← all stories

After the Web · Era 4 · Dot-coms and the crash

2003–2011bought by eBay for ~$2.5–2.6 billion (2005), 65% sold…

Skype

A program for free calls over the internet, launched in 2003 by the creators of the file-sharing network Kazaa. Calls between users went directly between computers, without expensive servers, and the money came from paid calls to ordinary phones. eBay bought Skype in 2005 for about $2.5–2.6 billion, wrote off $1.4 billion, sold most of it to investors in 2009, and in 2011 Microsoft bought the whole company for $8.5 billion.

Founders Niklas Zennström · Janus Friis
Domains skype.com
p2pfreemiumvoipacquisitionlicensing

In brief

Skype is a program for free calls over the internet, released in 2003 by the people who had created the file-sharing network Kazaa. Calls went directly between users' computers, so the service cost almost nothing to run, and the revenue came from paid calls to ordinary phones. Over eight years the company was sold three times: eBay bought it in 2005 for about $2.5–2.6 billion, two years later wrote off $1.4 billion, in 2009 sold most of it to investors at a valuation of $2.75 billion, and in 2011 Microsoft paid $8.5 billion for it. The story is useful because it is all about money: a free product, a paid add-on, someone else's license on the core, and the price of a buyer's mistake.

How it started (the founders)

The founders were the Swede Niklas Zennström and the Dane Janus Friis. Before Skype they had built Kazaa, a file-sharing network that by 2003 had already been sold to Sharman Networks; according to a press account, they first discussed the idea for the new project in a park in Copenhagen. The idea was simple: the same P2P technology, but voice instead of file sharing. The software core was written by the Estonian developers Ahti Heinla, Priit Kasesalu, and Jaan Tallinn; one retelling adds Toivo Annus to them, who is credited with the idea of making calls cheaper through P2P estimate.

The company was registered in Luxembourg, the marketing office was in London, and the main development team was in Tallinn estimate. An important detail that is rarely mentioned: Skype did not own the software foundation of the P2P network, the "Global Index." It was licensed from Joltid, according to press retellings a company of the same Zennström and Friis estimate, in connection with the founding of Skype, as the S-1 puts it. This detail came back to bite six years later.

The public beta came out, according to Wikipedia, on 2003-08-29 estimate. A document of the era is more reliable: the skype.com home page of 2003-10-05 describes Skype as the next phenomenon from the people who brought you KaZaA and shows a counter of 992,794 downloads and counting.

Year-by-year timeline

Lesser-known but significant facts

  1. The network's core did not belong to Skype. Until November 2009, Skype only licensed the "Global Index," on which the P2P network ran, from Joltid, the founders' company. In September 2009, in the middle of the sale to investors, Joltid sued, seeking to bar use of the technology and claiming damages that it put at more than $75 million a day: the service itself was called into question. The dispute was closed with money and a stake.
  2. The earn-out was closed at about a third of the cap. In 2005 the earn-out cap was about €1.2 billion; in 2007 the parties agreed on €375 million (~$530 million). The earn-out conditions (active users, gross profit, revenue) were not being met on schedule estimate.
  3. The 2003 site sold itself on Kazaa's reputation. The first line of the home page called Skype the next phenomenon from the people who brought you KaZaA. Skype itself pointed to Kazaa as proof that P2P works.
  4. eBay's forecast almost came true on revenue, and it did not help. In 2005 eBay publicly expected $60 million of revenue from Skype in 2005 and more than $200 million in 2006, against $7 million in 2004; per the S-1 the result was about $72 million for 2005 and $193.7 million for 2006, that is, the first target was beaten and the second almost reached. The $1.4 billion write-off came in 2007, when revenue was already $382 million: the problem was the price, not the growth estimate.
  5. Skype was not entirely "serverless." A 2004 technical analysis (Columbia University): the central element was the login server, and everything else was a decentralized network of ordinary computers and supernodes.

Legend vs. the record

1. Skype was made by Swedes. The press of 2003 calls both founders Swedes. Reference sources show otherwise: Zennström is Swedish, Friis is Danish; the code was written by Estonians; the company was registered in Luxembourg, and development was in Tallinn. Verdict: the legend is inaccurate: the founders were Scandinavians, the product was Estonian, and the legal entity was Luxembourgish estimate.

2. eBay bought Skype and lost money on it. On the books there were losses: a $1.4 billion write-off in 2007 and $343 million on Joltid in 2009. In cash flows the picture is different: eBay paid about $1.3 billion in cash, roughly $1.2 billion in stock (the difference between $2.5 billion and $1.3 billion), and $530 million for the earn-out, and it received $1.9 billion in 2009 and $2.3 billion in 2011; eBay itself spoke of a total return of $1.4 billion on the investment. Verdict: the legend is exaggerated: eBay got its money back with a profit but missed the growth, since Microsoft paid $8.5 billion for a company that had been valued at $2.75 billion when eBay sold its stake twenty months earlier estimate.

3. Skype's core was its own. Per the S-1, no: the core was licensed from Joltid in connection with the founding, and Skype got the rights to it only in 2009. Verdict: the legend is false: without that deal the sale to investors would not have happened in that form estimate.

4. Skype was free, and nobody made money on it. Calls between users were free, but the main revenue came from SkypeOut, prepaid minutes for calls to ordinary phones; in 2010 there were 8.1 million paying users against 124 million online (about 6.5%, my calculation). Verdict: the legend is false: what was free was the core, and what was paid was the part with a real cost (access to the operators' networks).

The first growth lever

The lever was the absence of costs for each new user. P2P (calls go through users' computers rather than through the company's servers) made it possible, according to Technology Review, to do without the cost of a centralized server and to give the core service away for free. In numbers: 2.6 million downloads in two and a half months against more than six months for Kazaa, about 993,000 by 2003-10-05, and almost 8 million registered users by the summer of 2004. Two cheap tricks were added to this: the reference to the creators of KaZaA on the home page and eight languages as early as the third week after launch.

The money came from a separate product: SkypeOut, summer 2004, prepaid minutes for calls to ordinary phones. Revenue was about $7 million in 2004, $382 million by 2007, and $719 million by 2009. What made people pay was not a paywall on the core but physics: a call to an ordinary number has a real cost.

The paired story

The pair is Kazaa (history-kazaa): the file-sharing network of the same Zennström and Friis, launched in 2001; the code of both products was written by the same Estonians, Jaan Tallinn, Heinla, and Kasesalu estimate.

Parallels today (projects from the catalog)

What a builder can take from this in 2026

  1. Make free what costs you nothing, and charge for what you pay for yourself. A call between users cost almost nothing, a call to a phone meant real money to an operator, and that is exactly what was sold on prepayment. Find a per-minute or per-unit item in your costs and make that item the paid one.
  2. Prepayment is working capital. The S-1 says it plainly: users pay in advance. If you have a usage-based plan, sell credits up front.
  3. Before scaling, check who owns the core. Skype lived for six years on a license from the founders' company, and in the middle of the sale the licensor went to court demanding that the technology be barred; settling it cost $343 million and ~10% of the equity. If you depend on someone else's code, model, or API, put the rights, the term, and the termination conditions in writing now.
  4. The reputation of a previous project is a lever. Telling visitors on the home page that Skype came from the creators of KaZaA gave the technology credibility from day one. If you have a well-known past product, name it.
  5. An earn-out shifts the risk but does not cancel a mistake in the valuation. eBay paid about a third of the earn-out cap and still wrote off $1.4 billion. Tie the payout to metrics that both sides can measure.

Discrepancies and what we could not verify

Sources (primary first)

Internet History Keeper

Every historical dossier here is free and open, and it will stay that way. If you want the series to keep going (new dossiers, checks against primary sources, the English edition), support it with a subscription for $9.90 a month. It is support, not access, and you can cancel at any time.

Become a keeper — $9.90 a month →

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 →