rewrite this title AI is pushing crypto media into a fight over trusted market data

Artificial intelligence is commoditizing news and routine research across the industry, and it’s pushing crypto media companies to rebuild themselves as data platforms, analytics providers, and institutional infrastructure. The firms racing to assemble AI-ready databases are positioning to become the industry’s reference layer, the source that investors, regulators, and algorithms rely on to understand digital assets.

On June 12, Blockworks acquired Messari, folding two of the largest data and research operations in crypto into a single platform covering more than 40,000 digital assets.

The Wall Street Journal put the price above $10 million, a steep markdown from the roughly $300 million valuation Messari carried after its 2022 Series B, and that discount tells you how much the economics of crypto information have changed in four years.

Blockworks raised money in April at a $192 million valuation in a round led by ParaFi Capital and Reciprocal Ventures, with Coinbase Ventures participating, and it openly said it intended to use that capital to buy competitors. Co-founder Jason Yanowitz has described the goal in plain terms: he wants to build the Bloomberg of crypto.

The Messari purchase reflects a shift that’s been building underneath the AI hype for a couple of years. The value in financial information is moving away from the article and toward the database on which the article was built. The companies positioned to dominate the next few years are the ones that own the canonical datasets institutions and machines treat as authoritative.

A research and reference operation of that kind earns its money from feeds, terminals, and API calls rather than from readers, and it answers to compliance officers and quants more than to an audience, which makes it a structurally different business than the newsroom it usually grows out of.

Why publishing stopped being an advantage in the age of AI

The pressure on media companies starts with distribution, where the traffic that funded digital publishing for two decades is steadily draining away across the broader media business. Google search referrals to publishers fell about 33% globally in the year leading up.

to November 2025, according to the Reuters Institute’s annual trends report, with US referrals down 38% and European referrals down 17%, while referrals from Google Discover dropped 21%. By early 2026, roughly 58% of Google searches ended without a click to any outside site, as AI-generated summaries answered the question on the results page.

Penske Media has taken Google to court over the change, arguing that the search company is cannibalizing the traffic that publishers were promised in exchange for allowing their work to be indexed.

For a crypto outlet, the consequence is that breaking news and routine explainers, the formats that carried the traffic model for years, are worth steadily less each quarter. A summary of a token unlock or a treasury disclosure gets generated in seconds and consumed inside a chat window, and the click that used to follow it is now completely gone.

Every financial market tends to move through the same sequence as it matures. It starts with reporting and opinion, when information is scarce, and simply explaining a new asset class builds an audience. It moves into research, as institutions arrive and want context and frameworks rather than headlines.

It then standardizes into data, when investors would rather query a database than read fifty notes on the same thing. And it ends in infrastructure, where that data essentially becomes the workflows that the market can’t operate without.

Bloomberg reached that final stage decades ago, which is why it earns somewhere around $11 billion a year, charges close to $31,980 for a single terminal seat in 2026, and keeps more than 325,000 subscribers wired into its system.

The journalism it produces is just a nice little bonus next to its core business. The reason markets can’t switch the terminal off is the information that feeds their models, pricing, and compliance systems.

Crypto is entering that fourth stage, and by Yanowitz’s reckoning, it could get there much faster than equities did. Building a research and reference operation in traditional markets required large teams of human analysts to key in filings by hand, whereas crypto generates structured, real-time, machine-readable information natively, both on-chain and through standardized disclosures, making it an ideal input for automated systems.

CryptoSlate’s own reporting has tracked corporate AI adoption climbing from 8.7% in 2023 to 14.2% in 2024 and 20.2% in 2025 on OECD figures, and the agents doing the consuming are beginning to transact on their own.

What does a reference layer control?

Once a market reaches that stage, whoever controls the reference data has leverage over everyone downstream, because asset managers price portfolios off it, index providers build products around it, exchanges wire it into their systems, regulators cite it, and AI models train on it.

A company that owns the canonical figure for a protocol’s circulating supply or a treasury’s holdings can shape how billions of dollars get allocated without ever publishing an opinion about any of it, and the future gatekeepers in this market are the database operators who sit further upstream than any editor ever did.

Consolidation of power and influence is already underway, and Blockworks’ acquisition of Messari is only the latest example. Paris-based Kaiko acquired Amberdata earlier in June to deepen its derivatives and on-chain coverage and add AI-focused research tools for banks, asset managers, and hedge funds.

In January, the oracle provider RedStone bought Security Token Market along with a dataset spanning more than 800 tokenized assets. Each of these deals pulled fragmented sources of very valuable information into fewer hands.

The reason this is more important than ordinary media consolidation is what institutions need before they can scale into digital assets. Large allocators require standardized disclosures, clean historical datasets, legal-entity mappings, governance archives, and risk metrics that they can defend to their own compliance committees.

Crypto has already institutionalized custody, settlement, and trading; information is the piece being institutionalized now, and the demand for trustworthy data grows alongside the demand for capital.

AI raises the stakes of all this rather than lowering them. In the very near future, an analyst will rarely open a protocol’s documentation by hand and will instead ask a model to compare every Layer 1 network on treasury composition, validator concentration, governance participation, and revenue.

The quality of that answer depends on which databases the model has been trained to trust, so whichever companies own those datasets are the chokepoint that every automated comparison has to pass through. That position compounds over time, because each new institutional or machine consumer makes the underlying data more valuable and a little harder to dislodge.

Established publications have been feeling this pressure for a while. The economics of publishing on its own keep getting tougher as distribution fragments and machines absorb routine reporting, eroding the advertising and referral revenue that funded newsrooms for years.

However, they are also sitting on years of reporting, structured metadata, proprietary research, and editorial credibility, and that archive can become the raw material for institutional intelligence products and the AI-ready knowledge bases that models depend on.

The durable position for crypto media may be to supply the trusted information layer that AI consumes, while holding onto the editorial judgment that decides what belongs inside it.

Crypto was built to remove trusted intermediaries from money and allow people to transact without a bank or a clearinghouse standing in the middle.

As institutions and AI move in, it’s begun assembling a fresh set of trusted intermediaries that sit over its information, and the companies that end up owning the canonical datasets, the supply figures, the governance records, and the on-chain metrics that every investor, regulator, exchange, and model treats as ground truth could hold more influence than any newsroom ever did.