History Repeating Itself with AI

Posted on by

History Repeating Itself with AI

In the 1800s, the U.S. railroads transformed everything. They didn’t just move goods—they moved economies. Fortunes were made and lost overnight. There was real innovation, but also wild speculation. Thousands of companies laid track; only a handful survived. And over time, the real winners weren’t always the rail lines—but those who sold the steel, controlled the infrastructure, or secured government contracts.

Fast forward to today. The AI revolution is playing out on the same track—just exponentially faster.

We’re seeing:

  • Big platforms building the new digital rails.

  • Startups pitching flashy demos with little revenue.

  • Governments stepping in as anchor customers.

  • And capital flooding in, chasing the next “OpenAI moment.”

But just like back then, not every train makes it to the station.

That’s why we need a new kind of investment lens—one that cuts through the hype and asks:
Is this a durable platform or just a passing feature?  Is value being created—or just funded?

Apply this lens to today’s AI leaders:

  • Palantir looks like 19th-century infrastructure: embedded, defensible, running on long-term government contracts. Low flash, high resilience.

  • OpenAI is the transcontinental line—ambitious, dominant, and costly. It’s powerful, but still racing ahead of its margin structure.

  • Anthropic? A promising route through tough terrain. Big contracts are coming, but the burn rate could derail the journey.

If AI is the new railroad, the same lessons apply:  Don’t just chase the trains. Own the tracks.

Robert Muir
Former Australian Investment Commissioner, North America
Former Director, Australian Nuclear Science and Technology Organisation (ANSTO)