On September 14 the European Union switched on its first consolidated tape for shares and ETFs, a regulated reference feed for European equity trading operated by EuroCTP. Goldman Sachs made the case for it in July: less than a third of the region's trading volume appears on exchange order books, so investors reading only those books saw a fraction of real activity, and European markets were consistently undersized as a result.
Onchain markets are more fragmented than that. Trading is spread across hundreds of venues on dozens of chains, and a large share of the order flow arrives through aggregators, solver systems and RFQ market makers that never post to a public pool. There is no primary listing venue, no consolidated best bid and offer, and no equivalent of the US securities information processor.
A desk watching one or two venues is seeing a sample, and it prices from that sample. It quotes wider than it needs to and trades smaller than the market would allow.
Dune's Market Data Feeds turn onchain trading data into market data. The Spot Feed and the Perpetuals Feed publish each trade about a second after the block that contains it, carrying the instrument, the venue, the side the taker took, the executed price in quote per base, and a US dollar value marked against an oracle price pinned to that block. When no reliable mark is available, the feed reports the trade as unpriced and gives the reason.
Three applications account for most of the demand so far: price discovery, execution review and adding onchain markets to an existing data product.
1. Onchain price discovery
Market makers quoting on centralized venues treat onchain prints as an independent signal on where an asset is trading. The Spot Feed carries prints across 200+ venues, resolved and priced on arrival, so cross-venue spread and venue concentration can be measured as they move.
Onchain markets also never close. Tokenized assets trade on both spot and perpetual decentralized exchanges around the clock, so a desk holding a listed position can follow its gap risk through the weekend and into Monday's open.
2. Prove best execution and mark positions in real time
Proprietary trading firms, market makers and asset managers executing onchain carry the same obligation as any broker. They have to document that each fill got the most favorable terms reasonably available, and the review is transaction cost analysis: sampled trades measured against the market at the moment each one executed.
Getting that market picture has been the difficult part. A fill on one DEX has to be judged against every other venue trading the same pair in the same second, which until now has meant rebuilding the comparison venue by venue, in formats that do not match, long after the trade settled.
The Spot Feed supplies that comparison already assembled and priced, arriving in the firm's own systems as trades happen. The same record reprices open positions between scheduled marks. When a chain reorganizes, the corrections arrive as messages, so figures a team has already filed are not silently edited.
3. Add onchain markets to an existing data product
Data vendors and internal market data platforms already carry European equities, US futures and FX. Clients are asking for onchain next.
Building that means writing a decoder for every venue, following protocols that upgrade on their own schedule, issuing identifiers that stay stable as venues change, and passing corrections to every downstream client without breaking anything.
Market Data Feeds offer one normalized data model across venues and chains, with venue and instrument identifiers from Dune's registry. Europe took the same route in September, when the firms redistributing EuroCTP put one consolidated feed into the products their clients already used instead of asking them to connect to 130 venues.
The feeds cover a specific instrument catalog
The feeds do not carry every trade that settles onchain, and the gap is deliberate. A pool has to meet thresholds on liquidity, pricing integrity and venue registration before its trades enter the universe. A pair that no oracle can price, or a pool that exists for an afternoon, produces a price nobody should mark a book against, and including it would distort every aggregate built on top of the feed.
A quant screening new launches in their first hours will find this restrictive, and Dune's DEX datasets suit that work better. For a firm pricing positions it intends to hold, the narrower universe is what makes the prices usable. Dune publishes the methodology, and when a customer needs something that currently sits outside the catalog, the universe can be reviewed.
Historical data built on the same data model
Research teams usually backtest on one dataset and trade on another. When the two disagree there is no way to tell which is wrong, and a model that tested well can behave differently in production for reasons nobody can trace.
Historical Spot Trades comes out of the same code as the Spot Feed, with the same universe, pricing and identifiers. Six months of history is available in a standard trial package across the EVM chains and Solana, extended nightly, with BNB Chain to follow.
Getting access
Dune streams trades from both feeds via WebSocket, and reference data is delivered via REST API. Historical data can be accessed via API or inside Snowflake, BigQuery, Databricks or S3 via datashare.
