Seven years after independence promise, GitHub becomes Microsoft AI division
Thomas Dohmke's departure and competitive pressure from Cursor reveal how AI transforms big tech acquisitions
Microsoft's 2018 promise that GitHub would "retain its developer-first ethos and will operate independently" died quietly this week. Thomas Dohmke's resignation as GitHub CEO and the platform's absorption into Microsoft's CoreAI division represent more than corporate restructuring—they mark the end of big tech's most visible experiment in acquisition independence.
The timing tells the story. When Microsoft acquired GitHub for $7.5 billion in 2018, preserving independence wasn't merely diplomatic language but economic necessity. Developer trust was fragile, alternatives like GitLab beckoned, and the acquisition faced fierce scepticism from the open source community that viewed Microsoft as fundamentally hostile to their values.
Seven years later, that calculation has reversed. Independence, once GitHub's greatest asset, has become a competitive liability in the age of artificial intelligence.
The promise that worked
Microsoft's original commitment was explicit and public. "GitHub will retain its developer-first ethos and will operate independently to provide an open platform for all developers in all industries," the company declared in its 2018 acquisition announcement. This wasn't corporate theatre—it reflected genuine strategic thinking about how to manage a culturally sensitive acquisition.
The approach succeeded brilliantly. Under Microsoft's ownership but independent operation, GitHub expanded from 28 million developers to over 150 million whilst growing from 85 million repositories to more than 1 billion. The platform launched GitHub Actions, deepened enterprise integration, and maintained its position as the world's dominant code hosting service.
Nat Friedman, who led GitHub as CEO from 2018 to 2021, embodied this balanced approach. A respected open source veteran, Friedman navigated the delicate task of integrating GitHub with Microsoft's enterprise capabilities whilst preserving developer trust. When Thomas Dohmke succeeded him in November 2021, the independence model appeared vindicated.
Yet Dohmke's departure announcement this week revealed how dramatically circumstances had shifted. "My startup roots have begun tugging on me and I've decided to leave GitHub to become a founder again," he wrote. More tellingly, he confirmed that GitHub would now operate "as part of Microsoft's CoreAI organization."
No replacement CEO will be named. After seven years of independence, GitHub becomes simply another Microsoft division.
The competitive storm
The catalyst isn't internal corporate dynamics but external pressure that reached crisis levels with startling speed. The AI coding tools market has exploded, creating existential threats to established players that seemed unassailable just months ago.
Consider Cursor's trajectory. The AI-powered code editor, developed by startup Anysphere, reached $200 million in annual recurring revenue by March 2025. By July, that figure had more than doubled to over $500 million. This isn't gradual market evolution—it's disruption compressed into months.
Cursor succeeds because it was built AI-first rather than retrofitting intelligence onto existing tools. Developers praise its seamless integration of large language models into coding workflows, creating an experience that feels fundamentally different from traditional IDEs enhanced with AI features.
Anthropic's Claude Code presents another significant challenge. Claude Sonnet 4's 72.7% performance on the SWE-bench coding benchmark outpaces many competitors, and developers consistently describe Claude's coding capabilities as superior to other models. When leading AI coding platforms like Cursor rely heavily on Anthropic's models, Microsoft faces the uncomfortable reality that its competitors depend on a third party for their most advanced capabilities.
Microsoft's response through GitHub Copilot has been commercially successful but strategically defensive. The tool now serves 20 million users and accounts for over 40% of GitHub's revenue growth, with 90% of Fortune 100 companies adopting it. Yet this success contains a troubling irony: GitHub Copilot increasingly relies on Anthropic's Claude models for advanced features, creating dependency on a competitor's technology.
Data as the new oil
GitHub's true value transcends its $2 billion annual revenue or even its massive user base. The platform hosts humanity's largest corpus of human-written code—an irreplaceable resource for training AI models that becomes more valuable with each repository added.
This data advantage explains Microsoft's formation of CoreAI in January 2025 under Jay Parikh, formerly Meta's global head of engineering. Satya Nadella positioned the new division to "build the end-to-end Copilot & AI stack for both our first-party and third-party customers," explicitly framing GitHub's future through an AI lens.
The strategic shift is profound. Rather than viewing GitHub as a neutral platform that happens to benefit Microsoft, CoreAI treats it as a core component of Microsoft's AI infrastructure. This isn't subtle integration—it's explicit repositioning of one of the world's most important developer platforms as an AI company asset.
The implications extend beyond corporate organisation charts. GitHub's long-cultivated reputation for platform neutrality becomes complicated when the platform serves dual roles as independent service and AI training ground. Developers uploading private repositories now contribute to datasets that may enhance Microsoft's competitive position against their preferred tools.
The independence paradox
The broader pattern extends across the technology industry. Google recently acquired Windsurf's leadership team whilst Cognition absorbed the remaining staff, demonstrating how AI competition drives consolidation. Amazon's CodeWhisperer competes directly with GitHub Copilot, whilst countless startups chase developer mindshare with AI-powered alternatives.
Each company recognises that developer tools represent more than productivity software—they're gateways to AI adoption across enterprises and foundations for future competitive advantage. In this context, maintaining platform independence becomes strategically untenable.
Microsoft's experience with GitHub illustrates the limitations of the LinkedIn model for AI-era acquisitions. LinkedIn could operate independently because social networking remained adjacent to Microsoft's core business. AI eliminates this adjacency. Every platform becomes a potential data source, every tool a component of the AI stack, every user relationship a competitive advantage.
The speed of transformation is remarkable. Dohmke appeared on The Verge's Decoder podcast just weeks before announcing his resignation, suggesting decisions were made rapidly rather than through careful long-term planning. When startups can achieve $500 million revenue runs within months, established players cannot afford gradual strategic evolution.
What developers inherit
The practical implications for GitHub's 150 million users will emerge gradually but predictably. Expect deeper Azure integration, more prominent Microsoft AI tool positioning, and potential restrictions on features that compete with Microsoft's offerings.
Enterprise teams should prepare for tighter Azure workflows whilst monitoring policy changes around data usage and AI feature integration. Organisations managing sensitive code should evaluate GitHub governance policies recognising the platform's evolution from neutral host to AI infrastructure component.
The developer community faces a more fundamental challenge: adapting to a world where truly neutral platforms become economically unsustainable. The promise of maintaining platform independence whilst pursuing AI supremacy appears inherently contradictory.
The new acquisition reality
GitHub's transformation signals the end of big tech's independence promises for strategically important acquisitions. Future deals will likely abandon such commitments, acknowledging that AI competition requires full integration of subsidiary capabilities.
This represents more than corporate strategy evolution—it reflects AI's impact on competitive dynamics across technology. When your primary advantage depends on data access and model training, you cannot afford arms-length relationships with critical assets.
For developers, this creates both opportunities and constraints. Enhanced AI capabilities and deeper tool integration may improve productivity, but platform choice and genuine neutrality seem destined to diminish. The challenge lies in preserving the collaborative innovation that made platforms like GitHub valuable whilst accepting their transformation into competitive weapons.
The AI era demands new models for platform governance and competitive strategy. GitHub's journey from independent acquisition to Microsoft AI division marks the beginning of this evolution, offering a preview of how artificial intelligence reshapes the relationship between big tech and the platforms that serve entire industries.
Dohmke's vision of "one billion developers enabled by billions of AI agents" may prove prescient. Whether those developers retain meaningful platform choice or become subject to the strategic imperatives of their AI-powered tools remains the crucial question for the industry's future.