AI coding tools are now the most contested market in enterprise software. Claude Code, Codex, Project Polaris, Antigravity, Cursor, Kiro all competing

[Nina81]

CNBC published analysis this morning noting that Microsoft and Google are taking on Anthropic and OpenAI in the AI coding market that Anthropic currently leads through Claude Code. Microsoft's Project Polaris announced yesterday at Build. Google's Antigravity 2.0 with Gemini 3.5 Flash. Amazon's Kiro. The seven-way competition is intensifying.

The strategic importance is clear: whoever captures developer daily workflow captures the enterprise relationship. Developers choose their tools and those choices become procurement decisions for the organisations they work in.

Microsoft and Google are late to AI coding, but 'absolutely critical' they compete for growth

https://www.cnbc.com/2026/06/01/microsoft-and-google-take-on-anthropic-and-openai-in-ai-coding-models.html

[NightHarbour]

The Microsoft angle is specifically interesting because GitHub gives them 150 million developers who already have Copilot in their editor. Distribution advantage before the model competition even starts

[StoneCold]

Project Polaris being Microsoft's own model is the hedge against Claude Code's success. If Anthropic is winning on coding capability, Microsoft builds its own rather than continuing to pay for OpenAI's

[KeyboardWarrior]

Enterprise IT departments want to approve one or two tools. The race right now is to be one of the approved tools before the list closes. After that, switching costs make displacement very hard

[TristanFenwick]

The developer who chose Cursor over Copilot a year ago and built their workflow around it is not switching to Polaris because Microsoft built a new model. Habit and integration depth matter more than benchmark scores at that point

[ParallelSelf90]

Antigravity 2.0 being 4x faster on output tokens is the differentiator for agentic tasks where iteration speed determines how productive a session feels