Arcade just raised $60M because your AI agents can't prove they're allowed to do anything

Your AI agents have access to everything. And nobody can prove they're supposed to.

Arcade.dev just closed a $60M Series A led by SYN Ventures with strategic checks from Morgan Stanley and Wipro.

The problem they're solving is simple and lethal: most companies can verify an agent is what it claims to be. Almost none can prove that a given agent, acting for a given user, is authorized to perform a specific action on a specific system.

"Agents don't fail in production because the model is wrong," said CEO Alex Salazar. "They fail because nobody can prove who is authorized to do what."

That gap is why most corporate agents never leave the pilot stage.

Arcade authored the MCP authorization specification that Anthropic adopted. They're running in production at a top US bank. Tool call volume is up 25x in six months.

Morgan Stanley doesn't write $60M checks for vaporware. They wrote one because every Fortune 500 deploying agents is hitting the same wall: agents exhaustively exploit every permission they inherit. Humans are restrained by fear of being fired. Agents are not.

Audit your agent permissions today. If your agents have standing service account access with no action-level authorization, you're one hallucination away from a breach.

SOURCE: https://siliconangle.com/2026/06/15/ai-agent-authorization-startup-arcade-nabs-60m-investment/
VERIFIED: BusinessWire press release June 15 2026, SiliconANGLE June 15 2026, The Next Web June 17 2026
SIGNAL: Agent authorization is becoming the governance layer enterprises can't skip. Morgan Stanley and Wipro investing signals that financial services and IT services are building agent infrastructure now.