Advertisement

Crypto Isn’t a Great Data Refinery, But It Could Be

Presented by Onchain Capital
Crypto has become exceptionally good at gathering information. Its next challenge lies in refining that information to a standard suitable for serious financial decisions.
By: Kha Nguyen •
Crypto Isn’t a Great Data Refinery, But It Could Be

“Data is the new oil” has circulated through technology conferences for close to two decades. Crypto adopted the phrase eagerly, usually as shorthand for the extraordinary quantity of information that blockchains produce.

Crude has limited value at the point of extraction. It contains impurities, varies considerably in quality and passes through an elaborate industrial process before becoming fuel capable of powering a car, turbine or aircraft.

Data follows a similar path. Extraction produces the raw material, while refining turns it into something people can rely on.

Crypto has become exceptionally good at gathering information. Its next challenge, however, lies in refining that information to a standard suitable for serious financial decisions.

Collecting blockchain data is the beginning

Public blockchains generate an immense volume of information. Every swap, transfer, liquidity change and wallet interaction leaves behind a record.

At sufficient scale, those records accumulate into hundreds of terabytes and eventually petabytes. A relatively small portion becomes information that a trader, fund manager or financial institution can readily use.

Creating that usable output demands a lot. A provider needs extensive coverage across decentralised exchanges and other markets. Conflicting records require reconciliation. Assets, markets and transactions require consistent definitions. Historical figures need continuity so that a metric calculated today retains the same meaning when somebody examines the series months later.

Then there is time. For professional users, refined data needs to reach their systems within seconds of a trade taking place on DEXes around the world. Freshness forms part of quality. A figure can be meticulously calculated and still lose much of its commercial value when it arrives after the decision it was meant to inform.

Combining wide coverage with very low latency makes refinement much harder. Cleaning an enormous dataset with generous processing time is one thing, but ingesting activity across fragmented markets, reconciling discrepancies and delivering dependable information a few seconds later is another level of difficulty entirely.

The truth about crypto data

I have sat with fund managers, traders and people from banks and asked them to walk me through how they assess the market. Several familiar public dashboards often appear on the screen. When we compare some of those figures with the data we see internally, the numbers can diverge substantially.

Indeed, people have adapted their workflows around uncertainty. They use a figure to understand direction rather than establish an exact number. A protocol appears to be growing, a particular DEX seems to be gaining market share, liquidity has risen sharply, and volume sits somewhere around the expected order of magnitude.

Whether the true figure is $1 billion or $2 billion can become secondary when the user mainly wants to know which protocol is attracting activity.

Over time, “relatively accurate” became an accepted standard.

For much of crypto’s history, that tolerance carried limited consequences. A trader scanning hundreds of small tokens could still learn plenty from an imperfect dashboard. The data acted as a rough map of where activity was concentrating and where interest was fading.

Professional users subsequently grew accustomed to lowering their expectations around precision. That habit, however, could become very problematic as the financial value resting on blockchain data grows.

The assets sitting on top of the data are changing

Crypto now supports a much wider range of financial assets.

For one, tokenised equities are entering regulated markets in the United States. In September 2026, the SEC granted temporary conditional relief designed to facilitate onchain trading of certain tokenised U.S. stocks. Government bonds have appeared in tokenised form, while tokenised gold and Treasury products have developed into established onchain categories.

Directional accuracy may be sufficient for spotting momentum around a memecoin, but a financial system containing tokenised stocks, government debt, commodities, stablecoins and crypto-native assets requires far greater precision from the data used to assess it.

The same applies to valuing a tokenised equity, assessing collateral, measuring liquidity or producing figures that feed into a fund’s NAV. Errors acquire financial consequences.

Professionals are already making decisions worth millions and billions of dollars using data whose provenance, coverage and methodology they may understand only partially.

The tolerance once associated with crypto analytics is becoming harder to sustain.

Why nobody fixed it

Refinement sits deep inside the data engine. It absorbs engineering time, computing resources and specialist attention while producing a number that users would expect to be right anyway.

Customers have grown accustomed to paying for access, speed and polished interfaces. Accuracy, by contrast, has proved harder to commercialise. When two dashboards produce different figures, the discrepancy can easily become an accepted feature of the market.

Money flowed into gathering more data and presenting it more attractively, while reconciliation received a smaller share of attention.

Birdeye spent much of its own history operating within the same incentives. We prioritised coverage, speed and access to enormous quantities of onchain activity. As our users became more sophisticated, and as larger financial decisions began to depend on the figures they consumed, the cost of imprecision became harder to dismiss.

High-quality refinement is expensive because most of the work disappears into the final result. A client sees a single figure. Behind it may sit millions of records, conflicting sources, classification decisions and repeated checks designed to ensure that the figure deserves to be trusted.

The race to refine

Oil became extraordinarily valuable because an industrial system emerged around turning inconsistent raw material into products manufactured to demanding specifications.

An airline buys fuel whose composition and performance can be trusted each time an aircraft leaves the ground. The value lies in the finished product and the reliability behind it.

Blockchains, meanwhile, produce more raw information than almost anyone can consume.

The kings and queens of crypto data will be those capable of turning that torrent into information that arrives within seconds, covering the market comprehensively and maintaining the consistency institutions require.

About the author: Kha Nguyen is the co-founder and CEO of Birdeye, an onchain market data company serving traders, developers and institutions.

Advertisement

Free newsletters

Good reporting. Delivered.

Choose the newsletters you want in your inbox.

Free. Unsubscribe anytime.

Choose your newsletters

No PRO membership needed. Privacy policy