#econoscope_markets
🔍 Did you know that the AI boom could trigger the next global recession?

⏩ Everyone’s talking about AI while companies are spending billions. But what if this spending is building a bubble that could pop?

The Bank for International Settlements (BIS) — the “central bank of central banks” — has issued a stark warning: the AI spending frenzy is the number one risk to global financial stability.

The numbers behind the gamble

💰 The five largest tech companies are about to commit over $1 trillion to AI-related data center capex in 2026, up from over $420 billion in 2025 — a 57% year-on-year jump.

But the returns? Not there yet.

1️⃣ Productivity gains are missing.
The St. Louis Fed analyzed 490,000 earnings calls since 2000. The result is that 95% of AI-related productivity talk is about what companies expect to happen in the future — not what's already happened. And that share has held steady since 2023.

2️⃣ Task-level wins, economy-wide nothing.
Goldman Sachs found that companies measuring AI at the task level reported roughly 30% median productivity gains. But for the whole economy? “Basically zero”.

3️⃣ The scary scenario.
Fitch Ratings modeled what happens if AI stocks drop 35% over six months: a US recession, GDP contracting up to 1.5%, private capital spending falling more than 6%, and global stagnation spreading within a year.

📎 This is Solow's Paradox 2.0.
Back in 1987, Robert Solow said: “You can see the computer age everywhere but in the productivity statistics.”
It took several decades of reorganizing factories, retraining workers, and changing workflows before electrification's productivity payoff showed up in the data. AI might be the same. So, the technology works. But companies need to reorganize before it pays off.

➡️ And that’s the risk. If the money keeps flowing but the expected returns don’t come, it won’t be just a tech crash. It will be a global economic story.

The bottom line is that AI isn’t failing. The diffusion is. Economy-wide gains require structural change — and that takes time the market might not have.

✍️ Written by Elmira Miftakhutdinova (3-IER)