← all stories

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

1997–2003acquired by Vivendi Universal for $372M · 2001

MP3.com

A California site where independent musicians uploaded songs for free and listeners downloaded them. In 1999 the company went public, in 2000 it lost a lawsuit to the largest labels over a service that stored copies of other people's CDs, and in 2001 it was bought by Vivendi Universal for $372 million.

Founders Michael Robertson · Greg Flores [unverified role]
Domains mp3.com
musicipocopyrightplatform

In brief

MP3.com was a California site where independent musicians uploaded songs for free, and listeners downloaded and played them. The company went public in July 1999 and raised about $346 million before fees, and in 2000 it was sued by the largest labels: its service for putting your CDs online stored copies of other people's discs without a license. In 2000 a $170 million reserve went to lawsuits and copyright disputes, against revenue of $80.1 million, and in 2001 the company was bought by Vivendi Universal for $372 million. It is a story about how a platform that hosts content from its own creators differs from a platform that copies other people's content itself.

How it started (the founders)

The founder of MP3.com was Michael Robertson. A company filing records that from September 1995 to March 1998 he "operated several web sites that focused on merging search technologies with commerce," that is, before music he ran sites at the intersection of search and commerce. Retellings name Greg Flores as a co-founder and give the site's launch date as December 17, 1997; the primary documents do not confirm that date: in the IPO prospectus the company "was incorporated in March 1998," and in the 10-K Robertson is listed as CEO "since March 1998." A reasonable reading is that the site started at the end of 1997 within another firm, and the MP3.com legal entity appeared in March 1998 estimate.

The idea was simple and rested on a lucky name: according to a retelling of an interview with Robertson, he bought the mp3.com domain in 1997 for about $1,000 from an owner who had registered it under his own handle at InterNIC rumor. The MP3 format was then a synonym for free music, and the domain worked as advertising on its own.

The early business was tiny: at the end of 1998 the company had 8 employees, and revenue for 1998 was $1.16 million. Growth came in 1999: by June 30 there were 142 employees, and revenue for the year rose to $21.9 million.

Year-by-year timeline

Lesser-known but significant facts

  1. The main site lived on advertising, not on music. In 2000 advertising brought in 96.5% of revenue, and in the fourth quarter a single customer accounted for about 61% of revenue. The money came not from CD sales or from My.MP3 but from the audience that musicians had brought to the site.
  2. Before the IPO the company itself warned investors about the hostility of the RIAA. Among the risks, the prospectus names possible opposition from "record companies and studios, the Recording Industry Association of America" because of the brand's association with mp3. The 2000 lawsuit was not a surprise; it was written into the risk factors.
  3. It paid artists for popularity, and that cost millions. Payback for Playback: $9.2 million from November 1999 to December 2000, later $1 million a month. For scale: revenue for all of 2000 was $80.1 million.
  4. The acquisition price was below the IPO price. In the IPO the shares cost $28; Vivendi paid $5 a share. The total value of the deal, $372 million, was only slightly more than the $345.7 million raised in the IPO before fees.

Legend vs. the record

Legend 1. MP3.com was a pirate service. Retellings lump MP3.com together with file sharing. The record shows two different services. The main site accepted music from the artists themselves, more than 125,000 artists by the end of 2000, with royalties of at least 50% on their CDs. The infringement concerned only My.MP3: the court found that the company had copied tens of thousands of CDs without permission and ruled this willful. The labels won the case precisely over My.MP3, not over the artists' site. Verdict: half true. The artists' site was legal; My.MP3 infringed copyright, according to the court's ruling.

Legend 2. One lawsuit killed MP3.com. The record says otherwise. The UMG judgment was $53.4 million, a modest sum next to the $345.7 million IPO. But it was not about one sum: eight of the ten plaintiffs settled, suits from music publishers (from March 2000) and independent labels ran in parallel, and the reserve for lawsuits came to $170 million against a loss of $279.5 million. The court counted $25,000 per CD, while the company itself, in its 10-K, recalculated the same rate for each work: on that basis its aggregate liability could exceed a billion dollars. The court did not "kill" the company: in 2000 it signed licenses with the majors, and in 2001 it was sold. Verdict: an oversimplification. The lawsuit did not kill MP3.com, but it ate up money and share value, after which the company was bought for $5 instead of $28.

Legend 3. Napster and MP3.com were the same thing. Retellings often merge the two. Napster did not store music: its server held an index, and the files went between users (per the Napster dossier). MP3.com stored copies and served them from its own servers, so it was itself the copying party. That is why MP3.com was tried as a direct infringer and the infringement ruling came three months after the suit, while in the Napster case the dispute was over an intermediary's liability for users' copies (per the Napster dossier). Verdict: false. What they shared was label content and label lawsuits; technically the models were different.

The first growth lever

The first lever was a catalog brought in by the creators themselves, and the money paid back to them for it. The artists uploaded the music themselves, and the company sold advertising. By July 1999 the site had music from more than 18,000 artists, more than 6 million visitors a month, and 44 million songs delivered. Advertising accounted for 89% of revenue from founding through March 31, 1999.

The second lever was Payback for Playback, from November 1999: artists were paid for the popularity of their tracks, and $9.2 million went to this over 14 months. By the end of 2000 the database had more than 125,000 artists and 13.5 million registered users. Revenue meanwhile grew from $1.2 million in 1998 to $21.9 million in 1999 and $80.1 million in 2000.

The paired story

The pair is Napster: two attempts at music on the network, and both ended in court.

Parallels today (projects from the catalog)

What a builder can take from this in 2026

  1. Tell apart "a platform for your own creators" and "a service that copies other people's work." At MP3.com the first worked and brought in an audience; the second lost in court within months. If your product stores or reproduces someone else's material, you need a license before launch, not after the lawsuit.
  2. Read the risk section of your own prospectus. In its 1999 prospectus MP3.com wrote that the RIAA might oppose it, and a year and a half later a $170 million reserve went to lawsuits, against revenue of $80.1 million.
  3. Do not build revenue on a single advertiser. 61% of quarterly revenue from one customer is a risk of the same order as a lawsuit.
  4. Paying creators for activity is possible, but count the cost. The $1 million a month for Payback for Playback was marketing that grows along with the audience.
  5. The licensing path is possible. Five majors signed licenses for 3–10 years with a fee for each added recording and royalties for plays; the "pay every time" model works for AI music today too.

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 →