Pyth Network (PYTH) : A complete overview of the decentralized oracle

Key Information

• Pyth Network is available on 78 blockchains, including Solana, Ethereum, Avalanche, BNB Chain, Polygon, and Aptos, using the Wormhole protocol to transmit data securely.
• The network is powered by 114 qualified data providers, such as Binance, Cboe, and Jump Trading, ensuring reliable price feeds for over 500 financial assets, including stocks, currencies, metals, and cryptocurrencies.
• Pyth Network has secured over $842 billion in transactions since its launch, holding a dominant position on networks such as Solana and Sui.
• The protocol offers a wide range of products, including Price Feeds, historical price archives (Benchmarks), Pyth Entropy for secure random number generation, and Express Relay to address the Maximal Extractable Value (MEV) issue in decentralized finance protocols.

---

What is Pyth Network?

Pyth Network is a decentralized oracle protocol designed to meet the growing demand for financial data among decentralized applications (dApps) in the blockchain industry. Developed on its own network, called Pythnet, the protocol provides on-demand access to reliable and secure financial data streams.

Launched in 2021, Pyth Network now provides over 500 financial feeds on various assets, ranging from stocks, currencies, and metals to cryptocurrencies. Its services are available on 76 blockchains, including Solana, Avalanche, Polygon, Arbitrum, Base, and Sui. Moreover, the data is sourced from 114 qualified entities, called publishers, including Cboe, Binance, and Jump Trading.

Unlike traditional oracles that rely on secondary data aggregators, Pyth uses a first-party data model, ensuring greater accuracy. With this model, prices are updated every 400 milliseconds and can be transmitted interoperably across blockchains via Pythnet, a dedicated appchain, and the Wormhole interoperability protocol.

---

Pyth Network’s Thesis

The oracle market is currently dominated by players who do not offer optimal accuracy due to their reliance on secondary sources. Data is retrieved and analyzed by nodes, which then reach a consensus before publishing it on the network and sending it to smart contracts. This push-based model causes relatively significant delays, which can lead to discrepancies between the real value of a data point and what is displayed to the user.

To address this issue, Pyth Network is innovating with a network of primary data providers and a pull-based model, allowing users to request price updates only when necessary. This system not only reduces latency but also lowers the costs associated with regular updates.

Furthermore, Pyth does not stop at providing financial data feeds but also offers services such as Benchmarks, historical price archives, Pyth Entropy, a secure random number generator, and Express Relay, a tool tackling the Maximal Extractable Value (MEV) issue for decentralized finance protocols.

---

Pyth Network Products

Price Feeds

Price Feeds are the core of Pyth Network. They are continuous streams of information that provide decentralized applications (dApps) with continuous access to the prices of a wide variety of assets, including cryptocurrencies, stocks, commodities, and currencies. Pyth Network’s Price Feeds are essential for automatically and accurately executing smart contracts in financial applications. For instance, a lending protocol depends on these feeds to calculate user collateralization levels in real time and trigger liquidation if a risk threshold is exceeded.

Each Price Feed is powered by a multitude of sources called data providers (over 110 at the time of writing). Among them are globally renowned firms such as Amber Group, BATS Global Markets, Binance, Bybit, Cboe, Coinshares, Flowdesk, Huobi Global, Kaiko, Jump Trading, OKX, and Wintermute.

The information is collected in real time from these different sources and then consolidated by Pyth’s aggregation algorithm to produce an accurate price estimate, accompanied by a confidence interval. This process ensures that each communicated price is representative of the market reality while minimizing the risks of manipulation or error.

Pyth Network’s Price Feeds stand out for their very high update frequency. Indeed, refresh rates can go up to once every 400 milliseconds on certain blockchains like Solana. Thanks to this speed, DeFi applications can access near-instantaneous market data, which is crucial for use cases where volatility can pose significant risks.

Benchmarks

The Benchmarks product offered by Pyth Network provides access to historical datasets derived from Price Feeds. This allows developers, researchers, and users to access past data for various use cases, such as backtesting investment strategies or evaluating asset performance. Unlike real-time feeds, Benchmarks provide price data over a broader period.

It is important to note that Benchmarks are built from information originating from Price Feeds and therefore offer the same quality and precision standards provided by Pyth Network’s data providers. Thus, each price displayed by the product at any given date corresponds to the price published by Pyth at that exact date in the past, as if a user had observed the feeds at that moment.

Pyth Entropy

Pyth Entropy is a secure random number generator designed to meet the needs of decentralized applications. Indeed, this tool is particularly useful for powering smart contracts that require unpredictable results, such as online games (GameFi), lotteries, prediction markets, or even non-fungible token (NFT) minting services.

As surprising as it may seem, generating random numbers securely on blockchains is particularly difficult due to their deterministic nature. At the same time, decentralized applications require a high level of unpredictability to ensure fairness and transparency in the services offered.

To address this problem, and unlike traditional random number generators, Pyth’s Entropy product creates an entropy source that is theoretically impossible to predict using a cryptographic protocol based on a commit-reveal mechanism, which operates in two phases. In the first, called the commitment phase, each party generates a secret random number that is hashed and recorded on the blockchain without revealing the initial number. Then, during the reveal phase, the parties reveal their respective numbers, which are combined to create the final random number. This process ensures that each participant cannot anticipate or influence the final result, making the process completely independent and secure.

Express Relay

Launched in July 2024, Express Relay is the latest product developed by Pyth Network. This solution is designed for developers and addresses the Maximal Extractable Value (MEV) issue. In a few words, MEV refers to the profits miners or validators gain from their role in block confirmation. Indeed, they can rearrange, include, or exclude transactions within a block, allowing them to generate profits at the expense of users and decentralized finance protocols.

With Express Relay, this dynamic is changed as it is now up to DeFi protocols to prioritize transactions via an auction system. To understand this, let’s take the example of a lending protocol that requires the intervention of external actors (searchers) to execute user liquidations. It can create auctions to obtain the priority to execute these transactions, leading to competition among searchers and allowing all actors to benefit from better profitability.

---

Pyth Network’s Architecture and Technology

General Operation

As a preamble, Pyth Network is divided into two relatively close segments, which are nonetheless important to distinguish. The first operates on the Solana blockchain and only serves smart contracts on it, while the second operates on Pythnet and provides data to all other blockchains. We will return to this later in the next sections of this report.

In any case, this does not fundamentally change the general operation of Pyth Network. Simply put, the network of publishers provides high-quality financial data, which is then retrieved by Pyth Network’s oracle algorithm. This data is then provided to decentralized applications’ smart contracts in two ways: either directly on the Solana network or on Pythnet and then to other blockchains through the Wormhole communication protocol.

Oracle Program

The Oracle Program of Pyth Network is the algorithm responsible for collecting data from publishers, aggregating it, and offering a final consolidated price, along with a confidence interval, to decentralized applications and developers. This program also maintains price feeds, stores a certain amount of data, respects strict quality criteria, and makes data available within appropriate timeframes. To ensure transparency of the process, the Oracle Program also records certain information associated with the data, including provider details.

Pyth on Solana

As explained earlier in this report, Pyth runs directly on the Solana blockchain to provide financial data feeds to decentralized applications. These Price Feeds store all necessary metadata, including price, asset ticker, a moving average over different time scales, and a confidence interval.

In the initial version of the oracle, developers could connect feeds to their smart contracts using a push-based model and receive an update every 400 ms (the duration of a slot on Solana). While this allowed Pyth Network to become the leader on Solana (95% of Total Value Secured and 100% of Total Volume Secured), it presents significant issues, as discussed in the “Thesis” section of this report.

For these reasons, on June 17, 2024, Pyth Network announced a major change with the introduction of a pull-based oracle. Thus, since June 30, 2024, decentralized applications on Solana access the oracle services by querying Price Feeds only when needed through the system established on Pythnet. In other words, protocols on Solana now use Pyth Network’s oracle in the same way as any protocol on another blockchain.

Pyth on Pythnet

Pythnet is an appchain, designed as a custom solution to serve Pyth Network’s needs. Simply put, the network is operated by data providers (who have a role similar to validators) and acts as a foundational computational layer that securely and reliably processes and aggregates the data they provide to form Pyth’s Price Feeds.

The technological infrastructure is a fork of Solana’s mainnet, whose consensus mechanism has been modified to a Proof Of Authority, a variant of Proof Of Stake where, instead of tokens, network participants stake their identity and reputation. Pythnet’s validators are the data providers themselves, and the Pyth Data Association DAO allows them to operate their node by delegating the necessary tokens to them.

It should be noted that Pythnet naturally benefits from Solana’s speed and efficiency but is not directly affected by bandwidth fluctuations or cost increases that may impact Solana during periods of high activity.

With Pythnet, any user can consult the blockchain’s transaction log to ensure that price calculations were performed correctly and that the price information provided by Pyth Network’s oracle is accurate and up-to-date. Moreover, Wormhole Guardians — who relay data to blockchains using Pyth’s services — also validate these calculations when forming their messages. As long as two-thirds of them are not malicious, users can trust the accuracy of the calculated prices.

Finally, Pythnet also ensures the security of the price aggregation and combination operation. Pythnet is designed to remain functional as long as two-thirds of the validators (data providers) are connected and operational. It should be noted that Pyth Network’s services are not dependent on the congestion of the blockchains receiving the price feeds.

Cross-Chain Messaging

Cross-Chain Messaging is an essential component of Pyth Network, as it enables the secure transmission of aggregated data from Pythnet to various blockchains. Currently, 76 blockchains are connected, including Ethereum, Polygon, Optimism, BNB Chain, Aptos, and, since June 30, Solana. To achieve this, the Wormhole inter-blockchain communication protocol is used.

Once Pythnet’s Oracle Algorithm has validated and aggregated the data, it produces a consolidated price along with a confidence interval. This is not sent directly to Wormhole but is transmitted in a “buffer” message, itself hashed in the form of a Merkle tree. This process is repeated every slot.

Subsequently, Pythnet validators communicate with the Wormhole smart contract deployed on Pythnet to send a message containing the root of this Merkle tree. Wormhole Guardians agree on the validity of the message and communicate it to the main network through a VAA, which will then disseminate the price information to blockchains.

“VAAs are Wormhole’s basic messaging primitive. They are inter-chain data packets issued whenever an inter-chain application contract interacts with the core contract.

Guardians must validate the messages issued by contracts before sending them to the target chain. Once a majority of guardians have observed the message and determined its finality, the guardians sign a keccak256 hash of the message body.

The message is wrapped in a structure called a VAA, which combines the message and guardian signatures to form proof.” Source: Wormhole Documentation

Hermes

Hermes is the final step in this long journey that data takes from its issuance by the provider to its reading by the decentralized application. Indeed, Hermes is the API developed by Pyth Network to simplify and automate the process of retrieving aggregated Price Feeds from Pythnet.

Hermes aims to prevent decentralized application developers from interacting with the Merkle proofs generated by Wormhole. The protocol monitors in real-time the messages communicated by Pythnet and the Merkle roots verified by the Wormhole protocol, manages translation, and extracts price data.

Then, Hermes allows developers to retrieve this information by the method of their choice: either via a simple REST request or through a WebSocket connection for a continuous data stream. Hermes’s main advantage is that it abstracts complex technical details and provides an easily usable service through a traditional API.

---

PYTH, Pyth Network’s Token

PYTH is the native cryptocurrency of the Pyth Network ecosystem, designed to be the key to decentralization and governance of the oracle. It allows users to contribute to the life of the Pyth DAO, whose rules are written on-chain in the Pyth DAO Constitution, and to delegate their tokens to the network of validators, i.e., the data providers.

PYTH Tokenomics

The initial launch of the PYTH token took place in November 2023, with a circulating supply of 1.5 billion tokens, or 15% of the total supply, set at 10 billion tokens. This limit is expected to be reached by around 2027, through a phased release schedule at 6, 18, 30, and 42 months after the initial launch. The distribution of PYTH supply is structured to favor the ecosystem’s expansion and the protocol’s continued growth. The token distribution table is organized as follows:

• Ecosystem Growth: 52%

5.2 billion PYTH are reserved for initiatives related to ecosystem development. This allocation is intended to support contributors, whether developers, researchers, media, or strategic players. Approximately 35% of this allocation (1.825 billion PYTH) has already been unlocked.

• Publisher Rewards: 22%

2.2 billion PYTH are allocated to data providers, also known as publishers. This allocation is designed to reward publishers for their crucial role in regularly publishing accurate and reliable data. Currently, about 26.7% of these tokens (587.5 million PYTH) have been unlocked.

• Protocol Development: 10%

1 billion PYTH are reserved for the main contributors of the protocol to create and develop the network’s infrastructure. This allocation aims to fund the design of new tools and products related to Pyth’s services. To date, 36.25% (362.5 million PYTH) of this allocation has been unlocked.

• Community & Launch: 6%

600 million PYTH are allocated to the community and launch initiatives. This allocation was distributed to reward active community users, members who contributed to the oracle’s development, and certain users of DeFi applications. All of these tokens were unlocked at the mainnet launch.

• Private Sales: 10%

1 billion PYTH were reserved for participants in a strategic funding round from December 2023, the details of which have not been disclosed. Among them are prominent investors such as Delphi Ventures, Wintermute Ventures, and Multicoin Capital. These tokens were partially unlocked for the first time in May 2024, at 25% (250 million PYTH).

PYTH Staking for Governance

The PYTH token is at the heart of the decentralized governance system of the Pyth DAO, allowing holders to actively participate in decision-making regarding Pyth Network’s evolution. By staking their PYTH tokens, users can submit proposals and vote on Pyth Improvement Proposals (PIPs). Each PYTH token represents one vote (at a 1:1 ratio).

Community proposals can be submitted provided that the proposer holds at least 0.25% of the total staked PYTH supply. To be approved, a proposal must receive more favorable votes than unfavorable ones and reach a minimum approval quorum at the end of the 7-day voting period.

Pyth’s governance system is organized around two main councils: the Pythian Council and the Price Feed Council. The Pythian Council, composed of 8 members, is responsible for protocol operations-related decisions. The Price Feed Council, composed of 7 members, oversees Price Feeds and data providers. In the governance system outlined in the Pyth DAO Constitution, these members are re-elected every six months (4 for the Pythian Council and 3 for the Price Feed Council).

PYTH Staking for the Oracle

The PYTH token can also be used to contribute to Oracle Integrity Staking (OIS), a mechanism aimed at ensuring the reliability and quality of the data provided by Pyth Network. Data providers, i.e., Pythnet’s validators, can stake PYTH tokens as collateral to demonstrate their commitment. In parallel, users can delegate their tokens to data providers, sharing the risks and rewards associated with this activity.

If a data provider submits erroneous or inconsistent information, part of their staked PYTH — as well as those of users — can be slashed, reinforcing the incentive to provide high-quality data. Logically, the reputation of validators is directly correlated to the quality of the data provided.

---

Pyth Network’s Key Metrics

Since its launch, Pyth Network has secured over $842 billion in transactions. In September 2024 alone, it secured $48.3 billion in total volume, making Pyth the leading oracle based on this metric. Additionally, the total value placed on decentralized applications using Pyth’s services (Total Value Secured) reached $7.5 billion at the time of this article’s writing.

In terms of adoption across various blockchains, Pyth Network serves 44% of the market on The Open Network (TON), 74% on Solana, and 91% on Sui. Two new blockchains subscribed to Pyth’s services in September 2024, ApeChain and Flow EVM, joining a total of 78 different networks, including Avalanche, BNB Chain, and many Ethereum layer 2 solutions.

Finally, Pyth Network is supported by 117 different data providers, allowing it to supply blockchains with price feeds for 512 different assets. As for partners connected to its services, there are over 420 decentralized applications and protocols. It is noteworthy that the majority of them are derivatives protocols, with $46.2 billion in volume in September versus only $2.1 billion for decentralized exchanges.

Update 11/05/2024

Pyth Network has its own ETN quoted on Euronext and emitted by VanEck.

---

Conclusion

Pyth Network has redefined the landscape of decentralized oracle protocols, with a pricing model based on expert data providers and a pull-based technology that enables clients to optimize their consumption and service costs. Furthermore, with the combination of Pythnet, Wormhole, and a suite of tools (Oracle Program, Hermes, etc.), Pyth provides precise and reliable price feeds to a wide range of blockchains.

Since its launch in 2021, the protocol has rapidly expanded to cover more than 500 financial asset feeds across 78 blockchains, with over a hundred renowned data providers such as Binance, Cboe, and Jump Trading. By continuing to expand its network of publishers and maintaining a strong commitment to data quality, Pyth Network is well-positioned to compete with the current leader in the field, Chainlink.