Building a scalable foundation for AI’s future

Every generation of computing faces the same pivotal choice: build walls or build bridges. When the internet was young, proprietary networks promised control and predictability—those systems are now museum exhibits. Open protocols succeeded by enabling global collaboration and innovation. When cloud computing emerged, open standards like Kubernetes won by creating an ecosystem where thousands of companies could build, innovate, and scale together. Now, as Agentic AI transforms computing once again, we face the same choice: build it open, or build it twice.

Why we’re at a crossroads

Today's infrastructure was built for humans sharing content like websites, emails, and videos. But artificial intelligence (AI) agents are fundamentally different. They don't just share information; they collaborate, reason, and take autonomous actions in real-time. More importantly, they introduce probabilistic computing into our technological foundation. This shift from deterministic to probabilistic computing represents a profound change in how we build and deploy technology. 

The real question isn't whether to build new infrastructure—it's whether we'll learn from history or repeat it. 

Consider two real-world examples emerging today: 

In enterprise IT, deploying a sales forecasting SaaS platform requires collaboration across multiple AI agents. A planner agent decomposes business requirements into specific build tasks, while ServiceNow agents coordinate with Salesforce and Hubspot systems for deployment. Infrastructure agents from Cisco and Microsoft simultaneously validate security, identity and access, costs, and SLO compliance. This all requires sophisticated agent-to-agent reasoning and without open standards, each of these interactions becomes a custom integration project. 

In drug discovery, for example, the complexity amplifies. AI agents analyzing protein structures collaborate with robotic systems running wet lab experiments. Each experiment triggers a cascade of agent interactions such as safety verification, compliance checks, and stability analysis. Building these interactions on closed systems means perpetually reinventing the wheel. 

These aren't hypothetical scenarios—they're happening now. But our current infrastructure forces these interactions through manual, inefficient, and potentially insecure channels. Organizations are already discovering that closed agentic software does not scale. We need something better. 

The open alternative: Internet of Agents 

We envision an open, interoperable, and quantum-safe Internet for agent-to-agent collaboration between AI agents across organizational boundaries. We're calling it the Internet of Agents. This isn't just another layer of infrastructure–it's the foundation for the next era of computing.  

This new layer will sit above today's cloud infrastructure, providing: 

• Agent discovery and authentication: Allowing AI agents to find and verify each other's capabilities
• Secure state exchange: Enabling agents to share complex context and reasoning
• Probabilistic intent resolution: Supporting non-deterministic goals and outcomes
• Cross-domain collaboration: Facilitating cooperation between agents from different vendors and domains
• Quantum-safe security: Future-proofing against quantum computing threats 

Why open standards matter 

Throughout my career building infrastructure for emerging technologies, one truth stands out: the most transformative innovations happen when we build open foundations that enable collaboration. AI agents can't transform the world if they can't talk to each other. Today's fragmented infrastructure is holding back a revolution in computing. Just as the Internet proved 40 years ago and cloud computing demonstrated 20 years later, we now face a similar pivotal moment with AI. 

While Cisco helped establish Transmission Control Protocol/Internet (TCP/IP) and Border Gateway Protocol (BGP) protocols that became the backbone of human-to-human internet connection, the Internet of Agents requires something bigger—it needs the entire technical community coming together to establish open standards for our AI-enabled future. The path forward is clear: no single company can or should control how AI agents communicate. We need community-driven standards, reference implementations, and shared infrastructure components that enable secure agent collaboration while preserving organizational autonomy. 

Open systems create maximum value for everyone in the ecosystem. Infrastructure vendors can innovate faster and reach broader markets. Developers gain the flexibility to create and customize solutions without vendor lock-in. Operators benefit from consistent tools and practices across different platforms. Most importantly, customers get access to the best solutions for their specific needs, regardless of who created them. 

“Organizations are already discovering that closed agentic software does not scale. We need something better.”

Just as open source transformed software development, open standards will accelerate agentic innovation by enabling developers worldwide to build on common foundations. As agentic software reshapes how we work and create, we face a critical opportunity to build the Internet of Agents. The next wave of AI innovation won't come from isolated systems, but from ensembles of agents collaborating together to solve complex problems. 

Building the future together 

This is a watershed moment for technology infrastructure. The transition to agent-based computing (where AI agents communicate and collaborate autonomously) will transform every industry, from software development to scientific research to physical manufacturing. Analysts project this market could reach trillions in the next five years.  But these projections assume we build the right foundation. Companies betting on closed systems today will likely find themselves rebuilding tomorrow—just as many did during the internet and cloud transitions. 

To share our complete vision, we've published a whitepaper on the Internet of Agents that explores: 

• The three-layer architecture needed for agent-to-agent collaboration
• How quantum-safe security must be built in from the start
• Real-world examples from IT operations to drug discovery
• Critical areas where industry collaboration is essential
• Technical requirements for secure agent collaboration 

We've already started working with industry leaders to develop reference implementations and contribute to open standards. But building this future requires broader collaboration. We need developers, enterprises, researchers, and technology providers to join us in creating the foundation for AI's future. 

The choice is yours. You can build it open now, or build it twice later. The Internet of Agents is happening. The only question is whether you'll help shape it or be forced to catch up to it. 

Read the whitepaper: The Internet of Agents

Let's build it right. The first time.