Microsoft – The Lost Decade That Wasn’t

The Number That Doesn’t Add Up

When Steve Ballmer announced in August 2013 that he would step down within a year, Microsoft’s shares jumped as much as 9 percent and closed at $34.75. The company was then worth about $269 billion, less than half of the $604 billion it had been worth when Ballmer took over as CEO, and its stock had lost 44 percent during his tenure. Read in isolation, those figures describe a failure so complete that the market cheered the departure of the man in charge. But set them beside the company’s own accounts and they stop making sense. In fiscal 2000, the year Ballmer inherited the job, Microsoft reported revenue of $22.96 billion
and net income of $9.42 billion. In fiscal 2013, the last full year before his exit announcement, revenue stood at $77.85 billion and net income at $21.86 billion. Revenue had more than tripled and profit had more than doubled, during a period in which the share
price went nowhere. Both facts are true, and the distance between them is the real subject of this story. Microsoft did not spend those years shrinking. It spent them being repriced, from a company the market treated as the inevitable owner of the future to one it treated as the aging owner of the past, while a handful of decisions, some bad, some quietly brilliant, determined which of those descriptions would turn out to be right. Thirteen years later the answer is on the income statement: fiscal 2026 revenue reached $331.8 billion, up 18 percent, with operating income of $155.2 billion. That is roughly fourteen and a half times the revenue of the year Ballmer started, and a market value that mid-2026 estimates put near $3.7 trillion. How a company gets from the first set of numbers to the second, without ever stopping being profitable, is not a story about rise and fall. It is a story about a platform, a valuation, and the difference between them.

A BASIC Interpreter and a Contract Written for Someone Else’s Hardware

Microsoft’s origin is smaller and stranger than its mythology. Bill Gates and Paul Allen founded it on April 4, 1975, with a single purpose: to write and sell a BASIC interpreter for the Altair 8800, in Albuquerque, New Mexico, because that was where the machine’s maker, MITS, was based. Microsoft did not invent the personal computer, and it did not invent BASIC; its first commercial act was to supply software to a hardware company that needed it.
According to one account, the pair sold their code to MITS for $3,000 plus a share of royalties capped at $180,000, a deal that looks modest until you notice what it established: a software company whose value depended on being the layer that other people’s machines ran on. The decisive test of that model came in 1980, when IBM, preparing its first personal computer, went looking for an operating system. Microsoft signed with IBM in November 1980 and, in July 1981, bought the rights to a CP/M-style system called QDOS, later 86-DOS, from Seattle Computer Products for $50,000. Accounts of the exact price differ. One history
says Microsoft first licensed the system for $25,000 before buying it outright, while another puts the total at $75,000, a figure that would equal the two payments combined, though the sources do not say so explicitly. What every account agrees on is the clause that mattered more than the price. Microsoft insisted on keeping the right to license the system to other manufacturers under its own name, MS-DOS, while IBM sold its version as PC-DOS. One historian’s reading is that IBM, whose culture favored hardware, undervalued software and effectively handed Microsoft the right to define the standard for every compatible PC that
followed. That is an interpretation, not a documented statement of IBM’s reasoning, and it should be held as one. But the outcome is not in dispute: when other companies built machines compatible with the IBM PC, every one of them needed MS-DOS, and Microsoft collected from all of them. This is the structural trick the rest of the story keeps returning to.
Microsoft rarely won by building the best product in a category. It won by owning the layer that everyone else’s products had to run on.

The Operating System as Toll Road

By the mid-1990s that layer had become Windows, and the toll road was fully built.
Microsoft launched Windows 95 in August 1995 and followed in September 1997 with Internet Explorer 4.0, a stepped-up challenge to Netscape’s browser. The browser fight is where the platform model collided with the law. The U.S. government argued that Microsoft was using its control of the operating system to choke competitors, and in the trial that followed, Judge Thomas Penfield Jackson declared Microsoft a monopoly and, on June 7, 2000, ordered the company broken up into an operating-system business and an applications business. Five months earlier, on January 13, 2000, Gates had stepped aside as chief executive and promoted Ballmer, so the new CEO’s first months were spent under the threat of dismemberment. The threat did not last. In a unanimous 7–0 decision in June 2001, the appeals court vacated the breakup, concluding that Jackson’s remarks to the press
outside the courtroom created the appearance of bias. The same court did not erase the underlying finding: it held that Microsoft had illegally maintained its monopoly in PC operating systems, but that it had not illegally attempted to monopolize the market for web browsers. The Justice Department then announced it would no longer seek a breakup, and the case ended in a settlement. The outcome is easy to misread in either direction. Microsoft was not exonerated, since the monopoly finding stood. But it was not dismantled either, and its operating-system franchise, the thing the market was most afraid of losing, survived intact. Whether the legal ordeal changed the company’s behavior or its ambitions is a question the available evidence does not settle, and it is worth resisting the common assumption that it did.

December 1999: The Valuation Peaks, Not the Business

The most important date in this story is not a product launch or a lawsuit but a stock price.
Sources disagree on the precise figure and day, which is itself worth knowing. The Associated Press put Microsoft’s value at $604 billion when Ballmer took over, another account gives roughly $600 billion, and a third pins the peak at $614 billion on December 27, 1999. The honest summary is that the company was worth about $600 billion at the end of 1999, at the top of the dot-com bubble. The more revealing number is the price investors were paying for each dollar of earnings. At the end of 1999 the stock traded at more than 74 times earnings, near $58 a share. A company at that multiple is not being valued on what it earns; it is being valued on the assumption that its growth will be near-limitless, that every new market from the internet to the living room will become another Microsoft franchise. And the business at that moment, strong as it was, did not justify such an assumption. Fiscal 2000 revenue grew 16 percent to $22.96 billion and net income was $9.42 billion, a superb result for a company of that size and nothing like the trajectory a 74-times multiple implied. This distinction is the key to everything that follows. The peak of 1999 was a peak of expectations, not of operations. A company can keep growing, keep earning and keep dominating its market while its shares fall, simply because the price it was assigned at the top required a future that no real business could deliver. Much of what the market later treated as Microsoft’s decline was the slow, painful correction of that gap, and much of the rest was real strategic damage layered on top of it. Separating the two is the most useful thing a reader can do with this history.

Fourteen Years of Growth Nobody Wanted to Pay For

The Ballmer years are the part of the story most distorted by hindsight, so it helps to begin with what the numbers show. One analysis describes the stock as flat for more than a decade even as revenue and profits tripled, but the filings tell a slightly different story: revenue grew about 3.4 times, from $22.96 billion to $77.85 billion, while net income grew about 2.3 times, from $9.42 billion to $21.86 billion. This was a company getting richer, not poorer, and the tension between its financial performance and its market reputation had real internal costs. The same analysis argues that a decade of flat shares turned internal competition zero-sum, as employees came to believe advancement could only come at colleagues’ expense, an interpretation consistent with the culture Satya Nadella would later describe himself as inheriting, but one that rests on a single secondary account and should be treated as a plausible diagnosis rather than a proven fact. The defining episode of the period is the iPhone.

In January 2007, Ballmer laughed at the device in a television interview, calling it the most expensive phone in the world and saying it would not appeal to business customers because it lacked a keyboard. Three months later, in an interview with USA Today, he went further: “There’s no chance that the iPhone is going to get any significant market share.” The quote has become the shorthand for Ballmer’s failure, and it deserves a fairer reading than that. His reasoning, as he laid it out, was a platform logic of the kind that had made Microsoft rich. He said he would rather have Microsoft’s software on 60 to 80 percent of the 1.3 billion phones sold each year than Apple’s expected 2 or 3 percent. That was a coherent strategy for a company whose model was licensing software to many manufacturers, and at the time Windows Mobile held about 6 percent of smartphones, far behind Symbian’s 72 percent, per Canalys figures cited at the time. One commentator argues that Ballmer was even partly right on the specifics, noting that the original iPhone did not work well with the Exchange systems common in businesses.

The mistake was not the arithmetic. It was the assumption that the phone market would behave like the PC market, with an open licensing layer on top of commodity hardware, when Apple had just shown that a tightly integrated device could capture the profit instead. Microsoft’s toll-road model had met a market where the road was not for sale.
The same years contain evidence of both the company’s blind spots and its foresight. In 2012 it took a $6.2 billion write-down on a Quantive, an online advertising business it had bought, at the time the largest in its history. And yet in October 2008, Ray Ozzie unveiled Windows Azure at the company’s developer conference, a cloud platform that became commercially available on February 1, 2010, four years before Nadella became CEO. This is the fact the simple chronology misses. The popular version credits Nadella with inventing Microsoft’s cloud business; the record shows that the seed was planted under Ballmer and Ozzie, and that what Nadella did was something different and arguably harder: he moved the cloud from the periphery to the center of the company.

The Phone Bet That Closed the Era

If the iPhone remark was a misjudgment of a market, the Nokia acquisition was a decision to double down on it. In the deal’s last act, Microsoft completed its purchase of Nokia’s phone business in April 2014 for $7.2 billion, when Windows Phone held roughly 3 percent of the market. The acquisition was the last major move of the Ballmer era, and one account says he pushed it through even though a number of executives warned against it. The logic was recognizable: if you cannot win the phone market by licensing software to other makers, buy a maker and control the whole stack, as Apple did. But the timing and the price proved costly. In July 2015, fifteen months after closing, Microsoft announced 7,800 job cuts, mostly in the phone unit, and a $7.6 billion impairment charge, the biggest in its history, with the company itself conceding that the unit’s prospects were below original expectations.

The write-down was larger than the original purchase price, in part because the final cost included assets and intellectual property beyond the headline figure. In an email to staff, the new CEO described the change in terms that quietly closed a chapter: Microsoft was moving from a strategy of growing a standalone phone business to one of building a vibrant Windows ecosystem that included its own devices. That sentence was a retreat, and it was also a decision. It meant Microsoft would stop trying to win a market it was not suited to win, and would redeploy its attention to the one where its structural advantage was real: software
that other people’s systems depend on.

The Reset: Less Windows, More Platform

Satya Nadella became CEO on February 4, 2014, with the company worth about $311 billion and widely regarded as a once-great institution whose best days were behind it. He was the first Microsoft CEO who was neither a founder nor an early employee, and he arrived in a company that one business-school case describes as fading toward irrelevance and plagued by internal fights and inertia. His first and least visible intervention was cultural. Drawing on a book on mindsets his wife had given him, he pushed employees to move away from a culture in which people proved they were right, toward one in which they proved they were learning, and the contrast between a know-it-all and a learn-it-all became the company’s shorthand for the change. Whether the culture shift caused the financial results, or merely accompanied them, is not something the evidence can separate cleanly, but the sequence of strategic choices that followed is easier to document.

Microsoft stopped treating Windows as the center of its universe and began treating it as one product among many that its services should serve. It bought the professional network LinkedIn in June 2016 for $26.2 billion and the code-hosting platform GitHub in June 2018 for $7.5 billion, the latter a bet on developers whom Microsoft had spent decades competing with. And it made Azure, the cloud service launched in the Ballmer era, the engine of the company. The market noticed.
Microsoft passed Apple to become the world’s most valuable company in the last week of November 2018, and its valuation crossed $1 trillion on April 25, 2019, up more than 230 percent since Nadella took over. Asked about the milestone, Nadella reacted with something close to irritation: “I would be disgusted if somebody ever celebrated our market cap.” The comment is revealing for what it shows about how he understood the lesson of the 1999 peak. A market value is a verdict on expectations, and the company had already learned what happens when the expectations are wrong.

The Bet on Someone Else’s Model

The most recent chapter repeats the pattern of the first, with an important twist. In artificial intelligence, Microsoft again chose to own the layer that others build on, rather than to build the best model itself. It invested its first $1 billion in OpenAI in 2019, and the commitment grew to about $13 billion in total. The relationship delivered what the Nokia deal had not: a way to add a new product category to the existing platform without fighting for share in a market the company did not control. By fiscal 2026 the results were visible. Azure revenue passed $100 billion for the first time, and Microsoft 365 Copilot reached more than 30 million paid seats.

Microsoft Cloud revenue exceeded $214 billion for the year, up 27 percent, and cash from operations rose by $46.8 billion to $182.9 billion. The company’s own AI business passed an annual revenue run rate of $37 billion in the third quarter, up 123 percent from a year earlier. Meanwhile the older franchises showed their age: in the fourth quarter, Windows OEM and Devices revenue fell 7 percent and Xbox content and services fell 10 percent. One detail in the accounts deserves care. Fiscal 2026 net income grew 31 percent on a GAAP basis, but only 22 percent on the non-GAAP basis that excludes the impact of Microsoft’s OpenAI investments, which is the company’s own way of telling readers that part of its reported profit growth reflects accounting for that stake rather than the operating business.

The partnership has also been renegotiated, and each version has given Microsoft somewhat less control. In October 2025, Microsoft agreed to hold 27 percent of the restructured OpenAI Group, valued at about $135 billion, but gave up its right of first refusal to be OpenAI’s cloud provider. Its rights to OpenAI’s intellectual property were preserved through 2032. A later amendment loosened the terms again: Microsoft’s license is no longer exclusive, OpenAI can serve customers across any cloud provider, and Microsoft no longer pays a revenue share to OpenAI, while OpenAI continues capped payments to Microsoft through 2030. The structural question this raises is the oldest one in the story. Microsoft’s greatest successes came when it controlled the layer other products depended on, MS-DOS and then Windows. In AI, it depends on a partner it does not control, and the terms of that
dependence have been changing. Investors are weighing it too: when the company recently unified its Copilot products, shares slipped about 1.2 percent to around $503, as traders questioned whether higher paid usage would cover the added compute and support costs.
That doubt is the same doubt that attached to the stock in 1999, in a different form: not whether the business is good, but whether the price already assumes more than it can deliver.

Half a Company in One Stake

The first number in this story was $269 billion, the value of Microsoft on the day Ballmer announced his exit. It is worth setting that figure against the most recent one. A single minority stake in another company, valued at $135 billion in October 2025, equalled roughly half of the entire market value of Microsoft in August 2013. In thirteen years the firm that the market had written off as a monument to the PC era grew to a size at which its investment in one partner is worth about what the whole company once was. That contrast does not prove that the market was wrong in 2013 or right in 2026. It shows how much of Microsoft’s history is the history of valuations chasing a business that had already changed underneath them, first too high in 1999, then too low in 2013, and now at a level that again requires believing in a future the company has not yet delivered.

THE CURVE IN 60 SECONDS

• 1975: Gates and Allen found Microsoft to sell BASIC for the Altair.
• 1980–81: IBM contract; Microsoft buys the rights to 86-DOS for $50,000 and keeps
  the right to license it to others.
• 1995–97: Windows 95 and Internet Explorer 4.0 launch.
• Dec 1999: Value near $600 billion at over 74 times earnings.
• Jan 2000: Gates hands the CEO role to Ballmer.
• 2000–01: Breakup ordered in June 2000, vacated by the appeals court in 2001.
• 2007: Ballmer dismisses the iPhone.
• Oct 2008: Azure unveiled; commercial launch February 2010.
• 2012: $6.2 billion aQuantive write-down.
• Aug 2013: Value of $269 billion as Ballmer announces his exit.
• Feb 2014: Nadella becomes CEO.
• Apr 2014: Nokia phone business bought for $7.2 billion.
• Jul 2015: $7.6 billion write-down, 7,800 job cuts.
• 2016, 2018: LinkedIn ($26.2 billion) and GitHub ($7.5 billion).
• Apr 2019: $1 trillion.
• Oct 2025: 27 percent OpenAI stake, valued at about $135 billion.
• FY2026: Revenue of $331.8 billion.

THE POSTMORTEM

What created the growth. A contract structure, not a product. Microsoft kept the right to license its operating system to every maker of compatible machines, and that made it the toll collector for an entire industry. What sustained the peak. Windows and Office extended that position through the 1990s, and the antitrust ordeal threatened the structure but, after appeal, did not break it. Where the vulnerability first appeared. In the gap between the business and its valuation: at over 74 times earnings in late 1999, the market was pricing a future the company could not deliver, while the business itself remained strong. What changed the trajectory. Not one event but a sequence: the bursting of the bubble, the antitrust threat, and then a shift in computing, from licensed software on commodity PCs to integrated devices, where Microsoft’s model did not translate. The decisive misstep was not any single remark about the iPhone but the repeated attempt to win phones with a PC-era strategy, which ended in the Nokia write-down.

Internal versus external. The valuation reset and the dot-com crash were largely external, and the platform shift to smartphones was a market change Microsoft did not cause. The decisions to dismiss the iPhone, to buy aQuantive, and to acquire Nokia were
internal and controllable. Azure, conversely, was an internal decision that paid off a decade later, which shows that the Ballmer era contained both the errors and the seeds of recovery.
What remains uncertain. Three things. First, how much of the later success is attributable to Nadella’s cultural reforms rather than to the timing of cloud and AI demand; the evidence supports both but cannot separate them. Second, whether the antitrust case
shaped Microsoft’s behavior in the 2000s, which is often asserted and rarely demonstrated.
Third, whether the current valuation reflects durable economics or the same kind of expectations that peaked in 1999. Confidence assessment. That the stock stagnated while the business grew: confirmed by filings. That Azure originated before Nadella: strongly
supported. That culture change drove the recovery: plausible but not demonstrated. That the OpenAI dependency is a strategic risk: plausible, given that the terms have repeatedly changed.

LESSONS FROM THE CURVE

A valuation is not a business. Microsoft’s revenue grew more than threefold while its market value fell by more than half, so reading either number alone misleads.
The strongest position is the layer others depend on, and the weakest is competing against it. Microsoft’s successes came from platforms; its largest write-downs came from trying to win categories where someone else owned the platform.
Retreat can be a strategic decision. The Nokia write-off was costly, but ending the standalone phone effort freed the company to concentrate where it had a real advantage.
Dependence changes character over time. What began as exclusivity with OpenAI has become a looser relationship, and a company that built its history on controlling the layer should expect scrutiny of any layer it does not control.

FAQ

What happened to Microsoft after its peak in 1999?
Its stock fell sharply and stayed below the 1999 peak for more than a decade, but its revenue and profit kept growing. The gap reflects a correction of expectations rather than a collapse of the business.
Did Microsoft really miss the iPhone?
Its CEO publicly dismissed it in 2007, and its own phone efforts later failed, culminating in a $7.6 billion write-down in 2015. His reasoning followed Microsoft’s licensing logic, which proved wrong for smartphones.
Who built Microsoft’s cloud business?
Azure was unveiled in 2008 under Ray Ozzie and launched commercially in 2010. Satya Nadella, CEO from 2014, made it central to the company.
Was Microsoft broken up?
A judge ordered a breakup in June 2000, but the appeals court vacated it in 2001, and the case ended in a settlement.
What is Microsoft’s stake in OpenAI worth?
It was valued at about $135 billion in the October 2025 restructuring; later estimates are higher but come from secondary trackers, so treat them as approximate.

SOURCES

SEC filings (Form 8-K and 10-K, fiscal 2026); Microsoft press releases (fiscal 2000, fiscal 2026 Q3 and Q4); Associated Press (2013 Ballmer coverage); The Register, TechCrunch, CIO and PCWorld (2015 Nokia write-down); Seattle Times and Fox News chronologies (antitrust case); HPCwire and Redmond Magazine (2001 appeals ruling); Bloomberg Businessweek (2019); MacRumors, Fortune and Daring Fireball (Ballmer on the iPhone); Directions on Microsoft, Yahoo Finance, The Hill and TradingView (OpenAI agreements); London Business School case study; Petri and Nira (early history).