I Did The Maths McKinsey Didn't, and Their "The state ofAI in 2025" Report Falls Apart

Spending More On AI Makes You Twice As Likely To Fail (McKinsey's Own Data). That's not my spin, but what happens when you actually do the maths McKinsey didn't do for you.

Read the full article: https://www.2ndorderthinkers.com/p/explain-the-2025-ai-report-from-mekinsey

Their 2025 State of AI report surveyed nearly 2,000 companies. Only 109—about 6%—qualified as "high performers." The rest? Strategy in place. Money spent. Boxes ticked. Nothing to show for it in profit.

McKinsey builds a compelling playbook from the winners. Ambitious vision. Workflow redesign. Senior leadership buy-in. Heavy investment.

But here's what they don't calculate: Of the companies that invested 20%+ of their digital budget on AI, only one-third became high performers. The majority failed.

In this video, I break down:

- Why the "88% AI adoption" headline is misleading

- How measurement bias determines where AI "creates value"

- Why the 10% agent adoption ceiling might not be early days

- The survivorship bias hiding in "best practices"

- The one calculation that flips McKinsey's narrative

0:00 – Hook: Only 6% are AI "high performers"
0:49 – Intro
0:58 – Why the "88% use AI" claim is misleading
3:20 – Where AI actually creates value (cost vs revenue)
10:09 – The calculation McKinsey buries: big spenders who failed
11:27 – Survivorship bias
13:09 – Key takeaways
15:26 – Final verdict

This is an analysis for executives who want to think critically—not just follow the hype.

Subscribe to Second Order Thinkers: https://www.2ndorderthinkers.com/