

2025) (collecting cases). SRS Op. Br. 90–91. SRS raises the argument again in its supplemental briefing. SRS 213

Op. Suppl. Br. 39. As the September Opinion requested supplemental briefing only on the

63 indemnification against “Losses . . . arising out of or resulting from . . . any breach

of any covenant” in the Merger Agreement. 214 “Losses” are defined to include

“reasonable attorneys’ fees and expenses.”215

The Merger Agreement requires written notice for any indemnification claim,

which must “state in reasonable detail the nature, basis and the amount of the Direct

Claim, to the extent known, along with copies of the relevant documents evidencing

such Direct Claim and the basis for indemnification sought.” 216 SRS makes no

argument that it has satisfied the notice requirements. 217 SRS is not entitled to

attorneys’ fees and expenses under the Merger Agreement’s indemnification

provisions at this point in time.

III. CONCLUSION

SRS is awarded $180,944,915.32 in damages, plus pre- and post-judgment

interest, for Alexion’s breach of its obligation to use CREs. Within 30 days, the

parties shall confer on an interest calculation consistent with the methodology

proper damages model, I do not consider SRS’s supplemental briefing regarding attorneys’ fees and expenses. 214 Merger Agr. § 8.2. 215 Id. § 8.1. 216 Id. § 8.3(d). 217 See Compl. at 104–05 (SRS’s complaint failing to reference Merger Agreement Section 8.2 and making only a passing mention of SRS’s alleged entitlement to attorneys’ fees); SRS Op. Br. 90–91; SRS Ans. Br. 69. 64 adopted in this opinion and submit a proposed stipulated order implementing this

opinion and the damages awarded in the September Opinion.