Leaving the dedicated IDE to live inside your usual editor — Google ships Antigravity to VS Code and JetBrains, and hands enterprises a "spending valve" and audit logs

Google frees its coding agent from a dedicated IDE, shipping it as an extension for VS Code, JetBrains, and Zed. At the same time, it bundles Antigravity into Gemini Enterprise and gives enterprises control through per-project budget caps and central audit logs. Concerns about lock-in and after-the-

Share
Leaving the dedicated IDE to live inside your usual editor — Google ships Antigravity to VS Code and JetBrains, and hands enterprises a "spending valve" and audit logs

Until now, Google's coding agent "Antigravity" ran only in its own workshop, so to speak: the dedicated Antigravity IDE (desktop app 2.0) and a CLI. This DeepMind-born agent has drawn attention as an "agent-first" tool that drives itself from planning through implementation and verification, but using it required developers to leave the environment they work in every day. The update Google announced for Gemini Enterprise on August 20–21 flips that premise. Antigravity has moved into the editors developers are already using.

It arrives as an extension in your usual editor

What's being shipped this time is an extension that bolts onto existing IDEs. Coverage is as follows, with a notable gap: only Visual Studio Code is generally available, while the rest are in preview.

  • Visual Studio Code: available now (general availability)
  • Visual Studio: preview
  • JetBrains: preview
  • Zed: preview

The existing desktop app (Antigravity 2.0) and the CLI remain available as well. In other words, developers can now choose between "switching to the dedicated IDE" and "calling it from their usual editor." Google explains that this extension is the result of customers asking to "be able to use Antigravity in the development tools they're already comfortable with." Agent-style tools tend to want their own container, but teams on the ground dislike the cost of switching tools — this is a compromise with that reality.

The other headline: administrator "control"

This update puts as much weight on "how enterprises hold the reins" as on distributing the extension. Antigravity is now bundled with Gemini Enterprise's Standard, Plus, and Standard Emerging Market licenses, so it can be used without a separate add-on contract. On top of that, spending and governance controls are consolidated into the admin console. Here are the key points.

AreaWhat you can do
Spend managementSet monthly budget caps per project in the billing console. Per-user and per-team controls are slated for the second half of 2026. Also supports a shared "pool" of tokens for teams, as well as capped overage usage.
Governance and securityConfigure security policies such as sandboxing workspaces and restricting access to MCP servers. Enable central audit logging with a single toggle to record prompts, agent responses, and metadata. Also supports Workforce Identity Federation (WIF) and Application Default Credentials (ADC).

What matters here are the "spending valve" and the "audit log." Left unchecked, an agent can consume tokens without limit, so a per-project cap serves as a practical brake against runaway billing. The audit log, for its part, provides a ledger for tracing after the fact who issued what instructions and what the agent returned — a move to meet the prerequisites that regulated industries and large enterprises require before deploying to production.

What changes on the business side

The practical benefit for enterprises is reduced friction in adoption. By folding Antigravity into an existing Gemini Enterprise subscription, there's no need to buy a separate license for the coding agent, and visibility into spending, security, and usage is unified in a single admin screen. Developers stay in their usual VS Code or JetBrains, while the IT department maintains control through budget caps and audit logs — a division of roles that has settled into a practical shape. Accenture and Wipro have contributed comments as adopters, suggesting that major systems integrators are taking this up in the context of building an "AI-first development organization."

Concerns worth watching behind the convenience

At the same time, critical perspectives have been raised. First is the whiff of lock-in. Extending Antigravity into existing editors while simultaneously bundling it into Gemini Enterprise is also a move that pulls development tools toward a single vendor's subscription and governance foundation. A mechanism that hands the reins to administrators is appealing, but the fact that you're entrusting those very reins — and everything with them — to Google's platform is something worth examining before adoption.

Second is the time lag: governance is being built out after the fact. The Register reports that this strengthening of controls came in response to customer requests, while also noting that it's part of a reorganization that "wires in" Antigravity — originally a standalone DeepMind product — to the main body. Against this backdrop it cites organizational turmoil such as the departure of Jeff Dean and others and Demis Hassabis's move to the chairman role, as well as competitive pressure from a new version shipping while Gemini 3.5 Pro remains unavailable. Since the enterprise management features are at a stage where they were "added on request" rather than "designed in from the start" — including the fact that per-user and per-team budget controls are planned for the second half of 2026 — it's reasonable to view the granularity of control as still a work in progress. There's no need to be alarmist, but the safe way to decide on production adoption is to actually verify that the audit logs and access controls meet your own requirements first.

From a closed, dedicated-IDE testing ground to both the editors developers use on the ground and the enterprise's management ledger — Antigravity's latest step is itself a reflection of the transitional period in which coding agents move from a "tool to try out" to a "tool operated across an organization."

References: Expanding Google Antigravity for enterprise customers (Google Cloud Blog) / Google tethers Antigravity to enterprise controls amid AI shakeup (The Register) / Google's AI coding agent just escaped its own IDE (The New Stack)

Read more

Making It Wait for "Jobs That Run Over an Hour": Codex 0.152 Adds Ceiling Dials for MCP Output Volume and Execution Time, and Turns the Planning Tool Off by Default

Making It Wait for "Jobs That Run Over an Hour": Codex 0.152 Adds Ceiling Dials for MCP Output Volume and Execution Time, and Turns the Planning Tool Off by Default

Codex v0.152.0 on August 31 and its next-day fix release added explicit ceilings on MCP tool output volume and execution time, and switched the planning tool off by default. Here's a rundown of the changes that matter for long-running unattended and semi-autonomous agent operation.

By FF
The CLI's Default Model Just Swapped In a Million-Token Brain — Claude Code v2.1.257 Makes Fable 5.1 the Standard and Adds a 'Containment Escape' Checkpoint to Auto Mode

The CLI's Default Model Just Swapped In a Million-Token Brain — Claude Code v2.1.257 Makes Fable 5.1 the Standard and Adds a 'Containment Escape' Checkpoint to Auto Mode

Claude Code v2.1.257, released September 1, 2026, swaps its default model to Fable 5.1 with its one-million-token context. It also adds guardrails to auto mode that stop credential retrieval and out-of-scope reads from slipping through. Here's a rundown of the changes that matter to developers.

By FF
"This Is an Authorized Exercise"—How the Aurora Ransomware Gang Insisted, While Making Cursor's AI Agent Do the Actual Intrusion Work

"This Is an Authorized Exercise"—How the Aurora Ransomware Gang Insisted, While Making Cursor's AI Agent Do the Actual Intrusion Work

Gambit Security and CloudSEK report that the ransomware group Aurora abused Cursor's AI agent for real intrusion work. Posing the tasks as an "authorized exercise" to slip past the safeguards, they had it handle reconnaissance and privilege takeover on the back of stolen credentials—a warning that a

By FF
One in Three Companies Now Choose to Build Rather Than Buy — McKinsey Measures How Coding Agents Are Reshaping the Procurement Decision

One in Three Companies Now Choose to Build Rather Than Buy — McKinsey Measures How Coding Agents Are Reshaping the Procurement Decision

McKinsey's annual survey found that about 30% of respondents passed on buying software because they could build it in-house with coding agents. We unpack the procurement shift from buying to building — and the current reality that productivity is up while profits stay flat.

By FF