The Future Arrived Early. The Infrastructure Didn’t.
Every generation falls for a story.
Not because people are stupid. Quite the opposite.
The smartest people in the room are often the most vulnerable to a compelling narrative because they can build entire worlds around a possibility. They can see around corners. They can imagine what could be.
The problem is that sometimes they become so obsessed with the destination that they stop paying attention to the road.
And so here we are with AI.
The story being sold is breathtaking.
AI will transform every industry. Software will write itself. Doctors will become superhuman. Every worker will have an army of digital assistants. Productivity will explode. Entire business models will disappear overnight.
Maybe.
The interesting question isn’t whether any of that is possible.
The interesting question is whether the infrastructure exists to support the future everyone is already pricing into the present.
Because those are two very different things.
For all the talk about intelligence, most of the conversation conveniently skips over the plumbing.
The data centres.
The power generation.
The transmission lines.
The semiconductor supply chains.
The cooling systems.
The networking infrastructure.
The billions of dollars required before a single additional token gets generated.
It’s a bit like announcing a city for ten million people while you’re still arguing over who pays for the roads.
The future being marketed today assumes a level of infrastructure deployment that would make previous industrial buildouts look modest.
Yet everywhere you look, bottlenecks remain.
Utilities are warning about power constraints.
Data centre projects face years-long timelines.
Grid upgrades move at the speed of bureaucracy.
Chip demand continues to outpace supply in critical areas.
The demand story is running at venture capital speed.
The infrastructure story is moving at government speed.
Those are not the same speeds.
And markets tend to notice eventually.
What’s fascinating is how familiar all of this feels.
The internet bubble wasn’t wrong about the internet.
It was early.
Railroads weren’t a bad idea.
Too much capital simply arrived before the economics matured.
The automobile changed everything.
Hundreds of car companies still disappeared along the way.
History is littered with technologies that transformed civilisation while simultaneously destroying investor capital.
People often assume those outcomes can’t happen together.
They absolutely can.
A technology can be revolutionary and wildly overvalued at the same time.
That’s the part many investors struggle with.
Especially when the story gets emotional.
And AI has become deeply emotional.
Nobody wants to be the person who missed the next internet.
Nobody wants to be the executive who dismissed the biggest technological shift of a generation.
Nobody wants to look foolish.
Ironically, that fear often creates the very behaviour people are trying to avoid.
When everyone is terrified of missing out, independent thinking becomes socially expensive.
Suddenly questioning assumptions feels riskier than accepting them.
And so analysts build models.
Founders raise capital.
Banks publish reports.
Consultants produce forecasts.
Media outlets amplify every milestone.
Each participant acting rationally within their own incentives.
The result becomes a self-reinforcing machine.
Not necessarily a conspiracy.
Just human nature.
Wall Street isn’t immune to this.
Never has been.
There’s a popular belief that markets are giant truth-seeking machines filled with cold, logical decision-makers.
Reality is much messier.
Markets are collections of human beings.
Human beings have careers.
Careers have incentives.
And incentives shape behaviour.
If an investor loses money doing what everyone else did, that’s often survivable.
If they miss a historic winner because they refused to participate, that can be career-ending.
Those are very different calculations.
The safest place to be professionally is often standing in a crowd.
Even when the crowd is wrong.
Especially when the crowd is wrong.
Because being wrong together feels safer than being right alone.
And that’s where things get interesting.
The truth is usually sitting in plain sight.
OpenAI and Anthropic are extraordinary organisations pushing the frontier of what’s possible.
That can be true.
It can also be true that their long-term economics remain difficult for outsiders to fully understand.
Both statements can coexist.
A company can create enormous value for society while still facing enormous questions about profitability.
A technology can be transformative while its business model remains unresolved.
Innovation doesn’t automatically guarantee sustainable economics.
Sometimes the market acts as if those questions have already been answered.
They haven’t.
Not yet.
The uncomfortable reality is that uncertainty doesn’t fit neatly into headlines.
Certainty sells.
Nuance doesn’t.
“Everything changes tomorrow” is a much better story than “this may take longer and cost more than expected.”
But history suggests the second headline is often closer to reality.
At the end of the day, the smoke and mirrors aren’t really the problem.
The mirrors simply reflect what people want to see.
And people desperately want to believe in the future.
Always have.
Always will.
The lesson isn’t to become cynical.
It’s not to dismiss AI.
It’s not to assume the whole thing is a bubble.
The lesson is simpler.
Separate technological possibility from economic reality.
Separate future outcomes from present capabilities.
Separate the story from the infrastructure required to make the story true.
Because eventually reality catches up to every narrative.
Sometimes reality exceeds the hype.
Sometimes it falls short.
Most often, it arrives much later than anyone expected.
The future may absolutely belong to AI.
The question nobody seems eager to ask is whether we’ve mistaken a twenty-year buildout for a five-year one.
And that distinction tends to matter a lot when real money is involved.
SPINMANTV.ME 
Leave a comment