At Mavria, one of our heroes is General Leslie Groves, the father of the Manhattan Project. Most people know Robert Oppenheimer, the man Groves plucked from obscurity and transformed into an American Prometheus. Far fewer know Groves.
But those brilliant minds and the geniuses who worked with them could not have harnessed atomic energy in time on their own. Instead, the Manhattan Project required an enormous machine around them: we’re talking factories, materials, infrastructure, logistics, government authority, and more than 100,000 people, ordinary people and extraordinary people, working across thirty sites and dozens of moonshot subprojects at once. Oak Ridge had to figure out how to enrich uranium, Hanford built entirely new reactors and factories to make plutonium, and Los Alamos was turning brand new physics into a working weapon.
Around all of them were problems involving power, construction, materials, labor, logistics, and government. Unique bottlenecks threatened each project, and American soldiers paid in blood for each day the work was delayed. Someone had to clear every bottleneck.
That was the brilliance of General Leslie Groves. America already had the geniuses, we needed someone willing to solve every problem standing between them and the mission.
Mavria exists to build the machine around the people building intelligence. The power, supply chains, factories, and infrastructure that put it into the physical world on a wartime clock.
Today, America’s geniuses are pushing the frontier of intelligence.
We think artificial intelligence shares a rare quality with nuclear energy. In the right hands, both can create almost absurd abundance. The same physics that ended a world war lit up the grid. Cheap and plentiful energy transformed the modern world. Cheap and plentiful intelligence could do the same thing again, giving ordinary people capabilities once reserved for the most wealthy and powerful companies and institutions on earth.
AI is power. Superior intelligence means better weapons, better science, stronger cyber capabilities, more productive industry, and better decisions. The next great war will not play out like Red Dawn. An adversary does not need to invade America with tanks and troops to dominate us. Whoever leads in AI will have enormous power to shape the world and we want that future shaped by a free society.
Facing the Nazis, America did not need to be convinced that the mission mattered. When a bottleneck appeared, we solved it. But we are not living through a World War. The AI race has no beaches to storm and no daily death count. That makes the urgency easier to ignore, but we believe the stakes are just as high.
Our founding team learned how to do this together in government. We helped lead implementation of the President’s nuclear executive orders at DOE and the NRC, another field where America already had plenty of experts and far too many things standing in their way. We did not need to design reactors or become nuclear physicists. We needed to clear the path around the people who could. Within a year, four advanced reactors reached criticality, the first new reactor designs to do so in America in more than forty years. The companies and technical teams built the reactors. That was the point.
It was a smaller version of the same machine. Now it has to be built around AI.
Earlier this year, our team left government to focus on a new challenge, the AI race. America invented artificial intelligence and now we have to industrialize it. That means rapidly building the power, sites, supply chains, manufacturing capacity, financing, and partnerships required to deploy intelligence in the physical world. Like the physicists of the Manhattan Project, America’s brightest minds need the machine around them to keep up.
This time the bottlenecks look different. You may have the chips but not the power. Then if you solve for power, you still might wait years for transformers, switchgear, or turbines. America spent decades offshoring manufacturing and now we’re scrambling to rebuild. Atoms can be very stubborn. Factories are slow to show up. Supply chains can break. Construction is messy and never goes according to plan. And even if you solve all of that, financing, permitting, labor, or a community that does not trust you can still stop a project.
The machine only works if everything moves together.
At DOE, we had a front row seat to the collision of these worlds. Software moves in weeks, the physical world moves in years. They may as well live on different planets, except now they depend on each other. Unfortunately, you cannot prompt a substation into existence.
And just like before, the bottleneck keeps moving. Solve one and another moves onto the critical path.
And in a race, every bottleneck becomes a vulnerability. China doesn’t have to beat us at the hardest part. They may only need to stay close enough and then win on deployment. We can have the best frontier models and chips in the world and still lose if China builds the power, factories, and infrastructure around intelligence faster than we do. The race is moving from inventing the best intelligence to deploying it at scale.
The answer is not to become more like China. America has the better model but we must prove that a free society can still build really, really fast. America already has what we need. Incredible people, free markets, and the deepest pools of capital. Most importantly, we live in a democratic system capable of unleashing all three. But that system must move a hell of a lot faster.
We have done this before. The difference is, this time the mobilization will not be hidden away at secret sites. It will happen in factories, power plants, substations, construction sites, machine shops, and small towns. Winning the AI race is not just a job for a few thousand people in Silicon Valley. We need electricians, welders, machinists, lineworkers, engineers, manufacturers, utilities, entrepreneurs, and builders in every part of the country.
If intelligence remains scarce, it will concentrate power. If we industrialize it and make it abundant, AI has a chance to be the great equalizer.
The AI race may have started in Silicon Valley but it will be won all across America.
America already has the geniuses. Now the rest of us have work to do.
If you are attracted to hard problems and want your work to matter, come build with us.
We don’t get to inherit the future we want. We have to build it.