Claude for Government Is Now Generally Available — FedRAMP Cleared for All US Agencies

Updated September 2026: Anthropic has also opened self-hosted Claude Code environments to public beta for Team and Enterprise plans (see the new section below). Claude for Government has moved from its July 2026 public beta to general availability. All US federal, state, and local government agencies can now access the service — it is no longer limited to invite-only participants. Federal agencies procure through Carahsoft via the GSA Schedule or by contacting Anthropic directly at claude.com/solutions/government.

Claude for Government is Anthropic’s FedRAMP High authorized product that brings Claude Code and Claude Cowork to US government agencies. Originally launched as an invite-only public beta on July 12, 2026, the product is now generally available to all eligible agencies across the federal, state, and local levels.

What Anthropic Launched — and What’s Changed

On July 12, 2026, Anthropic launched the public beta of Claude for Government, making Claude Code and Claude Cowork available through a FedRAMP High authorized environment. The product is now generally available and no longer invite-only — US federal, state, and local government agencies can purchase access directly through Anthropic’s sales team or via Carahsoft using the GSA Schedule.

FedRAMP High is the most rigorous tier of the Federal Risk and Authorization Management Program — the US government’s cloud security framework. Achieving it requires extensive third-party security audits and typically takes 12 to 18 months to complete. Systems carrying this certification are cleared to handle Controlled Unclassified Information (CUI), which covers a broad range of sensitive federal data.

The product bundles two distinct tools. Claude Code is Anthropic’s agentic coding assistant, capable of reading codebases, writing and editing files, running terminal commands, and executing multi-step development tasks autonomously. Claude Cowork is a file-based knowledge work tool that lets staff delegate document creation, RFP reviews, casework, and presentation drafts directly to Claude on their desktop.

According to Anthropic’s official blog post: “With Claude Code, public sector teams can build and modernize the software systems that underpin public services. Claude Cowork works directly with files on the desktop, allowing agency staff to delegate memo creation, RFP reviews, casework, and decks to Claude.”

Government-Specific Security Controls

The product ships with several controls designed specifically for the government context. Conversation history stays local on agency-managed devices rather than syncing to Anthropic’s cloud infrastructure — a key requirement for agencies handling sensitive information. Every administrative action is recorded in a hash-chained audit log, which makes it tamper-evident and suitable for compliance and oversight purposes.

Sensitive operations on Anthropic’s side require two-person approval, an additional safeguard against insider risk or unauthorized access. Agencies can also set department-level spending limits. Deployment goes through standard agency Mobile Device Management (MDM) platforms, and critically, agencies do not need a separate cloud-provider relationship — Anthropic is the contracted party directly.

This last point matters operationally. Prior government AI deployments through Microsoft’s Azure Government or AWS GovCloud required agencies to maintain a relationship with the cloud provider as well as the AI vendor. The Claude for Government model simplifies that procurement chain.

Self-Hosted Claude Code Environments: Public Beta for Team and Enterprise

Anthropic opened self-hosted Claude Code environments to public beta on August 6, 2026, for organizations on Claude Team and Enterprise plans. Sessions started from the web, mobile, desktop, or a scheduled routine run on the company’s own infrastructure — inside its network, next to internal services — instead of on Anthropic-hosted compute. The feature is off by default and is not available to organizations using zero data retention (ZDR).

The data boundary is specific. According to Anthropic’s announcement, repository checkouts, build artifacts, secrets, and any files a session creates or modifies stay on infrastructure the customer provisions. The conversation itself — prompts, responses, and tool results, which can include code Claude reads — is still sent to Anthropic for inference, and the session transcript is stored so it can be resumed from any surface. Teams evaluating this for compliance reasons should read that line carefully: source code at rest stays home, but code Claude reads during a session does not.

Deployment works through runners: long-lived processes that pick up sessions and start a Claude Code process for each one. Runners come in two modes. Fixed keeps a set number of runners running and distributes sessions across them. On-demand uses an orchestrator that watches for queued sessions, starts a runner as sessions arrive, and stops it when work finishes, so capacity tracks demand. A runner can serve multiple sessions, but each session runs in its own checkout, keeping work isolated between developers and accounts.

Anthropic is candid that this is not the default path. Its blog post states: “For most enterprises, we strongly recommend our hosted offering for operational simplicity with no infrastructure to run or maintain.” Self-hosting is aimed at teams whose network, tooling, or compliance requirements call for keeping agent execution on infrastructure they control, and Anthropic advises that a platform or developer-productivity team own setup and ongoing operation — building and maintaining the runner image, updating runners, and running the orchestrator in on-demand mode. George Jacob, Senior Engineering Manager at Faire, said in the announcement that the setup lets Claude “generate PRs, help fix CI issues, and respond to developer workflow events, with compute that can scale based on demand.”

Self-hosted environments differ from Remote Control, which lets a developer continue a session running on their own machine from a phone or browser. Remote Control sessions end when that machine stops running them and are tied to the user who ran claude; self-hosted sessions run on shared infrastructure operated by a platform team and can be used by any user in the organization. For agencies weighing Claude for Government, the combination is notable: FedRAMP High covers the hosted service, while self-hosted runners give teams with stricter network boundaries a way to keep code checkouts and secrets on their own servers. Teams comparing agentic coding options can see how Claude Code stacks up against Cursor, Codex, and others in our guide to the best AI coding assistant in 2026.

Why This Is a Significant Move for Anthropic

This launch puts Anthropic — which closed a $6.5 billion funding round at a $61.5 billion valuation — directly into competition with Microsoft and AWS in one of the highest-value segments of the enterprise software market. Government IT contracts are multi-year, sticky, and often scale to hundreds of millions of dollars.

Prior to this announcement, AI in government was largely routed through OpenAI’s products via Microsoft’s Azure Government offering, or through AWS Bedrock’s models on GovCloud. Anthropic entered this market directly with agentic tools rather than basic chat — a distinction that matters for productivity use cases in software development and document work.

The timing also aligns with active pressure from the Trump administration in early 2026 to accelerate AI adoption across federal agencies. Agentic AI security has become a central concern as agencies weigh adoption, and Anthropic’s audit trail and two-person-approval design directly addresses those concerns. Agencies can access the program at claude.com/solutions/government.

Anthropic already had government traction before this — the Government of Alberta previously used Claude to identify cybersecurity vulnerabilities. The FedRAMP product formalizes and scales that government presence into a full commercial offering.

Frequently Asked Questions

Sources: Anthropic official blog — Bringing Claude Code and Claude Cowork to Government | Claude for Government Help Center | Anthropic — Run Claude Code sessions on your own compute

Updated: September 10, 2026