The Great AI Reality Check: Why the Smartest Money Is Leaving Early

The Great AI Reality Check: Why the Smartest Money Is Leaving Early

There is always a moment at a party when a few people slip out quietly.

They are not being dramatic. They simply know what happens when the lights come on and the music stops.

That is what the last week felt like.

Peter Thiel sold out of Nvidia.

SoftBank pulled back.

Michael Burry pointed at the fine print.

And the ASX dropped more than 60 billion in one session.

Everyone reacted as if the world had suddenly shifted off its axis.

The truth is far less dramatic.

This is what happens when the story becomes louder than the substance.

We have spent two years speaking about AI with the confidence of a converted believer.

We convinced ourselves that earnings would grow forever, that compute was an infinite resource, that geopolitics did not affect technology, and that one company could carry the entire future without stress.

Now the real world is calling the bluff.

The brightest investors are leaving because they can see the cracks forming under the surface. Not because they think the future is gone.

Not cracks in the technology but cracks in the way we have been thinking.

We built this AI era on emotion.

• Excitement.

• Momentum.

• Fear of missing out.

• Confidence bending into certainty.

• Certainty pretending to be understanding.

Markets played along. And then the ASX fell and reminded everyone that balance sheets still matter, supply chains still matter, and geopolitics still shapes everything. Even the hype.

Nvidia is not the villain here.

The problem is that we turned Nvidia into the entire story.

We treated GPUs like toys instead of the new oil.

We built investment narratives based on belief instead of evidence.

We assumed human behaviour would behave itself.

It never does.

Thiel moving into Apple and Microsoft is about someone reading the enviornment instead of the headlines. This is a real reminder that the real risk now sits around AI and not inside it.

Export controls.

Energy demands.

Political pressure.

Vendor dependency.

A global race for compute.

And a market that behaves like every day is a sprint.

AI is running on a timeline measured in decades.

Markets are running on a timeline measured in quarters.

Those two rhythms are clashing, and last week was the first sign of it.

So where does this leave us?

Exactly where we should be.

Looking at risk with clearer eyes.

Treating AI as a global system rather than a feel-good storyline.

Understanding that technology does not escape economics or human behaviour, no matter how powerful it becomes.

This is not the end of the AI movement.

It is the moment it begins to grow up.

Over the next few days I will break down each part of this shift in more detail.

• Geopolitics.

• Balance sheets.

• Concentration risk.

• Supply chain pressure.

• Human psychology.

• And the two timelines that are pulling against each other.

  
  

There is a lot to talk about and the commentary is coming.

For now I will leave you with one question.

Are we finally ready to see AI for what it really is, instead of what we hoped it would be?