Table of Contents
Amazon reported $269.287 billion in Online Stores revenue in 2025, up from $247.029 billion in 2024. Books are included alongside other physical products and digital purchases; Amazon doesn’t disclose their share of that total. For book-market context, the House’s 2020 investigation reported Amazon’s US share as over half of print-book sales and over 80% of ebook sales. Those are historical findings, not current market-share measurements. Amazon’s 2025 filing and the House report, page 256, explain the figures.
Below, you’ll find the reported numbers, what they tell us about publishing, and how to check sales for your own books. Amazon’s company accounts, industry reports and your KDP dashboard each answer a different part of the question. Start with the market figures if you’re researching the industry, or jump to checking an individual book’s sales if you’re working on your catalog.
Amazon and the book market: the key figures
| Figure | Period | What it measures |
|---|---|---|
| Over 50% of print; over 80% of ebooks | 2020 House report | Historical US book-market shares reported by the investigation |
| $269.287 billion | 2025 | Amazon worldwide Online Stores net sales, across many product types |
| $49.619 billion | 2025 | Amazon worldwide Subscription Services net sales, including Prime and digital subscriptions |
| $12.3 billion | 2025 | Estimated US publisher revenue through online retailers across the industry |
| $2.2 billion ebooks; $2.5 billion digital audio | 2025 | Estimated US publishing revenue in these formats across the industry |
| $71 million | July 2026 | Total KDP Select author earnings displayed by Amazon |
The newer numbers come from Amazon’s filing, AAP’s 2025 annual report and Amazon’s Royalties Estimator page. The historical market shares come from the House report’s discussion of evidence gathered during its investigation, including earlier industry estimates and submissions. They help explain Amazon’s position in publishing at that time. They don’t establish its share today, and the current financial reports don’t supply a replacement percentage.
Amazon’s revenue in 2025
Amazon’s annual filing gives us two useful figures to follow: Online Stores revenue and Subscription Services revenue. Both include book-related purchases, but also much else. Read them as a picture of the businesses surrounding books, rather than a direct tally of what readers spent.
Online Stores revenue grew 9.01%
Amazon reported $269.287 billion in worldwide Online Stores net sales in 2025, compared with $247.029 billion in 2024. The category includes physical goods and transactional digital media. Here’s the year-over-year calculation using the amounts in its 2025 Form 10-K:
Online Stores growth = (2025 net sales - 2024 net sales) / 2024 net sales × 100
= ($269.287B - $247.029B) / $247.029B × 100
= approximately 9.01%That increase shows Amazon’s online retail business continued to grow. The filing doesn’t reveal whether book sales grew at the same pace. If your book’s sales increased by 5%, for example, comparing that result with 9.01% wouldn’t tell you whether you kept up with the book market: the larger figure includes purchases across many kinds of products.
For a chart or spreadsheet, call the series “Online Stores net sales, worldwide” and keep the financial year beside each amount. Someone reading the chart on its own can then understand it. To calculate growth, subtract the earlier amount from the later one and divide by the earlier amount, as shown above.
Subscription Services brought in $49.619 billion
Amazon reported $49.619 billion in worldwide Subscription Services net sales for 2025. This includes Prime membership fees and several digital subscriptions, including audiobooks and ebooks. The same annual filing doesn’t break out Kindle Unlimited or Audible revenue separately.
Adding that amount to Online Stores revenue won’t reveal book revenue. Both categories include unrelated purchases. There’s another distinction for marketplace sales: Amazon separately reports third-party seller services, including commissions and related fees. A seller’s full transaction value and the fee Amazon earns from it are different amounts.
Think about one book purchase. A reader pays the retail price, the retailer records its relevant revenue, the publisher receives money under its distribution agreement, and the author earns a royalty under theirs. All four amounts can describe the same book while answering different questions. That’s why a publisher-revenue estimate can’t simply be substituted for Amazon’s retail sales.
How big is the US publishing market?
AAP reports help put Amazon’s book business in context. They describe revenue received by US publishers; Circana’s BookScan follows retail print-book sales. Together, they help you understand the scale of publishing and how different formats are doing. Neither is a complete Amazon sales report.
US publishing revenue reached $33.4 billion in 2025
AAP estimated $33.4 billion in US publishing revenue for 2025, up 2.5% from 2024. Its annual report, released August 10, 2026, covers books and course materials across print and digital formats. Trade consumer books accounted for $21.7 billion. Ebook revenue was $2.2 billion, digital audio $2.5 billion, and revenue through online retailers $12.3 billion. See AAP’s annual results.
Industry Insights prepared the annual estimate using publisher submissions, Census information, historical AAP data and reference sources including Bowker, Circana and independent-publishing research. The $12.3 billion online-retail figure therefore describes a sales channel across the industry. It isn’t a figure for Amazon alone.
Trade, ebooks and online retail also overlap: one describes a publishing sector, another a format and another a route to the customer. An ebook sold online can belong to all three. Don’t add those figures to create a larger total. Choose the grouping that answers your question, such as format revenue when comparing ebooks with digital audio.
These totals are useful background for a publishing plan, but they don’t predict your title’s audience or earnings. A large ebook market doesn’t tell you how many readers will see your listing, buy your book or finish it. Those questions need evidence about your particular subject, audience and catalog.
Why another AAP report shows $9.8 billion
AAP’s monthly reporting series recorded $9.8 billion in Trade revenue for the full year 2025, down 0.5% from the comparable 2024 figures. It also reported $1 billion in ebook revenue, down 0.3%, and $1 billion in digital audio revenue, up 2.1%. These full-year figures appear in its December 2025 report; they aren’t sales for December alone.
This series describes approximately 1,300 participating publishers. The annual report uses a broader industry model, which explains why its totals differ. The gap between $21.7 billion and $9.8 billion isn’t a measure of Amazon sales or self-publishing sales. Subtracting the numbers would combine two different sets of information.
To follow a trend, compare the current and previous period in the same release. Participation can change, and publishers sometimes revise earlier figures. AAP’s report releases explain these comparisons. Use the published growth rate when dollar totals have been rounded: calculating from rounded billions can give a slightly different result.
BookScan tracks approximately 85% of US trade print sales
Circana describes BookScan as tracking approximately 85% of US trade print books sold through reporting retailers and other outlets. That tells you how much of the trade-print market its data covers. It isn’t Amazon’s share of that market.
The remaining coverage also won’t necessarily be distributed evenly among books. Dividing a title’s tracked sales by 0.85 assumes that the missing outlets sell that title in the same proportions as the outlets being measured. You need evidence about the title before making that adjustment. BookScan’s stated print coverage doesn’t include ebook purchases, audiobook activity or Kindle Unlimited reading.
Title counts answer another question again. Our self-publishing statistics guide explains why ISBN registrations count identifiers rather than purchases. KDP doesn’t require an ISBN for ebooks, so an ISBN-based count can’t be treated as a complete list of Kindle titles.
How many books does Amazon sell each year?
Amazon doesn’t disclose a complete annual book-unit total in its 2025 filing. That leaves no official figure here for all print books, Kindle purchases, Audible purchases and KDP titles combined. The table shows which questions you can answer from public information and which need additional records.
| Question | What the public sources support | What is missing |
|---|---|---|
| How many books does Amazon sell annually? | No complete company book-unit total in the annual filing | Units across relevant formats, stores, and transaction types |
| How much revenue does Amazon earn from books? | No books-only revenue line | The book component of its broader reporting categories |
| What is Amazon’s US print-book market share? | The sources below do not establish a current audited percentage | A matching Amazon numerator and US print-market denominator |
| What is Amazon’s US eBook market share? | The sources below do not establish a current audited percentage | A consistent definition covering purchases, subscriptions, and publisher coverage |
| How many titles does KDP publish each year? | No current annual total established by these sources | A platform-wide count with rules for titles, formats, and editions |
| How many copies did a KDP publisher sell? | Private account reports provide sales and royalty information | Access to that publisher’s account, not an estimate from its public rank |
The historical shares near the top of this guide are still useful when discussing the industry’s past. For a current forecast, use a clearly dated estimate with a method you can examine, or leave the total unknown. An unavailable figure doesn’t mean zero sales; it means the public records don’t answer that particular question.
How to check sales for an individual book
For your own KDP books, sign in to KDP Reports. Its reporting overview separates orders, KENP reads, estimated royalties, prior-month royalties and payments. Choose the book, marketplace, format and dates you want to examine. That gives you account records to work from, rather than guesses based on the public listing.
For another author’s book, you generally won’t have access to those records. The listing’s sales rank can help you compare relative activity, and a rank calculator can produce an estimate. It won’t show the author’s exact transactions. This distinction matters when you’re choosing comparable titles: an estimated sales range can guide research without becoming a claimed audited total.
Subtract refunds to find net sales
Amazon defines Net Units Sold as Units Sold minus Units Refunded in its Royalties Estimator documentation. For paid-sale analysis, select the relevant paid rows and keep free activity separate:
Net paid units = paid units sold - paid units refundedDownload the report for the title and period you need. Use the same marketplace and format filters when comparing months, and retain the refund fields. Save the export with its download date and note whether the figures are provisional. You can then repeat the calculation when the finalized report arrives.
For earnings records, Amazon directs publishers to Prior Months’ Royalties and Payments. The Royalties Estimator isn’t a payment statement. Keep “what I expect to earn” separate from “what Amazon has paid” when planning cash flow.
You can add matching unit counts across marketplaces to get worldwide sales. Adding one country’s paperback units to worldwide Kindle units, however, leaves an incomplete mixture. For money, keep currencies separate or convert them consistently and record the exchange-rate assumption before adding the amounts.
Kindle Unlimited: $71 million in July 2026 author earnings
Amazon’s Royalties Estimator page displayed $71.0 million in total KDP Select author earnings for July 2026 when checked for this article. That is a monthly author-earnings figure. It isn’t Amazon subscription revenue or a count of books sold.
Your own Kindle Unlimited reading appears in Kindle Edition Normalized Pages, or KENP. Amazon calculates the rate by marketplace and month. Its KDP Reports guidance says monthly KENP figures may change before finalization near the fifteenth of the following month.
Keep pages read and royalties together in your records, but separate from purchased copies. Dividing pages read by a book’s normalized length can give a reading-equivalent calculation; it doesn’t identify distinct people who finished the book. Likewise, multiplying July’s $71 million by twelve would create a scenario in which every month matched July, rather than an actual annual earnings total.
What an Amazon sales rank can tell you
Best Sellers Rank reflects relative activity, with more weight given to recent activity. Each format has its own Best Sellers Rank. Books category ranks can combine formats, while Kindle and Audible category ranks are format-specific. Amazon explains these differences in its Sales Ranking guidance.
If you use a rank-to-sales calculator, record the date, marketplace, format and exact rank you entered. Check what the model assumes and how it describes uncertainty. Amazon supplies no exact public rank-to-unit conversion and cautions against using rank to track sales precisely or compare books across categories.
Amazon also documents Book Trends Data, a tested feature that combines customers who purchased, read Kindle content or listened through Audible in the most recent 30 days. That can describe engagement around a book. Because it combines buying, reading and listening, it doesn’t give a paid-copy total.
A worked example: 112 net sales and $314 in royalties
Suppose one Kindle title records 120 paid units sold and 8 refunded in a finalized monthly report for one marketplace. That leaves 112 net paid units. Assume the account also shows $224 in paid-sale royalties after refunds, plus the KU activity below. These are hypothetical US-dollar figures for illustrating the calculations, not a claim about typical author earnings.
| Measure | Calculation or record | Result |
|---|---|---|
| Net paid units | 120 sold − 8 refunded | 112 units |
| Paid-sale royalty per net unit | $224 ÷ 112 | $2.00 per net paid unit |
| KU activity | 20,000 KENP read; $90 KU royalty | Pages and royalties, kept separate from sales units |
| Combined royalties | $224 paid-sale royalties + $90 KU royalties | $314 in this month, marketplace, and currency |
The $314 total tells you how much these two royalty streams produced that month. The $2.00 figure tells you how much paid-sale royalty you earned per net copy. Calculate that using $224 ÷ 112. Dividing the full $314 by 112 would mix Kindle Unlimited earnings into a calculation based only on purchased copies.
The KU row remains useful on its own: $90 ÷ 20,000 pages gives $0.0045 per KENP in this example. That’s the result of these invented inputs, rather than a quoted Amazon rate. The units and pages also stay separate; 112 purchases plus 20,000 pages doesn’t produce a meaningful book count.
Did your sales grow? Compare months and daily sales
Going from 100 to 112 net paid units is 12% monthly growth. If the earlier month had 30 days and the later month had 31, growth in sales per day is approximately 8.4%. Continue the same hypothetical example, keeping the title, marketplace, format and paid-sale definition unchanged:
Monthly total growth = (112 - 100) / 100 × 100 = 12%
Prior daily rate = 100 / 30 = approximately 3.33 units
Current daily rate = 112 / 31 = approximately 3.61 units
Daily-rate growth = ((112 / 31) / (100 / 30) - 1) × 100
= approximately 8.4%Both results are useful. The monthly total tells you how many copies sold during each reporting period. The daily rate accounts for the extra day in the later month. Use unrounded daily rates in the calculation, then round the final result for display.
Neither percentage tells you why sales changed. Advertising, a promotion, seasonality, a new edition or one unusually large order could contribute. Note relevant events and look at a longer run of comparable data before crediting a campaign. Adjusting for month length solves one comparison problem, while those other differences still need attention.
If the previous value is zero, report the change directly—for example, “0 to 12 net paid units.” Conventional percentage growth is undefined from zero. If either period remains provisional, label it and refresh the comparison when both reports have reached the same stage of finalization.
Five calculations worth keeping in your sales spreadsheet
A simple spreadsheet can answer most routine questions about your catalog. Keep columns for sold units, refunds, net units, paid-sale royalties, KENP and KU royalties. Separate currencies and make the report period clear. These calculations then follow from the numbers you already have:
- Net paid units: paid units sold minus paid units refunded.
- Sales growth: current net units minus previous net units, divided by previous net units, multiplied by 100. The earlier value must be nonzero.
- Paid-sale royalty per net copy: paid-sale royalties after relevant refunds divided by net paid units.
- Combined royalties: paid-sale royalties plus KU royalties in the same currency and period, while keeping copies and pages in separate columns.
- Published category growth: compare two periods in the same source’s reporting series, using its category name.
Market share needs a different pair of figures. For example, “Amazon US paid Kindle unit purchases in year X divided by all US paid ebook unit purchases in year X” compares like with like. Amazon worldwide revenue and US publisher receipts wouldn’t answer that question. The filing, AAP releases and BookScan coverage statement don’t provide the matching inputs needed to calculate a current Amazon book-market share.
Six easy mistakes when sharing book-sales statistics
A good number can become misleading when its description changes. For example, AAP’s $12.3 billion online-retail estimate is useful evidence about a publishing channel. Calling it “Amazon book revenue” would give it a meaning the source doesn’t support.
- Treating all retail revenue as book revenue. Amazon’s Online Stores category includes many products.
- Treating an industry channel as one retailer. AAP’s online-retail total extends across the industry.
- Confusing publisher receipts with customer spending. They represent different stages of a sale.
- Extending print coverage to every format. BookScan’s approximately 85% coverage describes US trade print.
- Calling reading activity copies sold. KENP and paid transactions count different things.
- Removing the year from an old estimate. Historical market shares remain historical when republished.
Before you use a number in a publishing plan
Open the original report and check what was measured: money, units, titles, downloads, subscribers, customers or pages. Note the geography, marketplace, format and transaction type. Then find the reporting period, including whether it covers a calendar year, part of a year, a rolling period or one month.
Read the coverage notes for participating publishers, reporting retailers, modeled estimates, excluded activity, refunds and revisions. If you’re calculating a share, both figures need to cover the same place, period, formats and kind of sale. A range can express uncertainty about an amount; it can’t repair a comparison between unrelated quantities.
If something is unclear, keep the figure in your research notes until you resolve it. Save the source or export filename, retrieval date, relevant filters, units and currency with any figure you use. Those details make it much easier to update your plan or explain a calculation to someone else.
Sources and downloadable data
The table below covers the main financial, industry and account sources. Dates in a report describe the activity it measured, which may be earlier than its publication date. The historical market-share discussion comes from the House report linked above; its supporting evidence is listed in footnote 1561.
| Source | Period and geography | Suitable use | Boundary |
|---|---|---|---|
| Amazon 2025 Form 10-K | Fiscal 2025; worldwide company categories | Online Stores and Subscription Services revenue comparisons | No books-only sales or unit total. (sec.gov) |
| AAP 2025 StatShot Annual | Calendar 2025; United States | Modeled industry revenue by sector, format, and channel | Not Amazon revenue or retailer-specific units. (publishers.org) |
| AAP December 2025 StatShot | Calendar 2025; participating publishers’ US sales | Reported net revenue and within-report comparisons | Not a complete market estimate. (publishers.org) |
| Circana BookScan methodology | US trade print; published coverage description | Understanding BookScan point-of-sale coverage | Not global or all-format coverage. (circana.com) |
| Amazon KDP Reports | Account-selected periods and marketplaces | Own-catalog orders, royalties, and KENP analysis | Not other publishers’ accounts or Amazon’s complete catalog. (kdp.amazon.com) |
Download the source data and notes as a CSV to keep the links, reporting periods and qualifications beside your spreadsheet. The account examples in this guide are worked calculations, not additional market observations. Use the public reports to understand the industry and your finalized KDP records to understand your own business.
