How Late-Night Pizza Orders Foreshadowed a Covert Strike

By Josh Baylin
Published June 25, 2025 |  Updated June 24, 2025

Last week, an anonymous, automated account on the social platform X noticed something strange...

Pizza shops around the Pentagon and critical U.S. Air Force bases were seeing a surge in late-night orders.

At the same time, foot traffic near D.C.-area bars and restaurants – typically full of defense staffers after hours – plummeted.

In other words... folks weren't out drinking. They were still at work.

Within 24 hours, U.S. B-2 stealth bombers carried out one of the most consequential covert strikes in recent memory. They hit Iranian nuclear infrastructure with precision – and utmost secrecy.

But clearly, something saw it coming.

Not a human spy. Not a whistleblower.

The machines.

Even Top-Secret Missions Aren't Safe From AI

The U.S. military can still encrypt, redact, and control its official channels. But it can't control Google Maps heat signatures... or Uber Eats logs... or movement data gathered by our smartphones.

A generation ago, exposing a classified mission would have taken espionage and lots of money.

Now, it only takes the right AI model and a thousand digital breadcrumbs.

We live in a world where AI, sensors, and ambient data can predict a covert military operation before it happens. They can guess our intent from foot traffic and food orders, long before human analysts catch on.

Internal security just isn't enough when outside observation is enough to crack the code. The break isn't happening inside the system... It's happening around it.

Last week, for example, all it took were those two key signals:

  • A spike in late-night pizza orders in the Pentagon area
  • A drop in foot traffic around Washington, D.C. bars

Each signal was meaningless alone. But taken together, they suggested something big was happening.

And they were right.

The Companies Guarding Our Information Edge

To protect their information edge, militaries and corporations are turning to AI as well – not just to improve their own operations, but to stop external AI from figuring them out.

If the machines can see you coming, you'd better build machines that can see them seeing you.

This isn't just about cybersecurity anymore. It's about cognitive stealth.

The winner of the new arms race will be whoever can act without being noticed... in a world where everything is noticed.

Despite a more than $800 billion budget, U.S. military decisions still flow through e-mail chains, PowerPoints, and outdated dashboards. Meanwhile, AI can gain strategic insights from metadata in real time.

That's the fracture. But here's the opportunity... This same technology can create a huge advantage for players that know how to use it.

Three companies are already building the digital backbone of this new defense era:

  • Palantir Technologies (PLTR) has grown beyond data analytics into "decision synthesis." Its new Gotham AI doesn't just show commanders what's happening – it recommends what to do next, with statistics and risk assessments updated in real time.
  • Anduril Industries – a private company – is constructing what it calls "autonomous command networks." These AI systems coordinate drone swarms, missile defense, and battlefield logistics without human intervention. Anduril's Lattice platform already manages perimeter security for military bases using ambient sensors and predictive algorithms.
  • Shield AI – another private firm – focuses on the "cognitive edge." AI pilots can fly missions, adapt to threats, and coordinate with human teams faster than any trained operator. Its systems already outperform human pilots in simulated combat scenarios.

These aren't traditional defense contractors.

They're building AI-native infrastructure. That means AI is seamlessly built into their operations... It's not just an added feature.

Palantir's shares are up a massive 85% year to date. These private firms are also big leaders today – and they're worth keeping an eye on if they enter the public scene.

The Pentagon didn't confirm Saturday's strike... The pizza did. And the companies reading those signals? They're about to run the modern world.

As I always say... you're either early... or you're obsolete.

Good investing,

Josh Baylin

P.S. If you're interested in AI stocks, tune in this morning... because a special talk is happening over at our corporate affiliate, Chaikin Analytics. At 10 a.m. Eastern time, Marc Chaikin is revealing a secret to separating the best AI companies from the overhyped traps. Watch it today right here.

Further Reading

"The smartphone itself must die," Josh writes. The same three-step pattern that predicted the rise of the smartphone now foreshadows its demise. And companies already investing in this revolution will shape the next generation of wealth.

As AI gets more popular, investors must watch out for companies that can't deliver. Some use futuristic promises to hide stale ideas – and even fraud. That's why the real players set to win the AI race are the hyperscalers... with the resources to build their own AI models.

Back to Top