Anthropic Expands Claude Enterprise AI Security Features
Anthropic introduced two new features for Claude Managed Agents during its “Code with Claude” event in London. At the same time, the company expanded its enterprise AI strategy through a global partnership with KPMG.
The announcements focus on helping enterprises run AI workloads securely within private infrastructure while maintaining centralized AI orchestration.
New Managed Agent Security Features
Anthropic launched MCP tunnels and self-hosted sandboxes to improve enterprise-grade security and infrastructure control. As a result, organizations can keep sensitive systems and internal data inside private environments while still using Claude Managed Agents.
MCP tunnels allow Claude agents to connect securely with Model Context Protocol servers inside corporate networks. Moreover, the system uses a lightweight gateway that creates a single outbound connection without requiring inbound firewall changes or public endpoints. The encrypted setup gives agents secure access to internal APIs, databases, ticketing platforms, and enterprise knowledge systems.
Meanwhile, self-hosted sandboxes shift tool execution into infrastructure controlled by the enterprise. However, Anthropic continues to manage orchestration, context handling, and error recovery. The public beta launch supports Cloudflare, Daytona, Modal, and Vercel environments, while companies can also deploy their own sandbox clients.
KPMG Partnership Expands Claude Deployment
Anthropic also formalized a global alliance with KPMG to integrate Claude into the firm’s client delivery operations. Consequently, all 276,000 KPMG employees worldwide will gain access to Claude’s AI tools.
The partnership introduces “Digital Gateway Powered by Claude,” which integrates Anthropic’s AI capabilities into KPMG’s Microsoft Azure-based delivery platform. Initially, the rollout will focus on tax clients and private equity firms. Additionally, the deployment aims to support agentic workflows, compliance operations, advisory services, and AI-powered enterprise products.
The broader expansion reflects rising competition in the enterprise AI market. Therefore, AI vendors increasingly focus on data residency, infrastructure control, and secure deployment models for regulated industries. Anthropic’s latest moves also strengthen its position against rivals pursuing similar partnerships across consulting and professional services sectors.
