Antigravity versus Cursor — the bundling pivot when the IDE meets the agent platform
Google's Antigravity CLI and SDK ship in the same week that Anthropic doubles Claude Code limits and removes peak-hour throttling. The agent-tool market is now tri-vendor — Cursor at the IDE layer, Claude Code at the senior-developer tier, Google Antigravity at the infrastructure-and-platform tier — with GitHub Copilot's AI Credits move and the open-weight stack as fourth-and-fifth positions. The bundling pressure on each is what shapes the next 18 months.
The structural state is a four-vendor market with one fifth wildcard. Cursor with Composer 2.5 dominates the mid-market IDE-plus-model tier — $20/month flat with in-house-model economics that don't require frontier-lab API spend. Claude Code with doubled limits and removed peak-hour throttling dominates the senior-developer cohort — 46% preference in JetBrains' May survey for developers with 10+ years of experience. Google Antigravity CLI in Go with public SDK dominates the infrastructure-and-platform tier — Go-language toolchain, multi-agent orchestration, $1.50/$9 token pricing. GitHub Copilot under AI Credits flex billing defends its installed-base position on usage-based pricing. The open-weight self-hosted stack (Continue, Aider, Cline, Tabnine with open models) is the fifth wildcard, gaining adopters who want full data-residency control.
The bundling pressure on each is what determines the next 18 months. Cursor's bundling is IDE-plus-model — the customer pays once for the IDE and gets the model included. Anthropic's bundling is Claude-platform-wide — Claude Code, Managed Agents, MCP tunnels, the Anthropic API all share infrastructure and authentication. Google Antigravity's bundling is Workspace-plus-GCP — the customer's existing Google identity, billing, and infrastructure relationships absorb Antigravity adoption frictionlessly. Microsoft GitHub Copilot's bundling is GitHub-plus-Office — the same Microsoft account, the same billing relationship, integration with Teams and Office productivity surfaces. Each bundle is real; each pulls customers toward a different default.
The competitive question is which bundle wins which customer segment. Mid-market shops without strong existing-vendor commitments default to Cursor — cheapest, most capable IDE-plus-model surface, no organizational dependency. Senior-developer-heavy shops (especially in financial services, AI labs, security-conscious enterprise) default to Claude Code — the model the senior developers prefer, supported by Anthropic's regulated-industry trust story. GCP-heavy shops default to Antigravity — the integrated stack with their existing infrastructure. Microsoft-heavy shops default to GitHub Copilot — the path of least resistance for organizations already paying Microsoft for everything else. The segmentation is durable; cross-segment switching is friction-heavy.
The economic pressure on Anthropic and OpenAI is mixed. On the one hand, Cursor's in-house model means one less customer for Anthropic and OpenAI's frontier-tier API spend. On the other hand, Anthropic's $30B raise from the AM cycle implies the lab does not need to defend every developer-cohort customer — the senior-developer tier (Claude Code's preferred-by-46% cohort) is the cohort that drives the revenue flywheel, and Cursor competes at the tier below. OpenAI's response is similar via ChatGPT distribution and the Realtime API — different developer segment than Cursor competes for.
The longer-arc question is whether the four-vendor market is stable or whether bundling pressure collapses one or more competitors over 18-24 months. The historical pattern in adjacent markets (CRM, productivity, video conferencing) is that bundling pressure consolidates to 2-3 vendors after 3-5 years of fragmentation. The AI coding tool market is currently in the fragmentation phase; whether it follows the adjacent-market pattern depends on whether the bundling moats (IDE, model preference, cloud integration, productivity-suite integration) prove durable or whether one of them collapses (e.g., if Google's Workspace bundling pulls Microsoft shops faster than expected). Watch the Q4 2026 procurement data for the early signal.
The line: the agent-tool market is mature enough to have four competitors but young enough that the segmentation is not yet locked in. The next 18 months decide.
Nipralo — Best AI Coding Tools 2026 → · Lushbinary — AI Coding Agents Comparison 2026 → · SitePoint — Claude Code vs Cursor vs Copilot 2026 →