I thought Pyth is more interested in supply the data onchain. Let it be delivered, used, manipulated, etc…how ever the subscriber wants.
True it leaves a large gap to be filled. But with so many protocols out there, its better not to concentrate on the UI, its more inportant to concentrate on connecting the UI (3rd party) with the data.
Couldn’t say anything better than you are doing here.
Two questions about this point “Operate a subscription API infrastructure”
Technicaly is there a Global Market Data Model (Norme/Standard) ?
The API CAN be consumed by a AI ? Do we need a specific tier for this ?
To serve different types of institutional clients without scaling up large teams, Pyth can add an AI layer on top of its core data APIs. This layer would:
Standardized Data Access
Provide all raw Pyth data through a unified, permissioned API gateway.
Ensure outputs are in consistent formats (JSON/CSV/Parquet).
AI Customization Layer
Use AI agents to transform the same raw data into client-specific outputs:
Quant funds: backtest datasets, volatility surfaces, correlation matrices.
Banks/Compliance teams: audit-ready, traceable reports with on-chain proofs.
Asset managers: macro dashboards, cross-asset analysis.
Self-Service Interaction
Offer an AI copilot interface where clients can request insights in natural language:
“Show me BTC/ETH correlation over the last 6 months with Fed meeting dates marked.”
AI fetches, processes, and delivers visualizations or downloadable files instantly.
Business Model
Tiered subscription:
Basic – raw data API only.
Advanced – AI analytics & visualization.
Premium – AI + human advisory support.
Trust & Compliance
Every AI output includes a hash of the original data for auditability.
Reports can be exported in compliance-friendly formats (e.g. XBRL, ISO 20022).
Summary:
By inserting an AI customization layer between Pyth’s APIs and institutional clients, the DAO can deliver differentiated, semi-custom services at scale. This approach makes Pyth more attractive to diverse client types while keeping operational overhead lean.
thanks for mentioning ser, tradfi data subscription is high currently, it migh expand pyth prices feed into the tradfi pipeline
build something josh come on , i m wif you on every steps ahead
i know you can you just gotta focus on buidl
I believe that the main focus of profits should be rapidly expanding the data publishers price feeds.
When the data set is too compelling it’ll grow into the single source that all players will be buying. At that time, pricing power will allow for subscription prices of $1,000,000 or more per month still substantially lower than the maximum monthly amount some institutions pay now according to pyth pro pricing webpage.
That would put revenues in the billions and the pyth token price would reflect that revenue.
I think that PYTH is on the right track.
I believe the goal is to provide every price Bloomburg provides and then some. Hence “The price of Everything, Everywhere” but in 40 milliseconds.