Agents Are Infrastructure Now
Google and Anthropic shipped managed agent compute in the same week - agents are a cloud primitive now.
Agents stopped being tools you prompt and became infrastructure you provision. Google and Anthropic proved it on the same day.
What Shipped
On May 19, Google launched Managed Agents in the Gemini API. A single API call spins up an agent with its own isolated Linux sandbox, reasoning loop, tool access, web search, and file system. The agent executes, returns results, and the sandbox tears down. No orchestration code. No infra management. One call.
The same day, Anthropic shipped self-hosted sandboxes and MCP tunnels for Claude Managed Agents. Self-hosted sandboxes let the agent loop run on Anthropic's infrastructure while tool execution happens on yours - Cloudflare, Modal, Vercel, or bare metal. MCP tunnels let managed agents reach private MCP servers without exposing them to the public internet. One outbound connection, no firewall rules, end-to-end encryption.
Two companies. Same week. Same conclusion: agents need managed compute, not just managed models.
The Architecture That Matters
Google's approach is fully managed. The Antigravity harness powers every surface - desktop app, CLI, SDK, IDE - and now powers the API too. Each managed agent gets a sandbox with no network access by default. You send a task, the agent reasons and acts, you get results back. The compute boundary is Google's problem.
Anthropic's approach splits the plane. Orchestration stays managed. Execution goes wherever your data lives. The agent reasons in Anthropic's cloud but runs tools inside your security perimeter. This is the hybrid model enterprises actually need - managed intelligence, self-hosted execution.
Both are the same architectural pattern: agents as a provisioned compute primitive. Like containers, but the unit of deployment is a reasoning loop instead of a process.
What Changes for Builders
I have been building agent systems where I manage the sandbox, the tool registry, the error recovery, and the execution lifecycle myself. That entire layer just became someone else's problem.
The shift mirrors what we saw with containers. First you ran processes on VMs you managed. Then Docker abstracted the runtime. Then Kubernetes abstracted the orchestration. Then the managed cloud ate all of it. Agents are following the exact same path - compressed into months instead of years.
The unit of work is no longer "call a model." It is "provision an agent." Budget it, scope it, tear it down.
Where This Heads
Agent provisioning will become as routine as container orchestration. The developers who treat agents as infrastructure - with resource limits, isolation policies, and lifecycle management - will be the ones shipping reliably. The rest will still be writing prompt chains.