Pyth Network (PYTH)

What Pyth Network is

Pyth Network is a decentralized oracle that publishes market prices to blockchains, and PYTH is the token that governs and secures it. Pyth's documentation describes PYTH as "the native token powering governance and staking across the Pyth Network," and describes its main product, Pyth Core, as "decentralized price feeds for DeFi and TradFi builders with deterministic on-chain delivery" (source: Pyth documentation, read September 2026).

What distinguishes Pyth from other oracle designs is where the data comes from. Its documentation states that "publishers submit pricing information to Pyth's oracle program" and that "Pyth has multiple data publishers for every product." The publishers are firms that see the market directly, rather than intermediaries scraping public endpoints, and the aggregation happens on Pyth's own chain rather than in an off-chain agreement.

For where to buy Pyth Network from your country or state, see the availability tables on this site and the venues listed at Exchanges.

How it works

Prices are constructed on a dedicated appchain. Pyth's documentation states that "Pyth's oracle program runs on Pythnet" and that "the prices constructed on Pythnet are transferred cross-chain to reach consumers on these blockchains." Consumers on other chains read the result rather than receiving it, which is the pull model: a contract that needs a price fetches and posts the latest update as part of its own transaction, instead of the oracle paying to push updates to every chain continuously.

The aggregation algorithm is documented precisely, and it is the most interesting part of the system. Each publisher submits a price and a confidence interval. The oracle then gives "each publisher three votes — one vote at their price and one vote at each of their price +/- their confidence interval — then taking the median of all the votes." The aggregate confidence interval is computed by taking "the distance from the aggregate price to the 25th and 75th percentiles of the votes" and selecting the larger of the two.

Three votes per publisher, spread across their own stated uncertainty, produces a specific and useful behavior. Pyth's documentation describes it as acting "like a hybrid between a mean and a median, giving confident publishers more influence, while still capping the maximum influence of any single publisher." A publisher who submits a tight interval concentrates its three votes and moves the median more; a publisher who is unsure spreads them and moves it less. Because the result is a median, no single publisher can drag the aggregate to an arbitrary value, and the documentation notes that the aggregate price always sits between the 25th and 75th percentiles of the votes.

The confidence interval is not decoration. It travels with the price, so a consuming smart contract can decline to act when uncertainty is wide, which matters most for liquidation logic where a bad price causes irreversible harm.

Aggregation is tied to slots. The documentation states that "the first price update in a slot additionally triggers price aggregation, which combines the price data from the previous slot into a single aggregate price and confidence interval."

Supply and tokenomics

Pyth's public documentation covers the oracle mechanics in detail but does not publish a PYTH total supply, allocation or unlock schedule on the pages read for this entry, so no supply figure is stated here. What it does document is what the token is for: governance, and Oracle Integrity Staking, described as staking "PYTH to support data publishers and secure the integrity of Pyth price feeds."

Pyth publishes more than one product, and the distinction is a commercial one worth knowing. Pyth Core is the permissionless on-chain feed set. Pyth Pro, formerly called Lazer, is a subscription product aimed at institutions and supported on a narrower set of chains. Entropy is a separate service, a "secure, verifiable random number generator for EVM-based smart contracts."

History

Oracle designs divide on two questions: who supplies the data, and who pays to move it. Pyth answered both differently from the incumbent model. Rather than a network of node operators reading public APIs, it recruited firms that observe markets directly as first-party publishers. Rather than pushing updates on a heartbeat to every supported chain, it made consumers pull the update they need, which is what allows a 400-millisecond cadence across more than a hundred chains without paying for gas on all of them continuously.

Publishing a confidence interval alongside every price was the second distinctive choice, and it followed from the first. If publishers are market participants with genuinely different views of a fast-moving price, the honest output is a range, and the aggregation algorithm is built to preserve that range rather than collapse it.

The August 26, 2026 upgrade to Pyth Core added API key authentication to Hermes, the service that serves price updates to consumers, which is an operational change integrators need to account for rather than a change to the oracle's economics.

Risks and what to watch

Publisher concentration is the risk the aggregation algorithm mitigates rather than removes. The median caps any single publisher's influence, but a feed with few publishers, or with publishers drawing on the same underlying venue, is thinner than a feed with many independent ones.

The pull model shifts a responsibility onto integrators. Because a consumer posts the update itself, a stale price is possible if the consuming contract does not enforce a staleness threshold. That is an integration decision, and errors in it have caused real losses across the oracle sector.

Cross-chain delivery is a dependency. Prices are built on Pythnet and transferred to consuming chains, so the availability and correctness of that transfer path sits between the aggregation and the contract that reads it.

Oracle Integrity Staking involves staking behind publishers, which implies exposure to their performance. Anyone staking should read the current slashing and reward parameters directly rather than relying on a summary.

Governance sets the parameters that shape all of this, so PYTH voting concentration is worth watching.

Frequently asked questions

Where does Pyth get its prices?

From first-party publishers. Pyth's documentation states that publishers submit pricing information to its oracle program and that there are multiple publishers for every product, rather than node operators reading third-party APIs.

How does Pyth combine publisher prices?

Each publisher gets three votes: one at its price and one at each end of its confidence interval. The aggregate is the median of all votes, which caps any one publisher's influence while giving more weight to publishers who state tighter intervals.

What is a Pyth confidence interval?

The oracle's own statement of uncertainty. It is computed as the distance from the aggregate price to the 25th and 75th percentiles of the votes, taking the larger of the two. A wide interval means publishers disagree, and a consuming contract can use it to decline to act.

How often do Pyth feeds update?

Pyth Core feeds update at a 400-millisecond frequency. Prices are aggregated on Pythnet and transferred cross-chain, with consumers pulling and posting the update they need.

Where can you buy Pyth Network?

Availability depends on your country or US state. See where to buy Pyth Network for the exchanges serving your jurisdiction, and Exchanges to compare fees, kyc requirements and payment methods.