10 facts that make AI look like a bubble ↓

1️⃣ 95% of enterprise AI pilots delivered zero measurable profit impact.
$30-40B spent across 300+ initiatives with nothing to show for it. (MIT, 2025)
2️⃣ $650-700B in AI capex planned for 2026 by Alphabet, Amazon, Meta, and Microsoft alone.
Similar to the telecom overbuilds that preceded the dot-com crash.
3️⃣ 75% of S&P 500 returns, 80% of earnings growth, and 90% of capital spending driven by AI stocks since ChatGPT launched.
That level of concentration hasn't been seen since the late '90s.
4️⃣ $300B Oracle cloud commitment from OpenAI. $100B OpenAI investment from Nvidia.
Circular mega-deals inflating valuations without real cash flow behind them.
5️⃣ 25-35x revenue multiples for AI companies, with limited profitability.
The last time multiples were this disconnected from fundamentals, it didn't end well.
6️⃣ 90% of firms report no meaningful productivity gains from AI.
Hallucinations, constant oversight, and AI fatigue are slowing adoption. (NBER)
7️⃣ Rising debt to fund AI infrastructure, with no quick returns in sight.
Moody's is warning that capex is straining budgets faster than returns can materialize.
8️⃣ $2.52T in global AI spending forecasted for 2026, driven by a handful of mega-caps.
If adoption doesn't catch up to infrastructure, the economics fall apart. (Gartner)
9️⃣ Every major tech wave followed the same arc: hype, overinvestment, euphoria, correction.
Railways in the 1840s. Internet in the 1990s. AI is tracking the same pattern.
1️⃣ 0️⃣ $1T+ wiped from software stocks in days after AI sentiment shifted in February 2026.