United States v. Microsoft and the Browser War That Made Room for Google

On May 18, 1998, Attorney General Janet Reno and Joel Klein, the head of the Justice Department's antitrust division, announced that the United States was suing Microsoft. A group of state attorneys general filed a parallel suit the same day. The charge was that Microsoft, which made the operating system on the great majority of the world's personal computers, had used that dominance to crush Netscape, the company whose Navigator browser had introduced millions of people to the World Wide Web. That same year, in a rented garage in Menlo Park, California, two Stanford graduate students incorporated a small company with an odd name to run their search engine. The two events had nothing to do with each other at the time. The argument of this post is that they had a great deal to do with each other later.

In short: The government proved that Microsoft illegally maintained its Windows monopoly, but an appeals court threw out the order to split the company, and a 2001 settlement imposed only conduct rules. Even so, the case changed how Microsoft behaved during the years when Google and other web companies were small and vulnerable. The lesson cuts both ways: litigation can protect competition even when it fails, and weak remedies leave the work unfinished.

Why a browser scared the biggest software company on earth

In the mid-1990s, Microsoft's power rested on Windows. Software developers wrote their programs for Windows because nearly every PC ran it, and nearly every PC ran it because that was where the programs were. Economists call that an applications barrier to entry. A rival operating system, no matter how good, would launch with no software, and without software nobody would buy it.

Netscape threatened that arrangement in an indirect way. A web browser, combined with a programming language like Java that could run on any machine, might let developers write programs once and run them anywhere. If the browser became the platform, the operating system underneath would matter much less. Bill Gates understood this early. In a widely circulated 1995 memo titled "The Internet Tidal Wave," he told Microsoft's executives that the internet was the most important development since the IBM PC and that the company had to respond.

Microsoft's response was Internet Explorer, which it bundled with Windows and gave away. It also pressed the companies that sold computers and the companies that provided internet access to favor Explorer and to drop or bury Navigator. It worked. Netscape, which had held most of the browser market in the mid-1990s, lost most of it within a few years, and AOL bought the company in a deal announced in late 1998.

What the trial showed

The trial opened in October 1998 before Judge Thomas Penfield Jackson of the federal court in Washington. It became famous partly for its theatre. The government played portions of Bill Gates's videotaped deposition, in which he quarreled over the meaning of ordinary words and claimed not to remember emails he had written. Microsoft's lawyers presented a video demonstration intended to show that removing Explorer from Windows broke the system, and government lawyers then raised questions about how that video had been made.

The substance mattered more. Witnesses from Netscape, Apple, Intel, AOL and computer makers described how Microsoft had used its control of Windows licensing to shape their choices: threats to raise prices, the withholding of technical cooperation, deals that rewarded partners for excluding Navigator. Internal emails showed executives talking about the browser not as a product to win customers with but as a means to protect Windows.

In November 1999, Judge Jackson issued findings of fact concluding that Microsoft held monopoly power in PC operating systems and had used it to harm consumers and competition. In April 2000 he ruled that Microsoft had violated the Sherman Act by illegally maintaining that monopoly, attempting to monopolize the browser market and unlawfully tying the browser to the operating system. In June 2000 he ordered that Microsoft be split into two companies, one for the operating system and one for applications.

The appeal: liability stands, the breakup falls

Microsoft appealed, and in June 2001 the full Court of Appeals for the District of Columbia Circuit issued a unanimous ruling that pleased neither side completely.

The court upheld the heart of the case: Microsoft had illegally maintained its monopoly in operating systems through a series of anticompetitive acts. That finding has been the foundation of monopoly law in the technology industry ever since.

But the court reversed the attempted monopolization finding, sent the tying claim back for a different, more flexible analysis, and vacated the breakup order. It also removed Judge Jackson from the case. While the trial was still going on, he had given secret interviews to reporters, to be published after his ruling, in which he made disparaging remarks about Microsoft and its executives. The appeals court said this created an appearance of bias that required a new judge to decide the remedy.

The case went to Judge Colleen Kollar-Kotelly. By then the political ground had shifted. The Bush administration took office in 2001, and after the September 11 attacks the Justice Department was eager to wrap things up. In November 2001 it reached a settlement with Microsoft. The deal did not break up the company. It required Microsoft to let computer makers install rival software and hide Microsoft's own, banned exclusive deals and retaliation against partners, and required the company to share certain technical information with competitors. Several states pressed for stronger remedies; the court largely approved the settlement in 2002. Oversight was extended several times and finally ended in 2011.

The argument for failure

It is easy to read this history as a defeat for the government, and plenty of people do. Here is the strongest version of that view.

Microsoft remained enormously profitable. Windows remained dominant for years. Internet Explorer's share of the browser market peaked around the time of the settlement, and Netscape never recovered. The conduct remedies were modest and, according to many critics, weakly enforced. If the purpose of the case was to restore competition in browsers and operating systems, it did not do that. Market changes, especially the rise of smartphones, eventually did more to dethrone Windows than any court order.

There is also a skeptical view from the right and from some economists: that Microsoft's browser was a natural feature of an operating system, that giving it away benefited consumers, and that the government was second-guessing product design in a fast-moving market where today's monopoly becomes tomorrow's has-been without help from lawyers.

Those critiques have merit. The breakup never happened, and the remedy was thin.

There is a cost argument too. The case consumed years of work by government lawyers and economists, and it absorbed the attention of the company's leadership at a time when the industry was shifting under everyone's feet. Critics on the right have long argued that this kind of litigation is a tax on success, punishing a firm for building a product people chose. That view deserves an answer rather than a sneer, and the answer has to come from what the case actually changed.

The argument for success, and why I find it more persuasive

The best case for the lawsuit is not about remedies at all. It is about behavior during the years the case hung over the company.

Here is the argument. From 1998 through the decade that followed, Microsoft was a defendant and then a company under court supervision, with every executive email potentially headed for a courtroom. Former Microsoft employees and outside observers have argued that this made the company far more cautious about using Windows to kill new rivals. During exactly those years, a cluster of internet companies grew from startups into giants: Google, which had launched its search engine in 1998; Amazon; and later Facebook and others.

Consider what Microsoft might have done to Google in a different climate. It controlled the operating system and the dominant browser. It could have made its own search engine the unchangeable default in Explorer, or made rivals harder to reach. In practice, Microsoft did make its own search engine the default in Explorer, and Google worried about this enough to complain to regulators. But the company could not use the kind of hardball it had used against Netscape without risking new findings against it. The settlement's ban on retaliation and exclusive dealing, and the reputational cost of looking like a repeat offender, created breathing room.

This is a counterfactual, and nobody can prove it. But it fits what people inside the industry have said, and it fits the timeline. The most valuable companies of the 2010s were born and grew up during the years when the most powerful software company on earth was being watched.

The circle closes

There is an irony at the end of this story. In 2020 the Justice Department, joined by many states, sued Google for illegally maintaining its monopoly in internet search. When Judge Amit Mehta ruled against Google in August 2024, he relied heavily on the Microsoft precedent. The legal framework the D.C. Circuit had built in 2001, for judging whether a dominant firm's conduct crosses the line, was the one used against the company that had grown up in Microsoft's shadow. And one of the central issues was the same kind of thing that had sparked the Microsoft case: control of the default, in Google's case through payments to Apple and others to be the preset search engine in browsers and phones.

Google, in other words, may have benefited from the Microsoft case twice: first from the breathing room, and second from the remedy question. Judge Mehta, like the courts in Microsoft, declined to order a breakup. Whether that caution is wisdom or repetition is the question our next decade of tech enforcement will answer.

What the case teaches

I take three lessons from it.

First, liability matters even when remedies disappoint. A judicial finding that a company broke the law changes how that company behaves and shapes how future cases are decided. Enforcers should bring cases that clarify the law, not only cases where a dramatic remedy is likely.

Second, remedies need to be built for the market's future, not its past. By the time the Microsoft settlement took effect, the browser fight was over. The settlement's most valuable provisions were the ones aimed at the next fight: no retaliation, no exclusive deals, open interfaces.

Third, politics shapes enforcement more than the law admits. The same case, with the same findings, ended in a breakup order under one administration and a negotiated settlement under the next. Antitrust is not self-executing. It depends on who is running the agencies, how much money they have and how much public attention their work gets. That makes it a subject for voters, not only for lawyers.

Europe took a different road

While the American case wound down, European regulators opened their own investigation of Microsoft, and the contrast is instructive.

In March 2004 the European Commission ruled that Microsoft had abused its dominant position. It fined the company about 497 million euros, then a record for a single company in an EU competition case. The commission ordered Microsoft to offer a version of Windows without its Windows Media Player and to share information that would let rival server software work with Windows. Microsoft fought the decision and lost on appeal in 2007, and the commission later imposed additional penalties for slow compliance.

The media-player remedy is a cautionary tale. Microsoft produced the stripped-down version, called Windows XP N, and sold it at the same price as the full edition. Almost nobody bought it. A remedy that gives consumers a choice they have no reason to make does little.

The browser fight in Europe ended differently. In 2009 Microsoft agreed to show European Windows users a choice screen listing rival browsers, and the commission accepted that commitment. Browser competition was already changing by then, with Firefox growing and Google having released Chrome. Later, after Microsoft failed for more than a year to display the screen as promised, the commission fined it again in 2013.

Put the two experiences side by side and a pattern appears. Remedies that require a company to offer an alternative product it does not want to sell tend to fail. Remedies that change what the user sees at the moment of choice, or what a partner is free to install, have a better chance. The American settlement's provisions on what computer makers could install were in the second category, which may be one reason they mattered more than their critics expected.

What you can do

The tech antitrust cases of this decade are being decided in courtrooms, at the agencies and in Congress's budget votes. Here is how to have a say:

• Ask your members of Congress to fund the antitrust agencies. The Justice Department's antitrust division and the FTC take on the largest companies in the world with a small fraction of their legal resources. Look up your representative and senators by address to find their phone numbers, contact forms and latest votes.
• Follow the cases as they reach the top. Appeals in the major tech cases can end up before the justices. Court Watch lists the Supreme Court's opinions this term with the Court's summaries.
• Track what Congress does on competition. The Congress tracker follows votes and bills, including proposals on app stores, data and mergers.
• Read the wider reporting. Our Big Tech and AI page collects coverage of the companies and the cases.

The Netscape engineers lost their browser war. The web they helped open survived it, partly because a federal court decided that winning a market and keeping it by any means are not the same thing.