The $37 Billion Lie

BLOG 1 OF 5

I'm going to share some numbers with you. They're all from the past six months and they're all from sources you trust — McKinsey, MIT, BCG, S&P Global. And they tell a story that nobody in enterprise technology wants to hear.

Enterprises poured $37 billion into generative AI in 2025 [1]. That's 3.2 times more than the year before. More than 6% of the entire global software market, captured in just three years. By any measure, this is the fastest enterprise technology adoption in history.

Now, the other side.

McKinsey surveyed 1,993 organisations across 105 countries [2]. They found that 88% now use AI in at least one business function. Sounds like a revolution, until you see the next number: only 6% qualify as "AI high performers" generating meaningful impact on their bottom line.

That's an 82-point gap between adoption and value. 82 points.

MIT's NANDA Initiative studied over 300 AI initiatives across 52 organisations [3]. Their finding was that 95% of enterprise AI pilots deliver zero measurable return on the P&L. Not disappointing returns, zero.

BCG's research across 1,250 companies [4] arrives at the same verdict from a different angle: 60% of enterprises generate no material value from their AI investments. The remaining 35% are "scaling" — a polite way of saying they're spending more money to find out if they'll ever get a return. Only 5% are actually winning.

And S&P Global [5] found that 42% of companies scrapped most of their AI initiatives in 2025. That's up from 17% the year before. Nearly half of enterprises actively abandoned their AI projects — in the same year that spending tripled.

Let that sink in. The market is simultaneously pouring in record money and pulling the plug on record numbers of projects. That's not investment. That's panic spending.

So what's actually happening?

Here's the number that connects all of this: Forrester estimates that generative AI currently orchestrates less than 1% of core business processes [10]. After three years. After $37 billion. Less than one percent of actual business operations run through AI.

Enterprises haven't spent $37 billion on AI. They've spent $37 billion on AI demos that never made it to production.

The adoption story is real. The technology works. The models are impressive. But somewhere between the demo and the daily workflow, something breaks. And it breaks at a rate that should alarm every board, every CTO, and every investor who's been told that AI transformation is well underway.

It isn't. Not for 94% of organisations.

The question isn't whether AI works. It's why enterprises can't make it work.

And the answer — supported by every serious study from the past 18 months — is not what most people think. It's not the models. It's not the data scientists. It's not even the data (though that's part of it).

It's something more fundamental. Something structural. Something that no amount of model improvement or prompt engineering will fix.

Next week, I'll break down the five specific failure modes that the research identifies — and why none of them are technology problems.

If any of these numbers surprised you, you're not alone. The gap between the AI narrative and the AI reality is the biggest untold story in enterprise technology. More to come.

Sources:

[1] Menlo Ventures, "2025: The State of Generative AI in the Enterprise," Dec 2025 → https://menlovc.com/perspective/2025-the-state-of-generative-ai-in-the-enterprise/

[2] McKinsey, "The State of AI: Global Survey 2025," Nov 2025 → https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai

[3] MIT NANDA, "The GenAI Divide," Jul 2025 → https://fortune.com/2025/08/18/mit-report-95-percent-generative-ai-pilots-at-companies-failing-cfo/

[4] BCG, "The Widening AI Value Gap," Sep 2025 → https://www.bcg.com/publications/2025/are-you-generating-value-from-ai-the-widening-gap

[5] S&P Global Market Intelligence, Enterprise AI Survey, 2025

[10] Forrester, "Predictions 2026: Enterprise Software," Nov 2025 → https://www.forrester.com/blogs/predictions-2026-ai-agents-changing-business-models-and-workplace-culture-impact-enterprise-software/