The CapEx Knight
Build the GPU fortress — it's the grown-up move
The CapEx Knight looks across the industry, sees everyone building GPU fortresses, and concludes that the grown-up move is to build one in-house. This character often understands data centers better than models, workloads, or market timing. They are the person who says 'we can always do it cheaper ourselves' even when mature cloud resources would let the organization learn faster, change direction, and avoid betting the treasury on a hardware guess.
Symptom
Every cloud cost discussion becomes a build-vs-buy argument that the Knight always wins — by defining 'win' as 'we own the hardware.' The business case assumes 100% utilization that never arrives.
Why It Matters
The CapEx Knight turns a capacity question into a capital commitment. The GPU fortress is impressive. It is also inflexible, expensive to staff, and locked to a hardware bet that the model market may invalidate in eighteen months.
What the Chapter Gives You
Why owning GPUs feels strategic but often isn't, the hidden costs of the fortress approach, and how to evaluate build-vs-buy without ego in the room.
Want the full chapter? Grab the free cheat sheet, read an excerpt, or get the book.
Recognize this one in your codebase?
Free cheat sheet, excerpts, and interactive diagnostics.