> For the complete documentation index, see [llms.txt](https://docs.thetpa.ai/deai-tokenized-processing-architecture/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.thetpa.ai/deai-tokenized-processing-architecture/tpai-chain-full-architecture/economics-where-every-tpc-goes.md).

# Economics: where every TPC goes

AI usage pays the TPA owner, the GPUs and the validators, and nobody else. Stage takes no protocol fee; its GPUs earn the 20% GPU share only for jobs they run, like any datacenter.

<figure><img src="https://2937558112-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F7bJZTBY9aHXlYfgnTW9F%2Fuploads%2Fnido5Eu4ZeN580GshFeR%2Fpayment-split.png?alt=media" alt="On-chain payment split · founder decisions 14–15 Nov 2026"><figcaption><p>On-chain payment split · founder decisions 14–15 Nov 2026</p></figcaption></figure>

**Pricing makes the 20% cover the GPUs.** Each model has a US dollar price per compute unit, set so the 20% share equals the real GPU cost. Users pay in TPC, converted at payment time by a price feed, so TPC price swings never break the split.

$$
\text{Price} = \frac{\text{GPU cost}}{20%} = 5 \times \text{GPU cost}
$$

* **GPU cost** comes from a cost table per GPU type that includes expected idle time, updated on a fixed schedule.
* **The gateway meters real GPU time**, and the receipt confirms it.
* **Market pricing** is added later, once many datacenters compete.
* **External API TPAs** set their own price; the owner pays their own compute and keeps 80%.

**The TPA owner sets the price.** For GPU pool TPAs, 20% covers GPU cost and 20% goes to validators, so the lowest allowed price is compute cost ÷ 20%; above that floor, owners compete on price. If the price check shows the floor is not competitive, the split for GPU pool TPAs is revisited before launch. External API TPAs have no GPU cost on TPAI and pay only the 20% validator share.

The business team should compare final prices with the market before launch.
