When I was a child in the 1950s, in the heart of the manufacturing era, a President could create a business policy around a few key men. Engine Charlie ran GM and, later, the Defense Department. In that post he succeeded Electric Charlie, who had run GE.
In the oil era, the Bush family was plugged into power through James A. Baker, whose Baker & Botts law firm represented the biggest players.
In 2024, leading AI companies elected Trump. In 2025 their plan to grow the industry, through that connection, fell apart instantly.
The reason is obvious to anyone who has covered tech for a while. Tech leadership can change quickly, and there are enormous advantages to being small. Small companies can find clues while big ones remain clueless.
Elon Musk isn’t guaranteed AI industry leadership because he made half a trillion dollars in cars and rocket ships. Legendary Intel CEO Andy Grove called his autobiography “Only the Paranoid Survive” and history has borne him out. Intel stopped being paranoid when he left and is now waiting to be broken up for parts.
Meta is now the world’s phone company. But it’s barely 20 years old. Amazon and Google didn’t exist when the Web was spun thirty years ago. Nvidia was founded in 1993. The industry leaders of the 1990s were AT&T, IBM, and GE.
Basing public policy on what tech leaders want is foolishness. You’re trying to entrench “leaders” whose own leadership is always suspect. They’re not where the industry’s dynamism comes from.
The DataSeek Lesson
This is the lesson of DeepSeek, and the great irony is it took a Chinese start-up to teach it.
China wrote a new tech policy in 2020, one that disadvantaged its biggest tech firms, but tech didn’t die as a result. China cut down its big trees and let 1,000 startups bloom.
An industrial policy for tech, if such existed, would favor startups, and I mean startups. The over-funded WeWork is not the model. It’s the little guys with barely enough funding to get by that will rule the future. They lack the bureaucracy that can inhibit change.
Defining tech’s future based on what a few key leaders want is like trying to hold water in your hands. Google and Microsoft may not exist in their present forms in the year 2050. They have enormous amounts of sunk capital, in data centers and fiber lines stretching around the world. But it’s what runs inside those data centers and over those fiber lines that will determine what the industry looks like even five years from now.
What Musk did, by taking over Washington, is put an unwelcome spotlight on himself. It guarantees the feet of 1,000 rivals will now come out of that light, looking to squash him like a cockroach. That’s the way the tech business works. Leadership is earned, and it must be re-earned every day, or it can disappear.
Given tech’s continued, and even growing, dominance over the global economy, this is a hard lesson to implement. But it still must be implemented. Thank DeepSeek for teaching it.
