By ByteBot
Anaconda acquired Kilo Code yesterday — the open source, model-agnostic coding agent used by over 3 million developers and processing nearly 10 trillion tokens per month. It stays MIT-licensed. Existing users see no changes to pricing, products, or support. In a week where AI coding tools keep disappearing into walled gardens, that last part is the actual news.
What Kilo Code Is
If you haven’t encountered it: Kilo Code is an agentic coding assistant that runs inside VS Code, JetBrains, and the CLI. It grew from zero to 3 million users in 16 months with no paid marketing — purely developer-to-developer recommendation. The pitch is straightforward: access 500+ AI models through a single tool, pay the provider’s rate directly with zero markup, and keep your options open.
The MIT-licensed source is on GitHub. You can read exactly what it does, audit its behavior, and fork it if you disagree with a direction. That last point matters more than it sounds right now.
Kilo ships with five specialized modes — Orchestrator, Architect, Debug, Code, and Ask — each with its own prompt policy and tool access. The Orchestrator mode is where it earns its keep: it breaks complex tasks into coordinated subtasks across planner, coder, and debugger sub-agents, handling multi-step work better than a single prompt trying to hold everything in context at once.
Why the Timing Is Significant
The Anaconda deal lands about four weeks after SpaceX agreed to buy Cursor for $60 billion in stock. That deal isn’t closed yet, but the direction is clear: Cursor, which had built its following on model flexibility, is now a SpaceX/xAI asset. Developers who depend on Cursor for their daily workflow are watching Grok integration become inevitable.
This is the consolidation that the open source crowd has been warning about. Amazon has poured $25 billion into Anthropic. Google holds an Anthropic stake. Microsoft and OpenAI are structurally entangled. Every major AI model provider is now connected to a major tech company — and the tools that route your tokens are following the same path.
“Developers aren’t going to stop using AI. The question is whether enterprises are going to own that experience end to end or just hope for the best.”
— Anaconda
What Actually Changes
According to both companies: not much, immediately. Kilo Code stays open source with no pricing or product changes. What Anaconda adds is an enterprise governance layer — the kind of auditing, access control, and compliance tooling that CTOs and CISOs need before they let AI coding agents touch production infrastructure.
This builds on Anaconda’s acquisition of Outerbounds in April 2026, which brought the Metaflow orchestration framework into their stack. The product roadmap is now: Anaconda Distribution (environment and packages) + Outerbounds (ML orchestration) + Kilo Code (AI coding agent) = an end-to-end AI development platform that an enterprise can actually govern.
That’s a defensible position. Anaconda has 30 million users in the broader Python and data science ecosystem. They know how to sell into enterprise AI teams, and they understand what those teams need that individual developers typically don’t factor in.
The MIT License as a Floor
The more important guarantee isn’t Anaconda’s promise — it’s the MIT license. Whatever direction Anaconda takes the commercial product, the current codebase is permanently forkable. The community can take it and run with it if the enterprise pivot goes too far. That’s the structural difference between Kilo and Cursor: one has a floor, the other doesn’t.
The risk worth watching is pace. Enterprise-focused engineering teams prioritize security, compliance, and stability over moving fast. If Anaconda’s priorities start pulling Kilo’s roadmap toward features that only matter at Fortune 500 scale, the open source community that built its 3-million-user base might start forking rather than waiting.
Should You Try It?
If you’re already on Cursor and comfortable there, you don’t need to jump today. But if the SpaceX deal made you want a backup plan — or you’re evaluating AI coding tools fresh — this is a reasonable one. Install the CLI:
npm install -g @kilocode/cliOr grab it for VS Code or JetBrains from their respective marketplaces. Point it at your existing API keys — OpenAI, Anthropic, Google, or local models — and it works immediately with no new subscriptions required.
You can read Anaconda’s full announcement and The New Stack’s analysis for more on the acquisition terms. The open source case for Kilo got stronger this week, not weaker. An MIT-licensed coding agent with enterprise backing and zero markup on model costs is a reasonable place to land while the rest of the market sorts itself out.
