From 'Linux Is a Cancer' to the World's Largest Open-Source Owner: Anatomy of Microsoft's Great Reversal

In 2001, Microsoft's CEO Steve Ballmer described Linux with a single, infamous word: cancer. It wasn't an offhand remark; it was doctrine. For years Microsoft treated free, open software as a mortal threat to the licensed-software empire that printed its money, and fought it with lawsuits, lobbying, and fear-uncertainty-and-doubt. Then, within roughly a decade, the same company open-sourced its crown-jewel developer framework, put Linux at the operating heart of its cloud, built Linux directly into Windows, and paid $7.5 billion for GitHub² — the beating heart of the open-source world. Same company. Complete inversion. The interesting question is never 'did they flip?' It's what changed underneath that made the old position untenable — and here the answer is unusually clean: the money moved.

Why hostility was once the rational strategy

It's easy to mock Ballmer in hindsight, and easy to get the lesson exactly backwards. In the Windows-and-Office era, Microsoft sold licenses, and its entire profit machine depended on proprietary software being scarce, paid-for, and locked in. Free software that was merely 'good enough' was a direct assault on the price of the product — not at the margin, but at the core. Under that model, calling Linux a cancer wasn't unhinged; it was a faithful description of the threat to the revenue line. Hostility was the correct strategy for the business Microsoft actually was. That's the part worth sitting with, because it means the later reversal wasn't a moral awakening — it was the same clear-eyed self-interest applied to a business that had changed shape underneath it.

Why the same hostility became suicidal

Then Microsoft's center of gravity shifted to Azure, its cloud, and the cloud business runs on the opposite incentive. A cloud provider doesn't sell licenses; it rents computing capacity by the minute, and it makes money when you run anything. The majority of the world's server workloads run on Linux. In that model, refusing to host Linux on Azure wouldn't protect Windows in the slightest — it would simply hand those customers, and their recurring monthly spend, directly to Amazon Web Services. The penguin that was a lost sale under the license model became a paying tenant under the cloud model — today more than 60% of customer compute cores on Azure run Linux.³ Same penguin. Opposite math. 'Microsoft loves Linux,' the company's later slogan, wasn't sentiment; it was an accounting identity wearing a friendly face.

The deeper shift: developers became the customer

Underneath the model change sits a deeper one that explains the GitHub purchase specifically. In the cloud era, the customer Microsoft most needs to win is the developer — because developers, not procurement departments, decide where workloads actually run. And developers had spent two decades migrating to open tools, to Linux, and to GitHub. A company whose instinct was to drag developers onto proprietary turf was, in cloud terms, fighting its own customers. The reversal is what it looks like when a firm finally accepts that it must meet its most important customer on that customer's home ground. Buying GitHub for $7.5 billion² wasn't nostalgia for code; it was acquiring the front door to every developer on earth — and, just as importantly, having the discipline to keep that door open rather than wall it off, because a closed GitHub would have driven the very developers it was meant to win straight to the competition.

This is why the timeline of the reversal matters. The shift wasn't a single announcement; it was a sustained pattern of costly, visible action — open-sourcing .NET, shipping the Windows Subsystem for Linux so developers could run Linux inside Windows, hosting Linux at scale on Azure, and leaving GitHub deliberately independent after the acquisition. That sequence is what made the U-turn credible. A reversal announced in slogans and contradicted in behavior reads as spin; one demonstrated in product, quarter after quarter, reads as strategy. Microsoft over-invested in the proof, which is precisely why the market believed it.

The cleanest way to summarize Microsoft's reversal is that nothing about its values needed to change. Linux went from threat to customer the moment the money moved from selling licenses to renting compute. The lesson isn't that Microsoft grew a conscience. It's that strategy is a function of the business model — and when the model inverts, the rare and valuable thing is having the stomach to invert with it, publicly, before the old position drags you under.

It's worth taking the skeptic's view seriously, because plenty of open-source veterans did. Microsoft's history included a documented playbook critics summarized as 'embrace, extend, extinguish' — adopt an open standard, add proprietary hooks, then use them to lock customers back in. From that vantage, buying GitHub looked less like peace than like a wolf acquiring the sheep's home address. The reason the fear has mostly not materialized is, again, the model: under cloud economics, extinguishing open source would destroy the very workloads Microsoft now profits from hosting. The old playbook made sense when Microsoft's goal was to protect license revenue; it would be self-harm when the goal is to maximize the compute customers rent. The incentives that once made Microsoft predatory are the same incentives that now keep it honest — which is a more durable guarantee than any promise, because it doesn't depend on goodwill, only on math. That, too, is the reversal's deepest tell: the behavior changed because the self-interest changed, not the other way around.