Something strange is happening in Silicon Valley. The most valuable companies in history are built on a technology that’s becoming free.

The big AI companies - OpenAI, Anthropic, Google - are valued at hundreds of billions because they have something nobody else has: massive language models that can write, think, and create. But here’s what’s keeping their investors awake at night: these models are becoming commodities at lightning speed.

The Pattern That Changes Everything

Just 12 months ago, OpenAI’s GPT-4 was miles ahead of anything else. Today? Free, open-source models are catching up. Not in years. In months. At a fraction of the cost.

This isn’t just another tech cycle. It’s a fundamental shift that’s played out before - with devastating consequences for the incumbents.

The Steam Engine Lesson

In 1769, James Watt patented his improved steam engine design. For the next 31 years, he and his partner Matthew Boulton had a monopoly. They charged enormous fees - equivalent to millions today - for anyone who wanted to use their superior technology.

Business boomed. They were the OpenAI of their era. Everyone needed their engines. They controlled the future.

Then 1800 arrived. The patents expired.

What happened next should terrify every AI investor.

Within months, competitors flooded the market. Richard Trevithick built high-pressure engines that were ten times more powerful than Watt’s designs. They were smaller, lighter, cheaper. The technology that Watt had kept locked away for decades suddenly became something anyone could build and improve.

The result? Steam engine adoption exploded. But Watt’s company? They had to pivot from being a technology company to just another manufacturer competing on price and service.

The Infrastructure Play

Here’s the part that should really worry proprietary AI companies: the winners weren’t other technology firms. They were infrastructure companies.

When steam engines became commoditized, the real money shifted to the companies that could build them cheaply and integrate them into existing systems. Railroad companies. Factory owners. Ship builders. They didn’t need to invest billions in R&D. They just downloaded the technology - literally took the blueprints - and built businesses around it.

We’re seeing the same pattern with AI. When Meta released LLaMA as open source, they weren’t being generous. They were executing the most brilliant business strategy of the decade: make the technology free, then sell the infrastructure to run it.

The Valuation Reality Check

Proprietary AI companies are valued as if they’ll maintain their technological moat forever. But history shows us what happens when technology becomes commoditized:

The valuation doesn’t just drop. It crashes.

Watt’s company went from having a monopoly on the most important technology of the Industrial Revolution to being just another steam engine maker. Their premium disappeared. Their moat evaporated.

Today’s AI leaders face the same fate, but faster. Technology moves at light speed compared to the 1800s. What took decades then happens in months now.

The New Winners

While investors pour money into proprietary AI companies, the real winners are emerging elsewhere:

  • Cloud providers who’ll run these models at scale
  • Companies that integrate AI into existing workflows
  • Businesses that use AI to cut costs rather than sell AI as a product
  • Infrastructure plays that benefit from AI becoming a utility

The pattern is clear: when revolutionary technology becomes free, value shifts to those who can deploy it efficiently, not those who invented it.

The Inevitable Math

Every month, open-source models get closer to proprietary ones. Every month, the cost to train and run them drops. Every month, more companies realize they can download powerful AI for free instead of paying premium prices.

This isn’t a speculation. It’s simple arithmetic.

The companies valued at hundreds of billions today are building their castles on sand that’s rapidly turning into water. The question isn’t whether the valuation crash will happen. It’s how fast.

The steam engine transformed the world. But it didn’t transform Watt’s bank account the way he expected. The same fate awaits today’s AI giants.

The future belongs to those who understand that the real value isn’t in owning the technology. It’s in knowing what to do with it when everyone has it.

History doesn’t repeat itself, but it rhymes. And right now, the AI industry is singing the exact same tune that played out two centuries ago - just at 1000x speed.

The great AI valuation crash is coming. The only question is: will you be surprised when it arrives?