When you buy an "AI product," you're rarely buying it from the people who built the model. You're buying through a stack of layers — compute, foundation model, platform, application, reseller — and every one of them buys from the layer below and sells to the layer above at a markup. This tool maps those layers, and lets you follow a single dollar of compute as it gets marked up on its way to the customer.
From the customer at the top down to raw compute at the bottom. Each layer resells what's beneath it, wrapped in more convenience, value, and margin. Tap a layer to see who plays there, what they buy and sell, and why they can charge more than they pay.
Start with the underlying compute cost of doing one unit of work, then set the markup each layer adds as it resells upward. Watch how a small base cost becomes a much larger customer price — and see which layer actually captures the margin. The defaults are illustrative, not anyone's real numbers; change them to model your own scenario.
Each layer's slice is the margin it captures — the difference between what it pays the layer below and what it charges the layer above. The customer pays the full bar.
Every layer is a different business: different margins, different moats, different risks. Buying at the top means speed and someone else's accountability; moving down the stack means more control and more margin captured, at the cost of more to build and own. Deciding where your organisation should sit — as a buyer, a builder, or a reseller — is exactly the kind of question we work through in our strategy sessions.
Talk to us about your AI strategy