United States v. Google LLC

United States Court of Appeals

FOR THE DISTRICT OF COLUMBIA CIRCUIT

No. 25-5016 September Term, 2024

FILED ON: MARCH 21, 2025

UNITED STATES OF AMERICA, ET AL.,

APPELLEES

v.

GOOGLE LLC,

APPELLEE

APPLE INC.,

APPELLANT

Appeal from the United States District Court

for the District of Columbia

(No. 1:20-cv-03010)

Before: CHILDS, PAN, and GARCIA, Circuit Judges

JUDGMENT

This appeal was considered on the record from the United States District Court for the

District of Columbia and the briefs of the parties. See D.C. CIR. R. 34(j). The Court has afforded

the issues full consideration and determined that they do not warrant a published opinion. See

D.C. CIR. R. 36(d). For the reasons stated below, it is

ORDERED and ADJUDGED that the district court’s order be AFFIRMED.

* * *

On December 23, 2024, Apple moved to intervene in this antitrust action against Google.

The sole question on appeal is whether the district court abused its discretion in denying that

motion as untimely. It did not. The most important timeliness consideration is whether Apple

moved “to intervene as soon as it became clear that its interests would no longer be protected by

the parties in the case.” Campaign Legal Ctr. v. FEC, 68 F.4th 607, 610 (D.C. Cir. 2023) (internal

quotation marks omitted) (quoting Cameron v. EMW Women’s Surgical Ctr., P.S.C., 595 U.S. 267,

279–80 (2022)). As the district court found, it was clear that Google would no longer protect

Apple’s interests at least by the time plaintiffs filed their proposed remedy framework on October

8. Yet Apple waited another seventy-six days to intervene. The district court also reasonably

found that the other timeliness factors disfavor Apple. We therefore affirm.

I

In 2020, the federal government and forty-nine states sued Google under the Sherman Act,

alleging that Google had violated Section 2 of the Act by “us[ing] anticompetitive tactics to

maintain and extend its monopolies in the markets for general search services, search advertising,

and general search text advertising.” A. 256 ¶ 1. Plaintiffs claimed that Google’s anticompetitive

tactics included “enter[ing] into exclusionary agreements” with “distributors—including popular-

device manufacturers such as Apple”—“to secure default status for its general search engine.”

A. 256–57 ¶ 4. The complaint particularly described Google’s distribution agreement with Apple,

the Internet Services Agreement (ISA). See, e.g., A. 280–81 ¶¶ 85–86; A. 290–92 ¶¶ 118–22; see

also A. 435 ¶ 290. Under that agreement, Apple preloads Google as the default search engine on

Apple devices in exchange for a portion of the advertising revenue Google generates from searches

on those devices (a practice known as “revenue sharing”). See A. 280–81 ¶ 86; A. 336–37; A. 538.

Plaintiffs alleged that through its ISA with Apple, Google “substantially forecloses [its] search

rivals from an important distribution channel,” and that “[b]y paying Apple a portion of the

monopoly rents extracted from advertisers, Google has aligned Apple’s financial incentives with

its own.” A. 291 ¶¶ 121–22.

Given the centrality of Apple’s ISA to the litigation, Apple closely monitored these

proceedings from the start. It participated in discovery. See A. 319. And two Apple executives

testified about the ISA at trial. See A. 612; A. 435–36 ¶¶ 291–95.

After four years of litigation and a forty-four-day bench trial, on August 5, 2024, the district

court found that Google “is a monopolist” and “has violated Section 2 of the Sherman Act.”

A. 338. Google has “achieved market dominance,” the court explained, in part “through [its]

distribution contracts.” A. 336–37. The court found that those “agreements are exclusive and

have anticompetitive effects.” A. 338. And it specifically discussed Google’s ISA with Apple,

noting that Google’s revenue-share payments to Apple—$20 billion in 2022 alone—

disincentivized Apple from developing a competing search engine. See A. 575–76; A. 550.

Following its liability decision, the district court in mid-September set a timeline for the

remedies phase of the trial. It scheduled discovery to begin in late September and calendared a

multi-week remedies hearing (essentially a second bench trial) for April 2025. In the lead-up to

the remedies hearing, on October 8, 2024, plaintiffs submitted a proposed remedy framework

outlining the remedies they planned to request. On November 20, plaintiffs filed a more detailed

proposed final judgment. And on December 20, Google submitted its own proposed final

judgment.

Apple moved to intervene as of right on December 23—140 days after the district court

issued its liability decision, 76 days after plaintiffs first outlined the remedies they would seek, and

33 days after plaintiffs submitted their proposed final judgment. The district court denied Apple’s

intervention motion as untimely, finding that Apple’s asserted need to intervene had become clear

by the time plaintiffs filed their proposed remedy framework on October 8. The court did,

however, permit Apple to participate as amicus and to file up to two fact-witness affidavits and a

2

post-hearing brief. Apple appealed and moved for a stay of the district court’s denial pending

appeal. The district court and our court denied Apple’s stay motion.

II

We review the district court’s denial of a motion to intervene as of right for abuse of

discretion. See Campaign Legal Ctr., 68 F.4th at 610. An abuse of discretion occurs if the district

court makes a legal error or relies on clearly erroneous factual findings. Id. Whether a motion to

intervene is timely must be “judged in consideration of all the circumstances.” Id. (cleaned up).

“The most important circumstance,” however, “is whether a party sought to intervene as soon as

it became clear that its interests would no longer be protected by the parties in the case.” Id.

(cleaned up).

A

To explain why Google can no longer defend its interests, Apple points to two ways in

which its and Google’s interests have diverged. Apple’s argument on appeal hinges on the

proposition that this divergence was not clear until November 20. The district court appropriately

found, however, that it was apparent by October 8.

First, Apple contends that the companies’ interests diverged once it became clear that

plaintiffs aimed to end not only the ISA’s guarantee of default status for Google but also its

guarantee of revenue-sharing payments to Apple. As Apple sees it, the companies have “vastly

different priorities in defending default status versus revenue share.” Appellant’s Brief 29. Google

would prioritize defending its ability to contract for default status over defending its ability to share

search revenue with Apple. And Google would defend the revenue-sharing arrangement only so

long as its search engine retained default status; “if, or once, the district court takes default status

off the table . . . , Google will then have little incentive to protect Apple’s interest in revenue

share.” Appellant’s Brief 30. Apple, by contrast, would argue that revenue sharing should be

permitted even if it could not make Google the default search option on its devices. Indeed, Apple

has long (unsuccessfully) sought the flexibility to collect revenue-sharing payments from Google

irrespective of Google’s default status. See Appellant’s Brief 17, 27, 29–30; see also A. 444–45

¶¶ 319–20.

The question, then, is when it became clear that plaintiffs would seek to end revenue-

sharing payments. Apple says that goal became clear only once plaintiffs filed their proposed final

judgment on November 20, because plaintiffs there proposed to prohibit Google from “offer[ing]

or provid[ing] anything of value to Apple . . . that in any way creates an economic disincentive for

Apple to compete in or enter the [general search engine] or Search Text Ad markets.” A. 682.

The district court reasonably rejected that view.

The August liability opinion arguably made clear that plaintiffs would pursue an end to

revenue sharing. In that opinion, the district court found it “unquestionabl[e]” that revenue-sharing

payments from Google to Apple “significantly contribut[e] to keeping Apple on the sidelines of

search, thus allowing Google to maintain its monopoly.” A. 576. The court acknowledged other

reasons Apple might be disinclined to build a competitor search engine. A. 575–76. The court

nonetheless concluded that the “prospect of losing tens of billions in guaranteed revenue from

3

Google” at least “disincentive[d] Apple from launching its own search engine when it otherwise

has built the capacity to do so”; and the court found that this factor alone was sufficient “to

constitute an anticompetitive effect.” A. 576. Apple conspicuously fails to address the liability

opinion’s discussion of the anticompetitive effects of revenue sharing, even though plaintiffs

repeatedly emphasize those findings in their brief. Compare Appellees’ Brief 2, 10, 32

(emphasizing the discussion at A. 576), with Reply Brief 11 (ignoring that portion of the district

court’s decision).

Given the district court’s discussion, it was quite likely that plaintiffs would pursue an

order enjoining revenue sharing between Google and Apple (and not just such payments as they

were tied to the ISA provisions protecting Google’s default placement). As the district court

reasonably concluded, the October proposed remedy framework removed any remaining doubt

about plaintiffs’ intentions. In that filing, plaintiffs stated that they were considering remedies that

“limit or prohibit default agreements, preinstallation agreements, and other revenue-sharing

arrangements related to search and search-related products.” A. 627. They explained they were

doing so in part “because Google’s monopoly-funded revenue share payments disincentivize its

partners from diverting queries to Google’s rivals.” A. 626. And they described that the remedies

“should account for alternative and future forms of monopoly maintenance.” A. 624; see also

Zenith Radio Corp. v. Hazeltine Rsch., Inc., 395 U.S. 100, 132–33 (1969) (explaining that in

designing remedies for anticompetitive conduct “it is not necessary that all of the untraveled roads

to that end be left open and that only the worn one be closed” (quoting Int’l Salt Co. v. United

States, 332 U.S. 392, 400 (1947)). As the district court put it, plaintiffs’ proposed remedy

framework “laid bare the very cleavage that Apple [now] claims is the reason Google can no longer

adequately represent its interests.” A. 977.

Apple’s answer is unresponsive. Apple repeatedly notes the breadth of the specific remedy

requested in plaintiffs’ November proposed final judgment. Apple may be right that the broad

remedy requested in November further emphasized the gap between Apple’s interests and

Google’s interests, but that does not mean it caused the divergence or made it “clear” in the first

instance. And Apple cannot persuasively explain how, after reviewing both the liability opinion

and the proposed remedy framework, it still reasonably expected that plaintiffs would not pursue

remedies targeted both at removing Google’s default status and at ending Google’s revenue-

sharing payments to Apple. To the contrary, as the district court found, it was clear by at least

October 8 that plaintiffs would pursue both of those remedies.

Second and alternatively, Apple contends that Google will now have less capacity and

incentive to advocate for any version of the ISA, because Google will have to defend against other

forms of relief that threaten major aspects of its business. See Appellant’s Brief 28. The district

court also reasonably concluded that this dynamic was present before plaintiffs filed their

November proposed final judgment. The district court’s August liability decision discussed

multiple aspects of Google’s business, including Google’s control of its Chrome browser, Android

operating system, and Pixel phones. See, e.g., A. 271 ¶ 53; A. 453–55 ¶¶ 351–56; A. 492–93.

Plaintiffs’ October framework then stated that they anticipated seeking “a number of mutually

reinforcing remedies” addressing “four categories of harm” that together generated Google’s

monopoly power. A. 625–26. For example, plaintiffs described that they were “considering

behavioral and structural remedies that would prevent Google from using products such as

4

Chrome, Play, and Android to advantage Google search” and emphasized that “Google’s

longstanding control of the Chrome browser, with its preinstalled Google search default,

significantly narrows the available channels of distribution and thus disincentivizes the emergence

of new competition.” A. 627 (internal quotation marks omitted). The framework also listed many

other remedies relating to data disclosure, independent licensing or syndication of Google’s ad

feed, and limitations on development of Google’s AI model. A. 627–29.

Apple counters that, before November 20, it was uncertain whether plaintiffs would seek

to force Google to divest from Chrome. See Appellant’s Brief 28–29, 32. Even if that were true,

however, it is unclear why the divergence of the companies’ interests would be triggered only once

plaintiffs requested that specific remedy. Had plaintiffs never requested divestiture, Google would

be faced with another version of the same problem; it would still need to defend against proposed

remedies targeting a range of its business practices, meaning Apple’s ISA would not be Google’s

sole or primary priority.

Lastly, even if Apple were right that the divergence of its and Google’s interests became

clear only on November 20, Apple has never explained why it waited thirty-three additional days

to intervene. In context, that delay seems difficult to justify; the remedies hearing was set for April

2025, and the district court repeatedly requested promptness given the case’s gravity and

complexity and the need to preserve institutional knowledge. See A. 978–79. Even if November

20 was the key date, then, the district court likely had discretion to deny the motion as untimely.

Cf., e.g., NAACP v. New York, 413 U.S. 345, 366–69 (1973) (finding motion untimely after

seventeen days, in case affecting “rapidly approaching primary elections”); Roeder v. Islamic

Republic of Iran, 333 F.3d 228, 233 (D.C. Cir. 2003) (finding that intervention was timely when

party intervened within “less than thirty days”). But cf., e.g., Nat’l Wildlife Fed’n v. Burford, 878

F.2d 422, 434 (D.C. Cir. 1989) (in slower administrative case, finding intervention timely after 73

days), rev’d on other grounds sub nom. Lujan v. Nat’l Wildlife Fed’n, 497 U.S. 871 (1990).

B

The district court also did not abuse its discretion in addressing the three other factors that

bear on timeliness—“the purpose for which intervention is sought, the need for intervention as a

means of preserving the applicant’s rights, and the probability of prejudice to those already parties

in the case.” Campaign Legal Ctr., 68 F.4th at 610 (cleaned up).

To start, the district court thoroughly recounted why the existing parties would be

prejudiced by Apple’s belated intervention. Parties may be prejudiced by delays that stall the start

of trial or shortchange their preparation time. See, e.g., United States v. Brit. Am. Tobacco Austl.

Servs., Ltd., 437 F.3d 1235, 1238–39 (D.C. Cir. 2006); Amador Cnty. v. Dep’t of the Interior, 772

F.3d 901, 905–06 (D.C. Cir. 2014). Here, as the district court explained, by the time it (promptly)

resolved Apple’s motion, only one month of fact discovery remained and there were just three

months before the remedies hearing. A. 987. If Apple intervened, the parties would need to

“request documents from and depose” Apple’s additional witnesses, including “an (as-yet-

unnamed) expert” witness to which the parties would “have to devote considerable time and

resources.” J.A. 986. The district court also explained that it could grant the intervention motion

only by taking hearing time away from the existing parties or moving the trial—and either option

would harm the existing parties. See A. 987–88. With little time left and less scheduling

5

flexibility, the court said, adding more days to the remedies hearing was no longer feasible; even

when the existing parties requested fourteen additional days, the district court could add only five

“without compromising” the decision deadline set months earlier. A. 987 & n.5. On abuse of

discretion review, we decline Apple’s invitation to question the district court’s well-explained

management of its own calendar. And we find, contrary to Apple’s claims, that the district court’s

analysis was not “untethered to [Apple’s] 76-day delay.” Appellant’s Brief 41. As the district

court wrote, it was that long delay that reduced scheduling “flexibility” and cut into preparation

time. A. 987–88. 1

Finally, the district court reasonably concluded that Apple did not show that the purpose

and need for party status favored intervention. Apple says that it must be permitted to intervene

so that it can cross-examine witnesses and present live testimony. The district court disagreed,

finding that Apple failed to demonstrate “that the information it wishes to present is any different

than that the court already considered—and largely credited—during the liability phase.” A. 982.

And as the court emphasized, Apple will be permitted to submit fact-witness affidavits and a post-

hearing brief. See A. 985.

Apple shows no abuse of discretion in those determinations. Apple argues in general terms

that it would offer testimony addressing the dynamics presented by plaintiffs’ proposed remedies

and that its executives’ earlier testimony was “historical” and needs updating. Appellant’s Brief

37. The district court responded that it was unclear why “party status is necessary to present such

evidence, especially when Google is likely to call [an Apple executive] as a witness” and noted

that Apple will be able to submit affidavits for the court’s consideration. A. 983; see A. 990. We

find no error in the district court’s analysis, much less one that would warrant reversal given the

other, more important timeliness factors: Apple’s delay in filing its motion and the prejudice to the

parties that would result.

* * *

For the foregoing reasons, we affirm the district court’s denial of Apple’s motion to

intervene.

Pursuant to D.C. Circuit Rule 36, this disposition will not be published. The Clerk is

directed to withhold issuance of the mandate herein until seven days after resolution of any timely

petition for rehearing or rehearing en banc. See FED. R. APP. P. 41(b); D.C. CIR. R. 41(a)(1).

Per Curiam

1

Apple also implies that it was an abuse of discretion for the district court to consider the potential

for additional, follow-on third parties to intervene if Apple’s motion were granted. But that consideration

was not improper—and Apple does not offer a case suggesting it was. To the contrary, our court has

endorsed that consideration as an aspect of the prejudice inquiry. See Deutsche Bank Nat’l Tr. Co. v. FDIC,

717 F.3d 189, 192 (D.C. Cir. 2013).

6

FOR THE COURT:

Clifton B. Cislak, Clerk

BY: /s/

Michael C. McGrail

Deputy Clerk

7