The point: information is the raw material
London Stock Exchange Group is not spending $3.535 billion to buy another exchange. It plans to acquire Octus, a company that collects data, documents and intelligence on public and private credit. The deal, announced on October 22 and still subject to approvals, makes a specific change visible: as artificial intelligence makes information cheaper to process, the value rises for those who own information that is difficult to obtain, verify and connect.
The problem is especially clear in private credit. A listed bond leaves observable prices and trades; a loan negotiated between a fund and a company may remain in documents, clauses and updates distributed among a few participants. Asking a model to write a summary is not enough to understand what it is worth and what risks it contains. Contracts, creditor priorities, collateral, maturities, covenants and the issuer’s history are required. If those data do not enter the system, AI can only organize a void.
LSEG is therefore buying a collection and workflow chain, not a static archive. Octus covers leveraged credit, distressed debt and private markets; it turns scattered sources into information that analysts, investors and advisers can compare. The company is expected to generate between $300 million and $320 million in revenue in 2026, according to LSEG. That is the buyer’s forecast, not an already certified result of the integration.
What LSEG is really buying
Octus, previously called Reorg, combines specialist journalism, data, research and workflow tools. A customer is not merely looking for news about an indebted company. They need to reconstruct who will be paid first, which conditions could trigger a renegotiation and how a change alters the value of different tranches. The useful product emerges when the original document, the update and the financial model remain connected.
LSEG already owns Refinitiv, FTSE Russell indices, market platforms and data services. With Octus it wants to extend its presence along the credit cycle: from discovering an opportunity to analysis, trading and portfolio monitoring. The move sits alongside the agreement under which LSEG-controlled Tradeweb plans to acquire Trumid, an electronic platform for corporate bonds. Together, the transactions bring content, analysis and the place of execution closer together.
This explains why the price cannot be judged simply by counting the files Octus owns. The value depends on how often professionals use those data, how difficult they are to replace and whether they can be distributed through products already on customers’ desks. LSEG will finance the purchase with new debt and expects the deal to be neutral to earnings per share at first and accretive thereafter. These too are corporate expectations: integration costs, customer renewals and approvals can change the outcome.
Why AI makes non-replicable data valuable
A generative model can read thousands of pages, extract clauses and compare scenarios faster than a human team. It cannot know that a new document exists if nobody has collected it, nor can it determine on its own which version is valid. It can also produce an answer that is plausible but wrong. In credit, where one word can change the order of repayment, the link to the source is not an editorial detail: it is part of the product.
The most useful combination is therefore less spectacular than an autonomous adviser. A system finds relevant documents, normalizes names and terms, suggests connections and presents the passage supporting a conclusion. The analyst checks provenance, date and context before acting. Proprietary data and AI reinforce each other: the former give the model current, traceable material; the latter reduces the time needed to work through it.
This complementarity also creates a barrier. Two providers may use similar models, but they do not necessarily hold the same document rights, correction history or network of journalists and analysts. Reuters reports that LSEG presents proprietary data precisely as a defense against AI-driven disintermediation. The defense holds only if customers continue to regard those data as more reliable and complete than cheaper alternatives.
Private credit remains private
More tools do not automatically mean more transparency for everyone. A platform can make a loan legible to its subscribers without turning it into a security with public prices. Private credit is growing precisely because companies and lenders value confidential negotiations, flexible structures and timelines different from those of listed markets. Octus can reduce the cost of finding information; it does not remove the economic reason why much information remains restricted.
For a fund or bank, a more complete database can reveal a fragile covenant or concentrated exposure earlier. For the borrowing company, it can increase creditors’ ability to compare terms and respond. For savers, the effect is indirect: pension funds, insurers and mutual funds may be exposed to vehicles investing in private assets. The quality of the data professionals use affects valuations, but it does not guarantee that a reported price reflects a sale that could take place tomorrow.
There is also a concentration issue. When the same group supplies indices, data, terminals and market access, it can provide a more efficient workflow; it also becomes a central gateway for many customers. Subscription prices, system compatibility and the ability to export data determine how much of the efficiency gain is shared and how much remains with the platform owner.
The evidence will come after the acquisition
The announcement already has a price and a strategy, but not yet an outcome. Approvals are required; LSEG will then have to integrate people, products and archives without losing the speed that makes Octus intelligence useful. It will also have to show that customers are buying more than a generated summary: exclusive data, traceable evidence and better decisions.
The signals to watch are concrete: how many customers renew, which products truly combine public and private data, how much revenue grows without price increases that push users elsewhere, and whether the tools preserve the link to original documents. LSEG is betting that in the AI era, scarcity is not the ability to write an answer. It is the ability to prove it.
