AI, Albert Einstein, and Anthropic

In my recent post on Anthropic vs. the U.S. military, I addressed what appears, at first glance, to be a dispute over contract language and corporate boundaries. In reality, it is something larger. It is about whether the United States and other democratic nations will lead in the most consequential strategic technology since the atomic age.

History teaches a hard lesson: nations do not survive by being morally earnest but strategically weak. They endure by mastering the decisive technologies of their time.

The pattern is clear across centuries. Iron weapons reshaped empires. Gunpowder destroyed old military hierarchies. Naval power and oceangoing fleets determined which nations ruled trade and which were ruled by others. Industrial production, railroads, and steel changed warfare forever. Aviation, radar, codebreaking, and mechanized armor helped decide the Second World War. Nuclear weapons then created a world in which a handful of states could end civilization itself.

Every era has its defining instrument of power. In our time, that instrument is artificial intelligence.

The central lesson is not merely that new weapons matter. It is that the nation that best integrates a new technology into industry, intelligence, logistics, command, and military doctrine gains an advantage far beyond the battlefield. It gains economic leverage, diplomatic influence, industrial leadership, and the power to shape the rules of the world.

That is where we are now with AI.

Albert Einstein provides a useful historical analogy, with one important clarification. He did not design the atomic bomb, nor was he a weapons engineer in the conventional sense. His role was more strategic and political. In 1939, Einstein signed the famous letter to President Franklin Roosevelt warning that Nazi Germany might be pursuing atomic weapons. He understood that a free nation could not afford to ignore the possibility that a ruthless adversary might obtain such a capability first.

That insight remains highly relevant today. It is not enough to know that a technology is dangerous. One must also ask who will possess it, under what system of government, and to what end.

After Hiroshima and Nagasaki, Einstein’s thinking became more cautionary. Like many serious men of his generation, he came to fear not only the enemy’s possession of catastrophic technology, but also mankind’s own success in creating it. He increasingly emphasized restraint, responsibility, and the need for moral seriousness in the face of scientific power.

That is where the AI analogy becomes especially powerful.

Today’s technology leaders are not altogether unlike the physicists of the 1930s and 1940s. They are creating systems with extraordinary commercial value, but also profound military significance. AI is not merely a better search engine, a more efficient coding assistant, or a labor-saving device. It is becoming the nervous system of intelligence collection, cyber operations, battlefield awareness, logistics, autonomous systems, and military decision speed.

That is why the present debate around AI, military use, and corporate red lines matters so much. The stakes are not only commercial. They are civilizational.

Much of the public discussion still treats AI as primarily an economic revolution. It certainly is that. AI will reshape software, medicine, finance, law, research, education, manufacturing, and customer service. Nations that lead in AI will almost certainly enjoy enormous economic gains.

But the deeper source of power lies in military application.

Whoever can field superior AI in intelligence analysis, cyber defense, autonomous systems, satellite interpretation, language exploitation, and command support will gain a decisive advantage in future conflict. More important still, the nation that controls the infrastructure beneath AI, chips, networking, fabrication, power systems, and secure deployment, will control the strategic foundation of the next era.

For that reason, democratic nations cannot retreat.

I am deeply patriotic. I believe in the American constitutional system, the free market, capitalism, and the liberties that define our country. Those principles are not separate from this discussion. They are the very reason the United States must lead. If the most powerful AI systems in the world are ultimately controlled by authoritarian regimes, the future will not simply be more technologically advanced. It will be far less free.

China understands this. Russia understands it in its own way. India and other major powers understand that AI is not optional. It is foundational. The next balance of power will not be shaped only by aircraft carriers, missiles, and submarines, but also by inference engines, autonomous systems, secure cloud infrastructure, and the ability to deploy intelligence at scale.

That reality places American technology companies in a difficult position. They may prefer to view themselves as neutral builders of commercial products. History may not allow them that luxury.

The proper answer, however, is not lawless acceleration. America should not win the AI race by surrendering the very principles that make America worth defending. The correct path is strength under law. The United States should move aggressively in AI for intelligence, cyber defense, logistics, battlefield support, missile defense, and secure command systems, while maintaining clear constitutional and legal limits around domestic abuse, mass surveillance, and unaccountable lethal autonomy.

A free people must remain strong enough to survive and disciplined enough to remain free.

From an investment standpoint, this race will likely reward the companies that provide the infrastructure of AI more reliably than the companies building today’s most fashionable applications. Over the next five years, the key opportunity may lie not only in model builders, but in the firms that supply the compute, manufacturing, and systems architecture needed to deploy AI at scale.

My three preferred long-term beneficiaries are Nvidia, Broadcom, and Taiwan Semiconductor Manufacturing Company.

Nvidia remains the central company in the AI hardware ecosystem. Its leadership in training is well known, but it is also positioned to benefit from the next wave of inference and real-world deployment. When governments, cloud providers, and defense contractors need proven capability at scale, Nvidia remains the standard.

Broadcom offers a different but equally important angle. AI does not run on chips alone. It depends on data movement, networking, switching, and specialized system design. Broadcom sits close to that essential connective tissue and should benefit as AI deployment becomes broader, more customized, and more infrastructure-intensive.

TSMC is the indispensable manufacturer. No matter which model company wins, or which chip architecture gains favor, advanced semiconductor fabrication remains a strategic chokepoint. TSMC’s importance is not theoretical. It is foundational.

The lesson of Einstein’s era is that free nations cannot afford strategic innocence. The lesson of our own era is the same. AI is not simply a business cycle or a software theme. It is the next contest over power, deterrence, national survival, and the future character of the civilized world.

America should lead that contest with confidence, discipline, patriotism, and strength.

And investors should remember that the companies most likely to benefit may not be the loudest voices in the room, but the firms building the engines, rails, and steel of the age to come.