Part 11 of a sourced series. This installment is openly my opinion — the analysis, not the reporting. The facts it rests on are documented in Parts 1–10 and in the evidence explorer. I name no crime here, because that was never the point.
What I went looking for, and what I found
I started this series angry. I expected to find villains.
What I found instead was a machine. Ten parts in, the through-line isn't a person — it's a scoreboard.
- A depreciation assumption that lifts reported profit by billions (Part 1).
- "Cloud revenue" defined so the number flatters the story (Part 2).
- Pricing where it's free to enter and metered to leave (Part 3).
- Licensing that quietly tilts the field toward the landlord's own cloud (Parts 4–5).
- A licensing overhaul that 10x'd some customers' bills (Part 6, per the complainants).
- Infrastructure used as leverage against a single vendor (Part 7).
- And financing that loops capital between the same handful of players (Part 8).
Almost all of it legal. Almost all of it disclosed. None of it requiring a single bad person.
The uncomfortable part
Here's what's harder to sit with than a villain: decent people, inside these systems, would mostly do the same things.
If your scoreboard rewards this quarter's profit, you will reach for the assumption that raises this quarter's profit. If it rewards growth optics, you will define the metric that shows growth. If it rewards retention, you will price the exit so leaving hurts. You don't need malice. You need a number on a wall and a bonus attached to it.
I think about organizations the way I think about minds. A nervous system acts to keep reality matching its prediction — to minimize surprise. A public company does the same thing, except its "prediction" is the consensus estimate, and the surprise it fears is a missed quarter. From that lens, everything in this series becomes predictable. Not excusable. Predictable.
Why "bad people" is the wrong frame
Naming villains feels great. It also fixes nothing.
Fire a CEO and the scoreboard hires the same behavior in a new suit. Fine a company and it prices the fine into next year's plan. The incentive outlives the individual every time. That's why I keep saying it: no bad people, bad systems. Not because no one ever acts badly — but because if you want different outcomes, the person is the weakest place to push and the system is the strongest.
This is also why I refused, from the first sentence of this series, to publish accusations of crime I couldn't prove. Not just because it's reckless — because it's the wrong diagnosis. The crime frame says "remove this person." The systems frame says "rewire this incentive." Only one of those changes what happens next quarter.
Opinion — Michael. I'll go further: the villain story is itself a product of the same broken scoreboard. Outrage is engagement, engagement is reach, reach is the metric — so the loudest version of every story is the one with a bad guy in it. I didn't want to feed that machine while criticizing it. The honest story is quieter and more useful: the system is doing exactly what it's paid to do.
What a better scoreboard looks like
If the problem is the incentive, the fix is too. Imagine cloud economics where the rewarded behavior is the trustworthy one:
- Profit that's durable, not borrowed from a future write-down.
- Metrics customers can actually read.
- Portability priced like a right, not a penalty.
- Procurement and policy that make lock-in expensive for the vendor instead of the customer.
That's not naïve. The EU already did one piece of it — banning the exit toll outright by 2027 (Part 3). The point of regenerative design is simple: make the move that's good for the customer also the move that's good for the quarter. Align those two and most of this series stops happening on its own.
The invitation
So this isn't a takedown. It's a map of where the repair lives.
If you run a company: audit your own scoreboard before you judge anyone else's. If you buy software: read the depreciation note and the egress table — Part 12 shows you how. If you build systems of any kind: assume people will follow the incentive, and design the incentive you'd be proud to be followed.
There are no bad people in this story. There's a machine we all built, and can all rebuild.
Next — Part 12: "What To Do." Practical moves for buyers, builders, and citizens.
🔎 Check it yourself: explore every claim →
Top comments (0)