AI agents are taking charge, but who’s managing the managers? The race to build safer, smarter, and fully controlled AI agents is heating up. Startups and investors are pouring millions to solve one key problem: How do businesses safely “parent” AI agents that act on their behalf?
Runta Raises $20 Million to Parent AI Agents Safely
Runta just closed a $20 million seed round led by a16z, pushing the company’s valuation past $100 million. That’s a big deal for a startup focused on something crucial—rebuilding the infrastructure where AI agents actually run. Runta’s founder, Guanlan Dai, aims to create a full operating system for agents, offering a new level of control and safety.
What does “parenting” AI agents mean? Runta isolates each AI agent and sets strict guardrails. This keeps agents from running wild or causing damage. The startup wants businesses to watch over their AI like a responsible parent, making sure every digital assistant stays on task and within limits.
Martin Casado, a16z partner, sees this as foundational. “We’re rebuilding the layer where agents actually run,” he says. This is the core of safe AI autonomy—providing the operating system that governs agent behavior.
Security Is the Name of the Game for AI Agents
AI agents can behave unpredictably. They might ignore guardrails or rewrite permissions. This makes security a top priority. Pulse Security, a fresh startup, raised $8 million in seed funding to tackle this challenge. Led by Foundation Capital and Zetta Venture Partners, Pulse Security’s co-founders Mike Armistead, Robert Hipps, and Nick Gilligan bring deep experience from companies like Google and Mandiant.
Pulse’s platform delivers program intelligence and agentic infrastructure tailored for security leaders. It integrates with existing security tools and acts as a “system of record” for managing security programs. Brett Wahlin sums it up: “Security teams often get lost in technical complexity, but CISOs must focus on business risk. Pulse transforms daily operations into meaningful, business-driven intelligence.”
Sid Trivedi from Foundation Capital adds, “Pulse’s agentic platform gives security leaders the strategic command they have been missing.” The goal is to provide CISOs a clear picture of risks and defenses in a world where agents act autonomously.
Zero Trust Principles Drive AI Agent Security
Andre Durand, CEO of Ping Identity, stresses zero trust as the future of AI agent security. “Zero trust really just says, just enough, just in time,” he explains. That means every agent action must be verified continuously—not just once at login.
Durand insists each AI agent must have its own identity. “It should not be impersonating the human,” he says. Agents can act on behalf of humans, but the line between human and agent actions must stay clear.
Shared secrets like API keys are risky and outdated. Instead, secure authentication methods tailored to agents are needed. Policy enforcement points—like API gateways—inspect every request and apply strict rules. “The goal is to move authorization from once at login to every consequential action,” Durand says.
He warns that AI agents can rewrite their permissions or ignore guardrails. This unpredictability demands frameworks that combine automated review with human accountability. “We probably will have to develop frameworks that we trust without verifying the output directly. It’s the best way to move at agent speed,” Durand notes.
Security leaders must cover the full lifecycle of agent management: discovery, registration, and policy enforcement. Durand challenges the industry: “Pause long enough to see the totality of what it would mean to secure multiple agents, both interacting with you from the outside as well as being deployed on the inside.”
Other Movers in AI Agents and Robotics
Alterion, another startup, announced Draco on July 16, 2026. Draco is a runtime control platform designed to govern AI agents in production, adding another layer of governance and control.
Meanwhile, Microagi raised an eye-popping $55 million seed round—the largest ever in Germany. Microagi focuses on robotics data. Since May 2026, it has offered a New York service recording customer data in exchange for free apartment cleanings. The startup licenses some data exclusively to frontier AI labs and helps shape the “brains” of robots used in factories.
Microagi buys hardware from Chinese robot-makers Unitree and UBTech and leases it to factories, feeding data back to improve AI-driven robotics on production lines. Bercan Kilic, Microagi’s founder and CEO, is building a new kind of robotics ecosystem powered by real-world data.
The AI Agent Revolution Is Just Beginning
AI agents are no longer sci-fi. They’re here and evolving fast. But they bring new risks and challenges. Runta, Pulse Security, Alterion, and Microagi are each building critical pieces of the puzzle. From infrastructure to security to real-world robotics, they’re shaping how AI agents safely integrate into business and daily life.
Will these new tools keep AI agents in check? The stakes could not be higher. The future belongs to the companies that can parent, govern, and secure AI agents at scale. And right now, that future is unfolding before our eyes.
Based on
- a16z is backing Runta, a startup that wants to ‘parent’ your AI agents — thenextweb.com
- Alterion Launches Draco, a Runtime Control Plane for Enterprise AI Agents | AP News — apnews.com
- Pulse Security Debuts Operational Management Platform Built for Security Leaders | Currency News | Financial and Business News | Markets Insider — markets.businessinsider.com
- Zero trust must now move at agent speed | VentureBeat — venturebeat.com
- Robot-data startup Microagi raises largest seed round ever in Germany | Semafor — semafor.com
