On June 4, 2018, Microsoft announced it was acquiring GitHub for $7.5 billion in stock. The developer community's response was immediate, passionate, and distributed across every platform developers use to be immediate and passionate: Twitter, Hacker News, Reddit, and — with a certain irony that was not lost on anyone — GitHub itself.
The reaction split broadly into three camps. The first said this was fine, Microsoft had changed, Satya Nadella was not Steve Ballmer, the company that had called open source a “cancer” in 2001 was now one of the largest contributors to the open source ecosystem, and GitHub under Microsoft would probably be better resourced than GitHub as an independent company burning through venture capital. The second said this was a catastrophe, that developer tools in the hands of a corporation with a history of embrace-extend-extinguish strategies could not be trusted, and that the only rational response was to migrate to GitLab immediately. The third said they had opinions but were too busy finishing a feature to act on them.
Eight years later, we have enough data to evaluate all three positions.
The Company GitHub Was Before the Acquisition
GitHub was founded in 2008 by Tom Preston-Werner, Chris Wanstrath, PJ Hyett, and Scott Chacon. It grew rapidly by solving a real problem: distributed version control with Git was powerful but socially awkward, and GitHub made it social. Pull requests, issues, the activity feed, the contribution graph — these were not technical innovations but social and UX innovations that transformed Git from a tool used by individuals into a platform used by communities.
By 2018, GitHub hosted over 85 million repositories and had more than 28 million developers on the platform. It had become, functionally, the public square of software development — the place where open source projects lived, where developers built their professional identities, where the collective output of the global software development community was stored and made accessible.
It was also, financially, not doing particularly well. GitHub had raised significant venture capital, had grown headcount aggressively, and had experienced internal turbulence — Preston-Werner had resigned in 2014 following harassment allegations, and the company had gone through a period of significant cultural difficulty. The business model — free for public repositories, paid for private ones — was being pressured by competitors offering more aggressive pricing. The acquisition provided both capital and strategic direction that the company was struggling to generate independently.
Why Microsoft Wanted It
The $7.5 billion price was not, primarily, a bet on GitHub's existing revenue. It was a bet on GitHub's position in the developer ecosystem — specifically, on the strategic value of being the company that owns the platform where most of the world's software development happens.
Microsoft under Satya Nadella had made a strategic pivot toward cloud services, with Azure as the central product. Developers are the primary buyers of cloud services, or at least the primary influencers of cloud purchasing decisions in organisations. A company that is trusted by developers, that is embedded in their daily workflow, that hosts their code and their professional identity — that company has a form of influence over developer preferences and consequently over cloud purchasing that is extremely difficult to replicate through conventional marketing.
GitHub was also, by 2018, the largest repository of code in the world. The AI applications of that code repository — which would materialise six years later in the form of GitHub Copilot — were already visible in outline to anyone paying attention to where machine learning research was heading. The acquisition was, in part, a bet on a future in which the ability to train AI systems on the world's code would be extraordinarily valuable. That bet paid off.
What Actually Changed
The developers who predicted catastrophe were largely wrong about the specifics, even if some of their underlying concerns were prescient.
GitHub did not become a Microsoft product in the sense of being absorbed into Microsoft's design language, pricing strategy, or corporate culture. It has maintained its own brand, its own interface, and a significant degree of operational independence. The free tier became more generous after the acquisition, not less — private repositories became free for all users in 2019, which was a direct competitive response to GitLab's offering but also a benefit to every developer on the platform.
The platform has improved in capability. Actions — GitHub's CI/CD automation system — launched in 2019 and became one of the most widely used automation tools in software development. Codespaces, GitHub's cloud development environment, launched in 2021. Copilot, the AI coding assistant trained on GitHub's code repository, launched in 2022 and has become one of the most significant productivity tools in software development, with adoption rates that suggest it has genuinely changed how many developers write code.
The concerns that were more justified: Copilot's training on open source code raised significant legal and ethical questions about whether using code licensed under copyleft licences to train a commercial AI product was consistent with those licences. This debate is ongoing and has not been resolved. The concentration of developer infrastructure in a single platform — and specifically in a platform owned by one of the world's largest corporations — is a legitimate concern about ecosystem resilience that the acquisition amplified rather than created.
What GitLab Did With the Panic
The immediate beneficiary of developer anxiety about the acquisition was GitLab, which reported a significant spike in repository migrations in the days following the announcement. GitLab was well-positioned to receive them: it offered a genuinely competitive product, a self-hosted option for organisations that did not want their code on any external platform, and an open core business model that gave it credibility with the open source community.
The migration wave was real but not as large as the initial social media noise suggested. Most developers who said they were moving to GitLab did not move to GitLab. The friction of migrating an established workflow, the network effects of a platform where your collaborators and your professional history already exist, and the absence of the catastrophe that had been predicted meant that the urgency faded before the action followed.
GitLab has grown substantially since 2018 regardless — it went public in 2021 — and the competitive pressure it exerts on GitHub has probably contributed to the platform improvements that followed the acquisition. The threat of migration, even when not acted upon, is a useful market discipline.
GitHub Copilot: The Thing That Changed Everything
The most significant post-acquisition development — and the one that has most fundamentally changed what GitHub is — is Copilot. Launched in technical preview in 2021 and generally available from 2022, Copilot is an AI pair programmer trained on GitHub's code repository that suggests code completions, generates functions from natural language descriptions, and increasingly operates as an autonomous coding agent for defined tasks.
Copilot's adoption has been faster than almost any developer tool in history. GitHub reported over 1.8 million paying Copilot subscribers by early 2024 and over 50,000 organisations using it. The productivity effects reported by users are real enough that they have begun to change how teams are structured, how much code individual developers are expected to produce, and what skills are most valued in software engineering.
Copilot is also the clearest expression of why Microsoft paid $7.5 billion for GitHub in 2018. The training data that powers Copilot is the code repository that GitHub had been accumulating for fifteen years. Without the acquisition, Microsoft would not have had access to that repository at the scale and terms that made Copilot possible. The AI future that was visible in outline in 2018 arrived, and GitHub's position at its centre was not accidental.
What Developers Think Now
Developer opinion on the acquisition in 2026 is, broadly, that it was fine. Not enthusiastically fine — the concerns about platform concentration and the Copilot licensing questions have not disappeared — but fine in the sense that the predicted catastrophes did not materialise and the platform has improved in ways that are used and appreciated by most of its users.
The developers who migrated to GitLab in 2018 are mostly still on GitLab and mostly satisfied with that choice. The developers who stayed on GitHub are mostly still on GitHub and mostly using Copilot. The open source community's relationship with GitHub is complicated in ways it has always been complicated, amplified by Copilot's training data questions but not fundamentally transformed by them.
The acquisition changed GitHub. It did not, as the most alarmed predictions suggested, destroy it. Whether the thing it changed GitHub into is better or worse than what GitHub would have become independently is a counterfactual that cannot be evaluated. What can be evaluated is the platform as it exists: the largest developer platform in the world, with more capabilities than it had in 2018, owned by a corporation whose interests are not always identical to the interests of the open source community it hosts, and used daily by the overwhelming majority of professional software developers regardless.
For the Developers With Opinions
The GitHub Double Meaning Coder T-shirt (India) and (International) — for developers who have used the platform, argued about the platform, threatened to leave the platform, and are still on the platform.
0 comments