Introducing Atlas: Infrastructure for Putting Entire Markets Onchain

Onchain finance can trade almost anything. But it still struggles to see the whole market.

One price feed may be enough for a simple swap. It is not enough for portfolio margin, tokenized funds, indexes, structured products, or liquidity tied to global markets. These applications need hundreds of prices at the same moment, not hundreds of disconnected updates they must assemble themselves.

Atlas is Steer’s multi-market oracle.

Its first market set brings together 620 equities, ETFs, and crypto assets as one synchronized onchain state. Prices are compressed without approximation, approved as one market view, and activated together.

That is the launch point, not the ceiling. Atlas is designed to scale far beyond it.

LitVM and Avalanche are the first chains where Atlas will be available, giving builders access to synchronized multi-market data from the start.

This is not another price feed. It is infrastructure for bringing entire markets onchain.

The next onchain market is bigger than one asset

The first generation of DeFi was built around crypto-native assets and relatively simple markets. Applications usually needed a small number of prices to lend, borrow, trade, or manage liquidity.

The next generation looks different.

Tokenized equities need to follow external markets. Funds and indexes need to understand baskets. Portfolio-margin systems need to evaluate many positions together. Structured products depend on multiple inputs. Market-aware liquidity needs to respond to benchmarks, net asset values, and prices beyond the local pool.

These products do not operate one asset at a time. They operate across a market.

Yet the data often still arrives feed by feed. Each value is delivered and verified separately, leaving the application to reconstruct a broader market view on its own.

That creates unnecessary cost and complexity. More importantly, it makes a basic question harder to answer:

Did every part of the application act on the same market?

Treat the market as the product

Atlas starts from a different assumption: if an application makes one decision across many assets, those prices should arrive as one market state.

Authorized operators approve the exact values and timing for an ordered set of markets. Atlas packages that state efficiently and delivers it onchain as one coordinated update.

The application can still read an individual price when it needs one. The difference is that every price belongs to the same known market context.

That means a risk engine does not have to wonder whether half of its collateral values came from one update and the rest came from another. An index does not have to assemble its basket from mismatched moments. A liquidity strategy does not have to infer whether the benchmark it follows is internally consistent.

The market arrives as a market.

620 prices, one synchronized view

Atlas launches with a 620-price market set.

The active set is designed to grow toward many thousands of prices that applications can read atomically as one market view. Beyond those hot prices, Atlas can support 20,000+ warm prices that can be brought into use when needed.

It includes 570 U.S. equities and ETFs alongside 50 crypto assets. During U.S. market hours, equity and ETF prices move with the regular market. After the New York close, those values hold at their regular-session close while crypto continues updating.

The result is a single view that respects the reality of each market. Closed assets do not pretend to keep trading. Continuous markets do not stop because New York has closed. Applications can see both through one consistent state.

Atlas is an early implementation, but it makes the larger idea tangible. Broad onchain markets do not have to be assembled from hundreds of isolated reports. They can be delivered, read, and updated as a unified system.

Compression without compromise

Delivering hundreds of prices together only works if the data can be packaged efficiently. But smaller payloads should not come at the expense of price precision.

Atlas never approximates an authorized price.

When market values move, the system can publish the exact changes from a known complete state. When that is no longer efficient, it publishes a new complete state instead. Compression changes how the data is represented, not what the data says.

Every decoded price remains identical to the authorized price. Market Codec was designed to introduce 0 bps of codec error.

That distinction matters. Models can be useful for forecasting markets, but an oracle delivery layer should not quietly estimate the values it was asked to carry. Atlas preserves them exactly.

One update means one update

A broad market view is only useful if it becomes active as a whole.

Atlas does not expose half of a new market alongside half of the old one. Every required part must be accepted before the new state becomes visible. If the update is incomplete, the previous market state remains active.

For applications, the outcome is clear: the complete new market is available, or it is not.

This creates a stronger foundation for decisions that span many assets. Portfolio health, index values, settlements, liquidations, and liquidity changes can all reference one active market state instead of a patchwork of updates.

What entire markets onchain can unlock

Atlas is designed for the point where onchain finance stops thinking only in pairs and starts operating across portfolios and markets.

That can include:

• cross-collateral lending and portfolio margin;
• tokenized funds, indexes, and baskets;
• equities and other externally priced assets;
• structured products and multi-asset settlement;
• treasury-wide monitoring and risk management; and
• liquidity that follows fair value, benchmarks, or net asset values.

It will not be the right delivery model for every use case. If an application occasionally needs one price, an individual feed may remain the simplest solution.

Atlas is built for something broader: applications that need to understand the state of a market before they can act.

From market data to market infrastructure

Steer has spent years building systems that allow onchain liquidity and execution to respond to external data. Atlas expands that work into a new layer.

It gives protocols, chains, issuers, and market operators a way to define the market their application needs and deliver that market onchain as one exact, synchronized state.

The goal is not simply to move more prices.

The goal is to make a broader class of financial products possible onchain.

The Atlas technical whitepaper will be published next, covering Market Codec, the authorization model, publication system, benchmarks, and security boundaries in detail.

For now, the idea is simple:

Onchain applications should not have to reconstruct the market before they can act on it.

Explore Atlas and see the whole market.

‍