In the 1960s, the U.S. government asked a chilling question: How do we keep communication alive after a nuclear strike? The answer became the foundation of the internet as we know it—resilient, decentralized, unstoppable.

But somewhere along the way, that vision changed.

A Cold War Origin Built on Redundancy

No single point of failure: The DARPA mandate

The original internet wasn’t made for memes, streaming, or even email. It was made for war.

Back in the Cold War, the U.S. Department of Defense’s research arm, DARPA, had one terrifying design requirement: make the network survive a nuclear strike. If Washington D.C. went dark, data needed to reroute through surviving nodes—maybe New York, maybe California, maybe somewhere unexpected. Redundancy wasn’t a feature; it was the point.

This was a decentralized dream: packets flying across redundant paths, networks resilient by design, and no single point of failure.

The Rise of Hyperscale Cloud Providers

66% of cloud infrastructure runs on three companies

Fast forward to 2025, and most of the internet runs on just three providers:

• Amazon Web Services (AWS) – ~32% global cloud market share
• Microsoft Azure – ~23%
• Google Cloud – ~11%

(Source: Synergy Research, Q1 2025)

That’s over two-thirds of the internet’s cloud layer managed by a handful of corporations. Instead of routing around failure, we’ve built a world where when one hyperscaler sneezes, the internet gets the flu.

Real-world outages with real consequences

Recent history has proven the risk is more than theoretical:

• Dec 2021 (AWS): Netflix, Disney+, Venmo, and even Amazon’s own logistics systems went offline.
• Nov 2023 (Azure): 911 call centers and mobile banking apps in multiple U.S. states were crippled.
• Feb 2024 (Google Cloud): Airline booking systems went down across major carriers.

These weren’t nation-state attacks. They were just normal cloud hiccups—with massive downstream effects.

What If Crisis Hits Today?

Single-region failure = national disruption

Imagine what happens if a major region—like AWS U.S.-East—is taken offline, whether by cyberattack, infrastructure failure, or even a targeted military strike.

That region alone supports thousands of critical services across the U.S. economy.

Even defense contractors rely on hyperscalers

Over half of the top 100 U.S. defense contractors run at least part of their infrastructure on AWS or Azure. One outage could disrupt supply chains, operations, and even national defense workflows.

The decentralized internet DARPA imagined was meant to fight back when punched. Today’s internet is more like a glass cannon—powerful, but fragile.

Reclaiming Resilience

Owning your infrastructure is making a comeback

For years, it was fashionable to say “on-prem is dead.” But that’s changing.

Companies like 37signals (makers of Basecamp and HEY) have moved off the cloud, citing lower costs, more control, and greater resilience. A hybrid model—where critical systems run on owned hardware or diversified providers—is gaining traction.

Should the cloud be regulated like power?

If cloud infrastructure is now critical to national safety—just like energy or water—should it be regulated as such?

Some experts argue for government oversight, redundancy mandates, or geographic diversity requirements.

Cloud independence as a security principle

Business continuity strategies must evolve.

Cloud independence should be treated like encryption or multi-factor authentication: a foundational security control, not a luxury. Ask: Can your business survive your cloud provider going down?

Final Thought: Did We Trade Resilience for Convenience?

The internet was once designed to survive a nuclear war. Today, its lifeblood flows through a few hyperscale campuses in Virginia, Ohio, and Oregon—and we hope nothing happens.

Redundancy was a principle. Convenience replaced it. The question now is: will we decentralize again before we’re forced to?