What Google Pays for Default Search Placement

By Bill Hartzer · Updated 17 September 2026

The number, and where it comes from

Google paid more than $26 billion in revenue-share payments for default search placement in 2021. That figure is a finding of the US District Court for the District of Columbia, recorded in its memorandum opinion in United States of America v. Google LLC, No. 1:20-cv-03010-APM, the redacted public version of which was released in 2024. It was sealed until that opinion published.

The comparison the court drew around it is the part that gets lost. The $26 billion was nearly four times more than all of Google's other search-specific costs combined. The same record puts the cost of actually operating the search engine — the crawling, the index, the serving infrastructure, the ranking systems — at $8.4 billion in 2020.

Item	Amount
Revenue-share payments for default placement	More than $26 billion (2021)
All other search-specific costs combined	Roughly a quarter of that, by the court's comparison (2021)
Cost of operating the search engine	$8.4 billion (2020)

Read the table carefully, because the two years are not the same and should not be netted against each other. What survives that caution is the shape: Google spends several times more on being placed in front of users than on the machinery that answers them. For a business commonly described as winning on technology, that is a striking allocation, and it is the single most informative fact in the case record about how search competition actually works.

A revenue-share payment is not a flat fee. The partner receives an agreed proportion of the search advertising revenue generated by the queries it sends, which means the payment scales with how much the placement is worth and aligns the partner's interest with keeping the default where it is. Alphabet reports the aggregate of these payments in its financial statements as traffic acquisition costs, a single line covering payments to distribution partners and to Google Network members, with no breakdown by partner. That is why the court record is the better source: the accounts show the total, the litigation shows what the total buys.

Two limits carry with every figure on this page from the case. The amounts are of 2020 and 2021 vintage, and the judgment is on appeal.

What the money buys, channel by channel

The Department of Justice Antitrust Division's closing argument presentation in U.S. and Plaintiff States v. Google LLC, a trial demonstrative filed in 2024, decomposes the placement into its three distribution channels. Google's default search contracts covered roughly 50 percent of all United States general search queries in 2020, split as follows:

The Apple Internet Services Agreement — 28 percent of all US general search queries. Safari on iPhone, iPad and Mac, plus Spotlight and Siri web results, all resolving to Google unless the user changes the setting.
Android manufacturer and carrier agreements — 19.4 percent. The contracts with handset makers and mobile network operators that decide what a new Android device searches with before anyone touches it.
Third-party browser agreements — 2.3 percent. Firefox, Opera and the rest. The smallest channel by query volume, and by a wide margin the most consequential to the companies on the receiving end.

In every one of those settings the alternatives are present and one tap away. A browser's search settings list Bing, DuckDuckGo, Ecosia and Yahoo among the engines a user can select; what the contracts buy is not exclusivity but the position nobody has to select.

The per-channel splits come from Google's own internal query data produced in discovery, which is the only reason they exist at this precision. No panel, tracker or survey can see which queries arrived through which contract; the distinction is invisible from outside the company.

Two things follow that are worth stating plainly. First, the Apple agreement is not one deal among several — it accounts for more than half of all contracted query volume on its own, and its 28 percent exceeds the entire Android channel by nine percentage points. Second, the browser channel's 2.3 percent looks negligible in query terms and is nothing of the kind in financial terms, because the recipients are small organizations for which a search revenue share is most or all of their income. The Mozilla accounts discussed further down this page are the clearest public illustration.

The 2020 data vintage matters here too. These shares describe a market before generative answer products appeared on every major search surface, and before the remedies phase of the case, whose September 2025 opinion addressed the future of these arrangements. The liability ruling is on appeal.

What the channel split does not tell you is how much of the $26 billion went down each pipe. The court's finding is an aggregate; the demonstrative's decomposition is of queries, not dollars. Placing the two side by side gives a strong indication of where the money concentrated, but an indication is what it is.

The Apple question, and why most answers to it are unsourced

"How much does Google pay Apple to be the default search engine?" is among the most-asked questions about search, and almost every page answering it cites a news recap of a news recap. It is worth separating what the record establishes from what circulates.

What the record establishes. The court found the aggregate: more than $26 billion in revenue-share payments for default placement in 2021, across all partners. The Department of Justice demonstrative establishes what the Apple agreement delivers in return: 28 percent of all US general search queries in 2020. Those two findings rest on Google's internal data, were tested at trial, and are citable.

What circulates. Per-partner dollar figures for the Apple arrangement are widely reported, generally traced to testimony, exhibits or unsealed passages surfacing during the litigation and then repeated at second and third hand. The commercial terms of the Internet Services Agreement were treated as confidential by both companies, and the amounts that reached the public did so in fragments, for particular years, sometimes in redacted or approximate form. A specific dollar amount for a specific year attributed to Apple may well be accurate; it does not have the provenance of the $26 billion aggregate, and a page that presents it as though it did is overstating what it knows.

The practical consequence for a reader is a small discipline. When you encounter a figure for what Google pays Apple, check whether the source is the filing or an article about an article. If the chain ends at a news site quoting an analyst, treat the number as an estimate, and note the year it purports to describe — these are annual revenue shares, so a 2021 figure and a 2022 figure are different quantities and are routinely quoted interchangeably.

The aggregate is in any case the more useful number for understanding the market. Whether one partner received $15 billion or $20 billion in a given year changes nothing structural. That total payments for placement exceeded $26 billion, and ran to nearly four times everything else Google spent specifically on search, tells you what the competitive constraint actually is: not the cost of building a better engine, but the cost of being the one the device asks.

What a default is worth to the company receiving it

The clearest public measurement of the other side of this transaction is not in the court record. It is in an audited financial statement.

The Mozilla Foundation and Subsidiaries Consolidated Financial Statements for the years ended December 31, 2024 and 2023 report that approximately 86 percent of Mozilla's revenues from customers with contracts were derived from a single customer in 2024, up from 85 percent in 2023, against total royalty revenue of $498.2 million. Royalty revenue is what Mozilla earns from search arrangements in Firefox.

State the limitation with the figure, because it is the whole reason the figure is interesting. The statements do not name the customer, and they do not disclose the expiry date of any search agreement. The concentration note discloses a dependency, not a counterparty. Mozilla's search relationship with Google is a matter of public record elsewhere and has been for many years, but the 86 percent figure itself is a statement about revenue concentration, and it should be quoted as one rather than converted into an assertion the auditors did not make.

Even stated that carefully, it is the most revealing number available about what a browser default is worth. An organization that ships one of the four browsers most people have heard of, that maintains an independent rendering engine, that employs several hundred people and funds a public-interest foundation, derives roughly six dollars in seven of its contracted revenue from one commercial relationship built on which engine the address bar queries.

Set that against the Department of Justice's channel split and the asymmetry is complete. Third-party browser agreements account for 2.3 percent of US general search queries — a rounding error to the payer, a rounding error's worth of traffic in a market of hundreds of billions of queries. To the recipient it is existential. That asymmetry is the mechanism by which a payment that is cheap for Google becomes decisive for everyone else's independence, and it applies well beyond Mozilla: most organizations capable of maintaining a browser or an alternative engine are orders of magnitude smaller than the company writing the checks.

The figures are for the financial years 2023 and 2024 as audited. Mozilla publishes these statements annually, and the concentration percentage moves year to year.

What the payer gets: Bing on Microsoft's own browser

None of this spending would happen if defaults did not work. The court record contains the cleanest available measurement of how well they work, and it arrives as an accident rather than an experiment.

The district court's 2024 memorandum opinion records that on Microsoft Edge — the browser where Microsoft sets the default — Bing's search share is approximately 80 percent and Google's approximately 20 percent. That is the mirror image of the ratio in every other setting. The same opinion's findings put Bing at 5.5 percent of all US general search queries in 2020, against nearly 90 percent for Google.

A natural experiment is a situation where something outside a researcher's control assigns people to different conditions, producing a comparison that would otherwise require a controlled trial. Edge is one. Its users are not a distinct population with unusual views about search quality; most did not choose the browser at all, since it ships with Windows. Flip the single variable of who sets the default and an engine at roughly a twentieth of national query share handles four searches in five.

That is the return on the $26 billion, expressed as a multiple. Whatever difference in quality exists between the engines, it is not large enough to make most people override a default that costs a few taps to change — and this site documents where that setting lives in every major browser. Defaults are therefore not a convenience; they are the distribution.

The figure is a US finding in a 2024 opinion describing earlier data, on a judgment under appeal. It also proves less than it is often made to prove: it establishes that defaults move share by a very large factor, not that the engines are of equal quality. Evidence bearing on the second question is a different matter and is set out on the page about why Google dominates search, where a randomized field experiment found that a third of people paid to try Bing for two weeks were still using it once the payments stopped.

The regulatory consequence: the regulator that declined to designate Bing

The distribution spending has a second-order effect that is easy to miss: it makes the scale gap between Google and its nearest competitor so wide that competition regulators reason about them as different kinds of object.

In 2024 the European Commission's Directorate-General for Competition decided Case DMA.100015, Microsoft / Online Search Engines, on whether Bing should be designated under Article 3(2) of the Digital Markets Act as a core platform service — a service large enough that its operator is subject to the Act's obligations as a gatekeeper. Bing crossed the Act's quantitative user thresholds. The Commission nonetheless declined to designate it, on the basis that it does not constitute an important gateway for business users to reach end users.

The reasoning rests on figures Microsoft itself submitted, which is what makes the decision unusual. It records that Bing's share of clicks across all devices in the European Union was approximately 3.3 percent for 2020 to 2022, and quotes Microsoft's own submission that Bing is "25 times smaller than Google Search, which has a share of 92%." A company arguing that it is too small to be regulated is an unreliable narrator by construction; that the Commission accepted the argument, and that the figures appear in a formal decision, is what gives them weight.

The consequence is worth holding still for a moment. The Digital Markets Act exists in part to constrain the search market. Applied to that market, it produced a finding that the second-largest general search engine in Europe is not an important gateway at all. Whatever one makes of the outcome, it is an official measurement of the distance between first and second place, and that distance is what default payments at the scale of $26 billion a year sustain.

Two limits on the figures. The data period is 2020 to 2022, so they do not describe the current market. And some of the figures in the decision appear in confidentiality bands rather than as point estimates, which is how the Commission publishes commercially sensitive data; bands should be cited as bands.

Reading these figures without overstating them

Four rules cover most of the mistakes made with the numbers on this page.

Quote the aggregate, not an inferred per-partner amount. More than $26 billion in revenue-share payments for default placement in 2021 is a court finding. A dollar figure attached to one named partner in one year is usually a press estimate or an unsealed fragment, and it should be labeled as one.
Keep the years attached. The placement payments are 2021. The operating cost of the search engine, $8.4 billion, is 2020. The query channel splits are 2020. Mozilla's 86 percent concentration is the 2024 financial year. The European Commission's click share covers 2020 to 2022. None of these describes 2026, and combining them into a single present-tense sentence produces a claim no source supports.
Say that the judgment is on appeal. The liability ruling in United States v. Google LLC has been appealed, and the remedies opinion of September 2025 addressed these arrangements further. The underlying query and payment data were drawn from Google's own records and were not seriously contested, but the legal conclusions built on them are not final.
Do not convert a revenue share into a subsidy. These are payments for a proportion of advertising revenue generated by queries the partner sends, not grants. The distinction matters when reasoning about what would happen if an agreement ended: the recipient loses a revenue stream tied to traffic it still has, not a gift.

The figures on this page were checked against the source documents in September 2026. The Mozilla statements are published annually and the concentration percentage moves; the court and Commission findings are fixed documents that do not update, and their age is part of what they say.

Frequently asked questions

How much does Google pay to be the default search engine?

More than $26 billion in revenue-share payments for default placement in 2021, according to the memorandum opinion of the US District Court for the District of Columbia in United States of America v. Google LLC. The court also found that this was nearly four times more than all of Google's other search-specific costs combined, and the same record puts the cost of operating the search engine itself at $8.4 billion in 2020. That total was sealed until the opinion published in 2024, and the judgment is on appeal.

How much does Google pay Apple specifically?

The court record establishes the aggregate across all partners rather than a per-partner amount, so no figure for Apple alone carries the same provenance as the $26 billion total. What the record does establish about the Apple arrangement is what it delivers: the Department of Justice's trial demonstrative records that the Apple Internet Services Agreement covered 28 percent of all US general search queries in 2020. Specific dollar amounts attributed to Apple circulate widely, generally traced to testimony or exhibits surfaced during the litigation and then repeated at second hand. Treat them as estimates and check which year they describe.

What does Google get for those payments?

Roughly half of all United States general search queries in 2020, according to the Department of Justice's closing argument demonstrative: 28 percent through the Apple Internet Services Agreement, 19.4 percent through Android manufacturer and carrier agreements, and 2.3 percent through third-party browser agreements. How much a default is worth in practice is visible in the court's finding that Bing holds approximately 80 percent search share on Microsoft Edge, where Microsoft sets the default, against approximately 20 percent for Google.

Does Google pay Mozilla for Firefox's default search engine?

Mozilla's audited accounts disclose the dependency without naming the counterparty. The Mozilla Foundation and Subsidiaries Consolidated Financial Statements for 2024 report that approximately 86 percent of Mozilla's revenues from customers with contracts came from a single customer, up from 85 percent in 2023, against total royalty revenue of $498.2 million. The statements do not name that customer and do not disclose any agreement's expiry date. The search relationship between Mozilla and Google is documented elsewhere, but the 86 percent figure itself is a revenue-concentration disclosure and should be quoted as one.

Why does Google spend more on placement than on running the search engine?

Because placement is where the competition happens. The court record puts default revenue-share payments above $26 billion in 2021 against $8.4 billion to operate the engine in 2020, and the Microsoft Edge comparison in the same opinion shows why: flipping which engine a browser queries by default moves share by a factor of roughly four, in the same country, among the same population. An engine that is not the default does not get the queries, and an engine that does not get the queries does not get the advertising revenue or the behavioral data.

What are traffic acquisition costs?

Traffic acquisition costs, or TAC, is the line in Alphabet's financial statements covering payments to distribution partners and to Google Network members for the traffic they send. It is reported as an aggregate, with no breakdown by partner or by channel, which is why the court record rather than the accounts is the better source for what any particular arrangement involves. The accounts show the size of the total; the litigation shows what the total buys.

Why did the European Commission decide Bing was not a gatekeeper?

In Case DMA.100015 the Commission found that although Bing crossed the Digital Markets Act's quantitative user thresholds, it does not constitute an important gateway for business users to reach end users, and so declined to designate it under Article 3(2). The decision records Bing's share of clicks across all devices in the European Union at approximately 3.3 percent for 2020 to 2022, and quotes Microsoft's own submission describing Bing as 25 times smaller than Google Search, which has a share of 92 percent. Some figures in the decision appear in confidentiality bands, and the data period ends in 2022.

What default search placement costs Google, on the record