Business Overview

PayPal Holdings, Inc. (PYPL)

17 September 2026 · Built from PayPal's 10-Ks FY2016–FY2025, 10-Qs, 8-Ks and call transcripts through Q2 2026, plus independent and competitor sources for industry structure · Not a valuation and not a recommendation.
PayPal is a trusted two-sided payments network that earns a spread on every dollar consumers pay merchants. How much it keeps depends on whether that dollar is funded by a card, which it must pay banks for, or by a balance or bank account, which costs almost nothing.
The engine · one dollar of payment volume (FY2025)

On $1.79tn of TPV, PayPal booked $33.2bn of revenue: about 1.85¢ per dollar. About 0.89¢ went to card issuers, networks and processors, and 0.10¢ to fraud and credit losses. ~0.86¢ per dollar, or ~$15.5bn, is transaction margin, which pays for a ~$9.4bn fixed cost base and leaves $6.1bn of operating income.

Merchant fee 1.85¢− funding cost 0.89¢− losses 0.10¢= margin 0.86¢− opex ~0.53¢= op. income ~0.34¢

Ratios computed from FY2025 10-K figures; TPV disclosed rounded.

~1.85%
Unit price: revenue per $ of TPV, FY2025 (3.01% in FY2016)
~0.86%
Unit profit: transaction margin per $ of TPV (1.78% in FY2016)
$6.07bn
GAAP operating income FY2025 · 18.3% margin · EPS $5.41
$13.4bn / $13.5bn
Debt vs cash & investments, Q2 2026 (10-Q)
+2%
Cycle signal: online branded checkout TPV growth, Q2 2026 (FXN)
+5% / +3%
Q2 2026 revenue growth reported / currency-neutral; organic ≈ reported (no material M&A since 2021)

1. Snapshot

BusinessTwo-sided payments network: 439m active accounts using the PayPal button, Venmo and the unbranded Braintree processor.
Makes money byCharging merchants a percentage-plus-fixed fee, paying card funding costs, absorbing losses, and keeping the spread; plus interest on balances and credit income.
Protected byStored credentials and buyer trust, fraud data from $1.79tn of volume, licences, $41.7bn of customer balances.
Earnings driversBranded checkout volume (over half of profit), funding mix, Braintree pricing, Venmo and credit monetization.
Cycle exposureMedium: consumer spending, credit losses and interest rates move against it together.

2. What the company does

An online merchant must accept payment from a stranger without being defrauded, and a consumer wants to pay without typing card details into an unfamiliar site. PayPal holds the consumer's funding sources, authenticates the buyer, protects both sides and moves the money. In FY2025 that meant $1.79tn of TPV across 25.4bn transactions (FY2025 10-K).

Trace a $100 checkout. The merchant is charged a fee with a fixed component and a percentage component. If the consumer pays with a stored credit card, PayPal pays interchange and network fees out of that fee, so most of it passes to the banks. If the consumer pays from a PayPal balance or bank account, PayPal describes the funding cost as "nominal", and nearly all of the fee is margin. Funding mix is the single largest determinant of what a dollar earns.

Transaction revenues were $29.8bn (90%) in FY2025; other value added services were $3.4bn (10%): loan interest and fees, partner-bank revenue share, referral fees, and interest on customer balances. That interest is not disclosed in dollars; 2026 guidance implies about $1.1bn (inferred).

The three engines earn very different amounts. Branded checkout is "over half our profit dollars". Braintree unbranded processing is almost entirely card-funded and carries "higher expense rates". Venmo and consumer financial services monetize a largely free P2P base through instant transfers, debit, Pay with Venmo and BNPL: Venmo revenue was ~$1.7bn in FY2025, and financial services are "close to 20%" of transaction margin.

PayPal has been exiting low-return activity. It pruned unprofitable Braintree volume, so FY2025 transactions fell 4% while TPV rose 7%. It sold its US credit portfolio to Synchrony in 2018 and now sells BNPL receivables forward: up to €65bn in Europe through March 2028, plus up to $7.0bn in the US. It keeps origination and servicing, not the balance sheet.

3. Industry, competitive position & moat

US merchants paid $187.2bn to accept $11.9tn of card volume in 2024, an average of 1.57% (Nilson Report). Most goes to issuers as interchange. The best slice of the profit pool belongs to the networks: Visa and Mastercard ran ~60% and 57.6% GAAP operating margins in FY2025 on roughly 28–31bp of volume, with no funding or credit risk. Pure processors earn far less. Adyen took ~17bp of volume in FY2025 at a 53% EBITDA margin, slipping to ~16.2bp and 49% in H1 2026.

PayPal's ~86bp of margin per dollar is far above a processor because branded checkout commands a premium. But a card-funded wallet is also the card acceptor, so it pays the bank's toll out of its own price.

Wallets are winning the category: they were 56% of global e-commerce value in 2025, up from 50% in 2023, and 40% in the US (Worldpay Global Payments Report). Which wallet wins is the open question, and no primary source publishes brand shares for PayPal against Apple Pay or Shop Pay. The market is national or regional: UPI, Pix and Europe's Wero (43.5m users) dominate their home markets.

Competitors by type.

What is hard to reproduce: 439m accounts with stored credentials and buyer protection; fraud outcomes on $1.79tn a year, holding the transaction loss rate to 0.07%; and $41.7bn of balances with global licensing. What it does not own: the phone, the merchant's checkout page, or the rails. Management conceded "more competitive intensity" in Europe and "competition from alternative payment methods" in Germany.

Winners in checkout tend to carry no funding cost, own the device, or own the merchant platform (inferred). PayPal is none of these. Its route toward that structure is cheaper funding through balances, bank, Venmo and debit.

4. Growth engine

From FY2016 to FY2025, TPV grew ~5.0× while revenue grew 3.1×, because revenue per dollar fell from 3.01% to ~1.85%. The filings blame mix. First came P2P and Venmo, then the end of high-rate eBay volume after July 2020, then low-rate Braintree processing. PayPal grew its lowest-margin dollars fastest.

In Q2 2026 TPV grew 10% but revenue grew 5% (3% currency-neutral), the transaction take rate fell 7bp to 1.61%, and TM$ grew just 1% (3% ex-interest). By product, Braintree added ~$400m, Venmo ~$60m and hedging ~$80m, while PayPal products lost ~$130m, partly to co-marketing booked as contra-revenue. With no acquisitions in 2022–2024 and only $153m of deals since, reported growth is essentially organic; FX is the relevant adjustment.

Structural · currently weak
1. Branded checkout volume

Over half of profit, but growing only 1–2% FXN in Q4 2025–Q2 2026 while wallets gain share. Each point here outweighs several points of Braintree.

Structural / cyclical
2. Consumer financial services & BNPL

BNPL TPV +26% in Q2 2026; ~20% of TM, growing double digits. Capital-light via forward-flow sales, but tied to credit quality.

Management-driven
3. Venmo monetization

~$1.7bn revenue FY2025 (+~20%); Pay with Venmo TPV +44% in Q2 2026. Converts free P2P users into merchant and card revenue.

Management-driven · low margin
4. Braintree / PSP volume

Largest revenue adder, repriced to "price-to-value". Lifts revenue far more than margin, and competes on price with Adyen and Stripe.

Cyclical
5. Interest on customer balances

~$1.1bn of 2026 TM (inferred). Falls with rate cuts.

The long-run unit metric: what a dollar of TPV earns

0.50%1.25%2.00%2.75%3.50%20162017201820192020202120222023202420253.01%1.85%Revenue per $ of TPV1.78%0.86%Transaction margin per $ of TPV
Revenue ÷ TPV and (revenue − transaction expense − transaction & credit losses) ÷ TPV, computed from 10-Ks FY2016–FY2025. FY2016–17 TPV as restated in the FY2018 10-K; TPV rounded from FY2020.

5. Margin, cash & capital allocation

Costs come in two layers. The variable layer was $16.0bn of transaction expense plus $1.7bn of losses in FY2025. The largely fixed layer was ~$9.4bn of support, marketing, technology, G&A and restructuring, 28% of revenue. Transaction expense rose from 31% of revenue to 48% over the decade as card-funded processing grew. Yet GAAP operating margin hit a ten-year high of 18.3%, because the fixed layer was cut: headcount fell from 30,900 in 2021 to 23,800 in 2025.

Mix vs margin

10%21%32%44%55%201620172018201920202021202220232024202531%48%Transaction expense, % of revenue15%18%GAAP operating margin
Transaction expense ÷ revenue and GAAP operating income ÷ revenue, computed from 10-Ks FY2016–FY2025.
MetricFY2016FY2019FY2022FY2025
Net revenue ($m)10,84217,77227,51833,172
TPV ($bn)360712~1,360~1,790
Revenue / TPV3.01%2.50%~2.02%~1.85%
Transaction margin ($m, computed)6,4089,60213,77315,465
GAAP operating margin14.6%15.3%13.9%18.3%
OCF − capex ($m, computed)2,4893,3675,1075,564
FY2016 TPV restated for the FY2018 KPI redefinition; losses renamed under CECL in FY2020; FY2019 cash flow restated in the FY2020 10-K.

Capex is only ~2.6% of revenue. Operating cash flow less capex was $5.6bn in FY2025, against $6.4bn of adjusted FCF; the gap is ~$1.2bn of loans originated for sale. Forward-flow sales ($26.7bn in FY2025) stop BNPL growth from absorbing cash, and $41.7bn of customer balances act as float.

Where the cash went, ranked:

2026 guidance pairs ≥$6bn FCF with ~$6bn of buybacks, while opex rises 7–8% and a ≥$1.5bn gross savings program begins. Pending, not in the figures: the Cymbio acquisition, the PayPal Bank (Utah ILC) application, and a $120–140m H2 2026 restructuring charge.

6. Cyclicality & what to monitor

57% of revenue is US, and management calls its base "really a middle-income demographic". Cross-border volume has fallen to 12% of TPV from 22% in FY2016. In a downturn three things compound: TPV slows, credit losses rise (partner-bank revenue share can drop to zero), and rate cuts shrink ~$1.1bn of balance interest.

Cycle position: margin per dollar of TPV (~0.86%) is at a ten-year trough, while GAAP operating margin (18.3%) is at a ten-year peak, built on cost removal. Credit is turning: consumer loan net charge-offs rose to 4.8% from 3.6% a year earlier, and merchant loans to 7.6% from 6.1% (Q2 2026 10-Q). The downside case is already visible in miniature: Q2 2026 TPV +10%, TM$ +1%.

Durable

  • 439m-account two-sided network with stored credentials
  • Fraud data on $1.79tn a year (0.07% loss rate)
  • Low-cost funding: balances, bank, Venmo debit
  • Capital-light: capex ~2.6% of revenue
  • Licensing footprint across markets

Borrowed

  • Interest on $41.7bn of balances (~$1.1bn)
  • Forward-flow buyers for up to €65bn of BNPL loans
  • Headcount and SBC cuts lifting margin
  • Share count down ~7% in FY2025
  • FX hedging gains (+$80m Q2 2026)
IndicatorWhere published
Online branded checkout TPV growth (FXN)PayPal quarterly call and release
Transaction expense rate; transaction take ratePayPal 10-Q MD&A; call
TM$ growth ex-interest on customer balancesPayPal call and release
Consumer and merchant loan net charge-offsPayPal 10-Q credit quality note
Wallet share of US and global e-commerceWorldpay Global Payments Report (annual)
Adyen take rate; Shopify Payments penetrationAdyen half-year results; Shopify 10-Q

7. Risks & unknowns

"Branded checkout represents over half our profit dollars."Management, Q4 2025 earnings call. The segment is growing ~2%.

What the sources could not answer:

8. Investor takeaways