AMAZON RESEARCHERS USE ANTHROPIC AI TO FIND CYBERSECURITY WEAKNESSES

That is a mistake.

The United States will not win the next technological era simply because it develops the smartest algorithms. It will win because it builds and controls the infrastructure that allows those algorithms to operate at scale.

When I think about artificial intelligence, I think about four pillars: power, compute, connectivity, and distribution.

Most of the conversation centers on compute. That is where the money is flowing and where many of today’s largest AI companies operate. Compute is important, but it is only one piece of a much larger system.

Artificial intelligence requires electricity. It requires communications networks. It requires physical infrastructure capable of moving information across the globe. Without those things, even the most advanced model remains little more than software sitting inside a data center.

America’s long-term advantage will depend on more than who writes the best code.

It will depend on who controls the bottlenecks.

History offers a useful lesson.

The industrial age was not won by inventors alone. It was enabled and won by those who built railroads, power systems, factories, and supply chains. The internet era was not won simply by creating websites. It was won by building fiber networks, cloud infrastructure, and semiconductor manufacturing capacity.

The AI era will follow the same pattern.

As a former SpaceX leader who led Commercial Crew and Starlink initiatives, I have watched many policymakers and investors continue to view infrastructure companies through outdated lenses.

SpaceX is a useful example.

Most people still think of SpaceX as a rocket company. Rockets are certainly part of the business, but they no longer tell the full story.

Starlink has become one of the largest communications networks in the world. It provides connectivity to rural communities, military operations, disaster response efforts, maritime users, and regions where traditional infrastructure struggles to reach.

In an increasingly AI-driven economy, connectivity becomes more valuable, not less.

The same is true for distribution.

Much of today’s AI discussion assumes that intelligence itself will be the most valuable product. That may prove correct. But intelligence still must be delivered to users, businesses, governments, and consumers. Infrastructure determines whether that happens efficiently and at scale.

That reality should shape both public policy and private investment.

The United States should absolutely support AI innovation. But innovation without infrastructure is not a strategy.

Washington should spend as much time discussing energy generation, communications networks, manufacturing capacity, and infrastructure deployment as it spends discussing the latest AI model.

The countries that dominate the next technological era will not simply be the ones that invent the best tools.

They will be the ones that build the systems required to power, connect, and distribute those tools to the world.

The AI race is not just a software competition.

ANTHROPIC BLOCKS MOST ADVANCED AI MODELS AFTER US BARS FOREIGN USE

It is an infrastructure competition.

America should act like it.

Vincent Peters is an AI researcher, entrepreneur, and former SpaceX leader. He holds graduate certificates in Artificial Intelligence from Oxford and Blockchain Technologies from MIT Sloan and studied engineering at West Point.