Why Edge AI Will Survive the Cloud Bubble

US stock valuations are now higher than before the 1929 Wall Street crash. The AI sector, driven by data center investments and circular financing, shows all the classic signs of a bubble. But when it bursts, one segment will remain standing: Edge AI.

The Numbers Don't Lie

According to the Financial Times, the cyclically adjusted price-to-earnings ratio of US stocks now exceeds levels seen before the 1929 crash. In data going back to the 1840s, only the dotcom bubble of 1999 saw more stretched valuations.

The Productivity Reality Check

According to the Financial Times, citing OBR and McKinsey data, even the most optimistic AI productivity estimates barely exceed historical norms:

Source: Financial Times, OBR, McKinsey — December 2025

The takeaway? Even McKinsey's bullish 3.4% estimate represents only a +1.3% marginal gain over the 2002-2012 baseline—hardly justification for 1929-level valuations.

Three Bubbles Compared

The Financial Times research shows how the AI boom compares to previous technology manias in terms of market capitalization as share of GDP:

AI Drove 100% of GDP Growth

Here's the most alarming statistic: AI-related investment drove all of US GDP growth in H1 2025. Not manufacturing. Not services. Just AI.

The Magnificent Concentration

A handful of companies now represent a dangerous concentration of market value:

NVIDIA alone is worth more than the entire GDP of Germany. When a single chip company's valuation exceeds the economic output of Europe's largest economy, something is deeply disconnected from reality.

The Railway Parallel

Economic historian Gareth Campbell of Queen's University Belfast draws a fascinating parallel to the Railway Mania of 1845. At its peak, railway investment reached 6% of UK GDP. Hundreds of new lines were proposed.

"New railways connecting smaller towns would fail to ever find enough passengers."

By 1850, railway stocks had tumbled to one-third of their peak. Sound familiar? Replace "smaller towns" with "enterprise AI use cases" and you have today's data center overcapacity problem.

The Circular Financing Trap

Here's how the AI bubble inflates itself:

When analysts report that AI drove all GDP growth, they're counting the same money multiple times as it circulates between interconnected tech giants. This isn't real economic output—it's financial engineering.

Why Edge AI Is Different

Edge AI operates on fundamentally different economics:

The European Opportunity

Europe's regulatory environment—often criticized as innovation-hostile—becomes a strategic advantage when the bubble bursts:

• GDPR:Forces data locality, making edge AI the only compliant option for sensitive workloads
• NIS2:Critical infrastructure can't depend on foreign cloud providers
• AI Act:High-risk AI systems require transparency that cloud black-boxes can't provide

Organizations that build on edge AI now won't need to scramble when cloud costs explode or providers fail. They'll already be sovereign.

What History Teaches Us

The railways didn't disappear after 1850. The internet didn't vanish after 2001. But the companies that survived were those with:

• • Sustainable unit economics (not growth-at-all-costs)
• • Real customers paying real money (not circular financing)
• • Infrastructure that serves genuine demand (not speculation)

Edge AI checks all three boxes. When organizations need AI that works offline, processes sensitive data locally, and costs nothing to operate after initial deployment—they'll find edge solutions waiting.

Positioning for the Correction

We're not predicting when the bubble bursts—timing markets is a fool's errand. But we are saying: build on foundations that don't depend on the bubble continuing.

At FFT Cognitive, we build sovereign edge AI for organizations that can't afford to be wrong about which side of history they're on.