Every transformative technology produces two competitions: the race to build it, and the race to use it to build everything else. America has a long history of winning the first. The second is far less certain.

America has a lot to celebrate on the first count. More venture capital is flowing into the American AI economy than anywhere else in the world, orbital data centers included. We are building one of the most ambitious computing infrastructures ever attempted. Companies like OpenAI and Anthropic have created extraordinary value at extraordinary speed. But in the context of winning the second space race, we are watching the wrong scoreboard.

Winning is not decided by how much capital a company attracts or the early valuations of the companies developing a technology. America wins when technology drives real productivity gains across an economic sector. America may be winning the race to build AI. China may be winning the race to use it to build everything else. And while American investors fixate on the value being created by AI companies, China has been relentlessly focused on something different: turning AI into industrial productivity.

Nowhere is that more apparent than in automobile manufacturing. China transformed itself from an automotive also-ran into the world's largest producer and exporter in a single generation. Humanoid robots are already being prototyped to very soon assist humans on factory floors, performing quality inspections and moving materials. AI optimizes supply chains, improves quality and reduces the labor required to manufacture increasingly sophisticated products. The space industry faces the same trajectory.

The results are already showing up in global markets. The long-dominant German automobile now faces a fundamentally different kind of competitor. Chinese cars are appearing across European showrooms: high quality, beautifully designed, and remarkably inexpensive. The shift resembles what happened to the landline telephone in the 1990s: gradually at first, and then seemingly all at once.

China's approach has been relentlessly practical, taking advances from the tech world and applying them to the fundamentals: availability, quality, and cost. AI is becoming another tool for driving all three. But what does any of this have to do with the Space Force?

In the past decade, military and civil space have transitioned into an industrial enterprise much like the automotive industry did over the last 50 years. Satellites are increasingly software-defined and manufactured at scale rather than handcrafted. Ground networks and user equipment are following the same path. Missions generate and transmit enormous amounts of data, requiring decisions to be made on orbit and on the ground at machine speed. AI will transform all of it, from engineering design and supply chains to constellation operations.

But there is a single, critical difference between the automobile industry and the national security space industry, which is the customer. Carmakers compete for millions of them. Those customers reward companies that deliver better products or lower prices. Uncompetitive companies lose market share or disappear.

National security space is largely a monopsony, a market with one dominant buyer also known as the government. Through every acquisition decision, our government shapes what kind of industry will exist in the future. The Pentagon and Capitol Hill rarely frame it this way, but acquisition policy is industrial policy. Because of that, acquisition strategy must always emphasize competition to prevent the creation of monopolies. AI alone guarantees nothing. Only competitive pressure creates the incentive to apply it to win.

A company competing against four or five capable rivals has enormous incentives to deliver better products at lower cost. Awarding tens of billions in sole-source contracts to a single company will not produce the same incentives. And if that continues across larger portions of Space Force and intelligence community spending, the consequences will be worse than what taxpayers pay for satellites and launches. All too soon, a dominant supplier will begin dictating warfighters’ requirements and costs rather than the customer. It is dangerous for taxpayers, it is dangerous for the industrial base, but most importantly, it is dangerous for the warfighter.

The Space Force should not be asking how to incorporate more AI into existing systems. It should be asking how to restructure competition among providers to get the most capability to warfighters quickly, without being locked into any single one of them for 20 years.

In any future conflict, America will not fight solely with the satellites it has in orbit on the first day. Systems will fail. Others will be jammed, disabled, or destroyed. Capabilities will have to be replaced, upgraded, and reconstituted while the conflict is underway. The country that can do those things faster and cheaper possesses an enormous strategic advantage.

China's automobile industry should serve as a warning for America’s space industry. Once China begins to apply the same combination of AI, automation, scale, and relentless cost reduction to spacecraft, America's technological lead in AI will not automatically preserve its military advantage in space.

Hope is never a winning strategy, and hoping that a company chosen by one Pentagon bureaucrat today will win our space wars for the next 25 years is a losing strategy. The next space race will not be won by the country with today’s most impressive AI model or even the most capable individual satellite. It will be won by the country that learns to use AI to continuously produce, improve and operate space capability faster and cheaper than its adversary.