In 1992, a seemingly small moment would mark the end of one of technology’s greatest battles. Vint Cerf, wearing a T-shirt that read “IP on Everything,” stood before a room of engineers. That simple statement represented victory in what historians now call the Protocol Wars—a decades-long fight over how computers would talk to each other.
The stakes were clear. Telecommunications giants and computing behemoths pushed for closed, proprietary standards. They argued this would ensure reliability, security, and profit. On the other side stood researchers and smaller companies advocating for open protocols that anyone could build upon. Their argument? Innovation happens faster in the open. More importantly, it creates foundations that everyone—not just those who can afford expensive licenses—can build upon.
This wasn’t just a technical debate—it was a philosophical one. A small networking company called Cisco bet big on the open approach, helping develop software that could discover, authenticate, and connect computers using open standards. At the time, this seemed foolish. Industry giants like IBM were pushing proprietary protocols and conventional wisdom suggested they would win.
But they didn’t. TCP/IP, the open protocol suite, emerged victorious not because it was perfect—it wasn’t—but because its openness enabled continuous innovation and improvement. Anyone could build on it, improve it, adapt it. The result was an explosion of creativity that gave us the internet as we know it today.
HOW OPEN STANDARDS HAVE DRIVEN INNOVATION
This pattern repeated itself twice in the cloud era. First, Linux became the foundation for cloud compute servers, outpacing proprietary systems through open-source innovation and community development. Then in 2014, as containerization grew, the industry needed a way to manage applications at scale.
While proprietary solutions competed for control, Kubernetes was released as an open-source project and transformed cloud infrastructure. Within just three years, this open approach had become the industry standard for container management, proving once again that open beats closed when it comes to infrastructure.
THE CHALLENGE AHEAD
Now, in 2024, we face perhaps the most consequential standards battle yet—one that will determine how artificial intelligence evolves.
In the next few years, the average company could deploy over 20,000 AI agents—specialized programs that can reason, learn, and take autonomous actions. These agents will need to collaborate across organizational boundaries to solve complex problems.
But our current infrastructure wasn’t built for this kind of collaboration. Today’s internet was designed for humans to share content like websites, emails, and videos. AI agents need something fundamentally different. They must share complex states, reason about uncertainties, and make coordinated decisions at machine speed.
BUILDING WALLED GARDENS OR AN OPEN ECOSYSTEM?
Once again, we’re seeing the same debate play out. Some are building closed, proprietary systems for AI collaboration. But history suggests a different path. Just look at how open source transformed cloud computing: What started as community projects like Linux and Kubernetes are now power trillion-dollar businesses.
The next layer of infrastructure—what we are calling an “Internet of Agents”—must be built on the same principles that made the original internet and cloud computing successful: openness, interoperability, and security by design.
Think about your organization’s future: Will your AI agents need to work with partner AI systems? Will they need to tap into existing software or customer data? Every closed system you build today becomes a barrier tomorrow. The real question isn’t whether to build open or closed systems—it’s whether to build it once with open standards or twice when closed systems fail.
Three key questions to consider:
- How will your AI systems grow beyond your walls?
- Where are you accidentally creating tomorrow’s silos?
- Is your infrastructure ready for secure agent collaboration at scale?
WHY AI NEEDS AN OPEN FOUNDATION
For leaders skeptical about open systems, consider the parallel with cloud computing. Just as companies that prepared early for cloud had significant advantages, organizations that build the right foundation for AI agent collaboration will be better positioned for the future.
The stakes are enormous. AI will transform every industry, from health care to manufacturing. Analysts project this market could reach trillions in the next five years. But that transformation can only happen if AI agents can collaborate effectively across every organizational, industry, and vendor boundary.
SHAPING THE COLLECTIVE FUTURE OF AI AGENTS
The Protocol Wars taught us that open standards unleash innovation in ways closed systems never can. Kubernetes showed us that community-driven development can outpace proprietary solutions. Now, as we build the foundation for AI’s future, these lessons are more relevant than ever.
Business leaders can take concrete steps now to help shape this future:
1. Contribute strategically to the foundation: Engage your development, operations, and IT teams in open-source AI collaboration projects like AGNTCY (full disclosure: This is Outshift by Cisco’s open-source initiative). Just as early contributors to internet protocols shaped decades of innovation, organizations that help build the Internet of Agents will influence AI’s evolution.
2. Map your collaboration opportunities: Examine how your existing applications and workflows could benefit from AI agent collaboration. This isn’t just about individual AI capabilities—it’s about specialized agents working together to solve complex business problems.
3. Make interoperability a competitive advantage: Rather than building proprietary walled gardens, prioritize solutions that work with emerging open standards. Organizations that embrace interoperability will create more flexible, adaptable systems that can evolve with the rapidly changing AI landscape.
STRIKE WHILE THE IRON IS HOT
The key is to engage now while these standards are still taking shape. Just as the early days of the internet created opportunities we couldn’t imagine in 1993, the Internet of Agents will enable innovations we can’t yet envision. The organizations that help build this foundation will be best positioned to take advantage of those opportunities.
We’ve seen this movie before. Closed systems promise control but deliver lock-in. Open systems unleash innovation. With AI, the stakes are too high to get this wrong. The question isn’t whether open standards will win—it’s whether your organization will help shape them or be shaped by them.