← all stories
1998–2002acquired by eBay for $1.5B · 2002

PayPal

Confinity → X.com → PayPal

An online payment service that ran on email, grown out of a company that had tried to sell encryption for the Palm Pilot handheld. Between 1998 and 2002 it went from beaming money between PDAs to the standard way to pay on eBay, survived a merger with the rival bank X.com run by Elon Musk and a war with eBay's own payment service (Billpoint) — and was bought by eBay for $1.5 billion six months after its IPO.

Founders Max Levchin · Peter Thiel · Luke Nosek · Ken Howery · Elon Musk (X.com)
Domains paypal.com · x.com · confinity.com (historical name)
paymentsmarketplaceviralreferralfraud-preventionfintechp2p-payments

In brief

PayPal did not begin as a payment service but as a company encrypting data on Palm Pilot handheld computers, an idea nobody needed. The real product was found almost by accident: a fallback feature for sending money by email instead of passing it over a PDA's infrared port turned out to be exactly what sellers on eBay auctions needed, tired as they were of waiting weeks for paper checks. In four years the company went from a garage startup to an IPO and was bought by eBay itself for $1.5 billion — even though for two years before that eBay had been trying to kill PayPal with its own payment service, Billpoint, and lost.

How it started (the founders)

Max Levchin, a 23-year-old engineer who had emigrated with his family from Kyiv to Chicago in 1991 on political asylum, had managed during his years at the University of Illinois to launch four startups: three failed, one (ListBot/PositionAgent) was bought by LinkExchange, and with that money Levchin moved to Silicon Valley in August 1998. There he got in touch with Peter Thiel, a 30-year-old hedge fund manager and lawyer by training, and in September 1998 the two of them, together with Luke Nosek and Ken Howery, founded Fieldlink: security software for the Palm Pilot that turned the handheld into a "digital wallet" with encryption estimate. The pitch to Palm CEO Jeff Hawkins failed: Palm already had a security partner, Certicom, and the corporate market for mobile security turned out to be too early estimate.

In December 1998 the company was formally incorporated under the name Confinity, Inc. (from "confidence" + "infinity") fact and turned toward consumers: first mobile wallets for buying things with a Palm, then the "beam money" idea (send money to a friend over the infrared port to split a restaurant check) estimate. In February 1999 a seed round of ~$500K closed, after which the company locked in the consumer pivot estimate. The original business plan aimed at the market for small non-card C2C payments: "$100 billion" a year by the 1999 estimate, with a target of 1 million users and $3–4 million in revenue by the end of 2000. The demo at the financial cryptography conference in 1999 got a cool reception (a year earlier DigiCash, a direct competitor in digital money, had gone bankrupt), but at the public launch a Nokia Ventures representative beamed Thiel a symbolic $4.5 million over the infrared port, and Nokia Ventures did in fact put that money into the company estimate. Board member Reid Hoffman openly doubted the practicality of beaming at the time: by his estimate, there was less than one Palm user per restaurant-dinner cycle estimate.

The feature for sending money by email was conceived by Levchin as a fallback in case a user had left the PDA at home, but it was the one that outlived all the other pivots: David Sacks, hired as product lead, convinced Thiel while still in his job interview to make it the priority (Palm had only 5 million users, while email had potentially anyone with a mailbox) estimate. Levchin himself later counted six consecutive turns of the idea before the company found money in email payments for eBay sellers estimate. PayPal's first recorded transaction was October 16, 1999. In parallel and independently, in March 1999, Elon Musk, who had just sold Zip2, incorporated X.com — an ambitious online financial supermarket bringing transfers, investments, loans, and insurance together in one place: "I do not fit the picture of a banker," he said of himself that same year estimate.

Year-by-year timeline

Lesser-known but significant facts

  1. The "Pay with PayPal" button was invented not by the company but by an eBay seller. In November 1999 a support employee forwarded to David Sacks (who was temporarily covering legal matters at the time) a message from a PowerSeller who had made a button with the PayPal logo for her own auctions and was asking permission to use it. The team searched eBay for "paypal" and found hundreds of listings where the service was already mentioned as a payment method, even though there was no formal integration at all.
  2. The main competitor belonged to eBay itself — and eBay lost to it on its own turf. Billpoint, a joint venture of eBay and Wells Fargo, was promoted on every eBay page, yet sellers and buyers still preferred PayPal, and Billpoint was costing eBay losses on the order of $10–15 million a year estimate; in the end eBay made PayPal at least three lowball acquisition offers ($300M, then $500M, then $800M) before agreeing to $1.5 billion estimate.
  3. Elon Musk was removed from the CEO post while he was on his honeymoon. In September 2000 Musk flew with his wife Justine to the Olympics in Sydney; in his absence the board of directors (with Thiel taking part) held a vote on replacing the CEO, and by the time Musk returned the decision could no longer be reversed; in October 2000 Thiel took the post.
  4. In its IPO filing the company itself warned investors that online casinos were paying through it. In the 2001 S-1 PayPal disclosed the risk of civil or criminal prosecution for serving potentially illegal gambling, the risk of being deemed an unauthorized bank in certain states, and a fine it had already received from MasterCard for excessive chargebacks — regulatory self-criticism unusually candid for a company going public six months later.
  5. The Igor anti-fraud system was born of a willingness to knowingly lose money in order to study the fraudsters. By Levchin's own account, he let part of the fraud go through in order to understand the attackers' methods and train Igor — a system built on early machine learning that, working alongside a team of ~100 fraud analysts, cut losses from more than 1% to 0.5% of payment volume.

The first growth lever

The first tangible growth came not from a single mechanic but from a bundle of three parts. At the core was the two-sided referral bonus devised by Luke Nosek: $10 to a new user for signing up and another $10 to the person who invited them once the invitee had registered. It was this mechanic that produced the viral coefficient behind growth of ~7–10% a day at the end of 1999: 1,000 users in mid-November became 12,000 by the end of December. But paying for growth into a void would have been expensive and pointless. The second part of the bundle supplied the channel: eBay sellers themselves turned the PayPal logo into a clickable button on their auctions, and the company merely legitimized and refined what the market was already doing on its own. By April 2000 PayPal was accepted on 20% of eBay auctions, by the end of June on 40%, and by 2001 on more than 70% estimate.

The economics were not free: by Thiel's own estimate, the referral program consumed $60–70 million over the first years, which produced 5–6 million active payers estimate; banks were spending an order of magnitude more to acquire a customer estimate. As the organic channel gathered pace, the bonus was reduced from $10/$10 to $5/$5, and verification barriers were added later. Corrected during verification: individual records gave $20/$20 as the starting figure, but the direct quote from the company's COO and an independent retelling of Luke Nosek's interview converge on $10/$10 as the original amount; the $20 in those records is probably a confusion with the total cost of acquiring a single user ($10+$10), which Sacks himself gave the same way in another interview. The third part of the bundle was direct competition with eBay's own payment service, Billpoint: despite Billpoint being promoted on the eBay site itself, by October 2000 five times as many sellers were using PayPal estimate.

Parallels today (projects from the catalog)

What a builder can take from this in 2026

  1. Look not for demand but for an already warmed-up market, and embed yourself in it. PayPal did not create the need to pay for goods online; it already existed on eBay in the form of agonizing paper checks. All of the company's successful pivots were moves toward an already boiling market, not the invention of a new one.
  2. Virality has to be paid for and counted as unit economics, not waited on in the hope that it happens for free. The $60–70 million spent on referral bonuses was not charity but a calculated acquisition cost, which was lowered as organic growth built up ($10/$10→$5/$5). The question is not "is this viral or not" but "how much does one active payer cost."
  3. The product that takes off rarely matches the one the first round was raised for. Out of six ideas (from corporate PDA security to email payments), the money was raised for cryptography on handhelds, and what worked was a feature conceived as a secondary fallback.
  4. Trust and security are a product and a PR asset too, not just a cost line. The investment in anti-fraud (Igor, CAPTCHA) became a public story about the company's security: it appeared in the press (Newsweek, MIT Technology Review) alongside advertising for the service.
  5. Honesty about regulatory risk at the IPO does not kill the deal. PayPal wrote plainly in its S-1 that illegal casinos were using its product and that it might be deemed an unlicensed bank — and it still went public and rose 55% on the first day of trading.

Discrepancies and what we could not verify

Updated by a second pass (verification, 2026-09-06): what changed is marked with a ✅/⚠️ at the start of the line.

Sources (primary first)

Primary (period documents and the founders' own words):

Secondary (analysis and retellings, some citing Jimmy Soni's book "The Founders," 2022, 260+ interviews):

The same thing, about today

The same breakdowns, but of projects launching right now: what the product is, where the first users came from, how they charge. The card is free, the full dossier is $5 (the dossier itself is written in Russian).

Browse the dossier catalog →