Top
These companies have a decent opportunity to win from Advanced AI adoption. Tools that augment analysts’ productivity, accelerate credit assessment, or improve client interfaces have the potential to enhance a company’s revenues and reduce costs rather than threaten margins."
In recent months, investors have become preoccupied with the question: will advanced artificial intelligence (AI), be it generative (GenAI) or agentic, disrupt the data-rich industries that underpin our modern world? From software through consulting to information services, credit bureaux, exchanges and insurance brokers, few data-rich industries have escaped the market’s blanket derating.
At the time of writing, the sell-off in financial information companies such as S&P Global, MSCI, Moody’s and LSEG has been swift and indiscriminate, reflecting a growing fear that Advanced AI could erode these companies’ competitive advantages. In this piece we home in on this single industry and our assessment of companies’ moats in the face of potential disruption.
As we see it, such sweeping reactions often mark moments of opportunity. For example, MSCI’s share price pullback early in the third quarter offered an attractive entry point into a company we believe can offer steady margin expansion and earnings growth, fuelled by ongoing revenue increases and operating leverage. The market’s instinct to “act first, analyse later” can overlook the complexity of what actually constitutes a durable moat in the world of data. As bottom-up high quality investors, our task is to assess, on a case-by-case basis, where new technology can and cannot erode the foundations of enduring franchises.
Three shades of disruption
We find it helpful to categorise Advanced AI-related disruption into three broad types. The first and most severe – what we call systemic disruption – occurs when a new technology enables the almost overnight creation of a credible “new tech” competitor, undermining market share and network effects. In software, this might mean a small, agile competitor suddenly matching an incumbent’s capabilities with minimal investment. Google’s success with Nano Banana against advanced Photoshop products is a recent example.
The second, material disruption, reshapes business models rather than eliminating them. Here, incumbents may need to reinvent pricing structures or delivery models to stay relevant – as we saw in software companies with the rise of the cloud.
Finally, there is secondary disruption, where Advanced AI simply lowers input costs or accelerates product development. This is the bucket into which most financial data and analytics companies fall. GenAI makes data collection and code writing cheaper, but the underlying business model and client demand – for trusted, accurate, regulated information – remains largely unchanged. Provided the company successfully innovates, Advanced AI may be an advantage rather than a threat.
The real moats: data, ecosystems, and integration
We believe financial information services are not easily replicated. Their moats are built on several interlocking defences.
The first is data ownership. Proprietary or branded datasets, like MSCI’s indices or Moody’s credit ratings, are effectively irreplaceable. While anyone can, in principle, build equity indices or opine on corporate creditworthiness (and many have tried!), it is the trust and familiarity of MSCI and Moody’s branded datasets that form an enduring moat. However, even ownership of clean public data can be a source of competitive differentiation. Re-creating decades of historical, verified data – with consistent identifiers and linkages – is both technically challenging and prohibitively expensive.
The second is the ecosystem effect. Ratings and indices function as a common language for global capital markets; they work because everyone else uses them. Once embedded in regulatory frameworks, benchmarks, and investment mandates, they become part of the financial system’s infrastructure – as is the case with S&P Global.
Third, integration into client workflows creates powerful switching costs. Platforms are wired into the daily operations of asset managers, traders and analysts – via application programming interfaces (APIs), terminals, and risk systems – making disruption a costly and risky endeavour.
Other defences include global distribution networks, enterprise-grade security that satisfies regulatory scrutiny, and increasingly frictionless user interfaces (UI). While UI innovation may be easy for start-ups to copy, integration, trust and compliance are not.
“Data activators”: moats built on language and trust
S&P Global, Moody’s and MSCI – what we call “data activators” – sit at the top of the information services value chain. Their key branded franchises convert raw data into actionable insight and, crucially, commonly understood vernacular. A Moody’s credit rating or an MSCI index does not derive its value from the difficulty of computation but from its universal recognition and acceptance – a good example of the ecosystem effect in action.
Over 70% of the profits of these firms stem from areas that are, in our view, largely insulated from Advanced AI disruption. While some secondary exposure exists – such as MSCI’s ESG data aggregation or S&P Global’s financial desktop offering – we believe these risks are modest compared with the market’s recent reaction.
Equally, these companies have a decent opportunity to win from Advanced AI adoption. Tools that augment analysts’ productivity, accelerate credit assessment, or improve client interfaces have the potential to enhance a company’s revenues and reduce costs rather than threaten margins. The ability to train models on a company’s extensive proprietary data is also an under-appreciated advantage. When the likes of S&P or MSCI embed Advanced AI within their own infrastructure, it raises the competitive bar for would-be disruptors rather than lowering it.
“Data aggregators”: shallower moats, still defensible
By contrast, firms such as FactSet or parts of the London Stock Exchange Group (LSEG) act as “data aggregators”, curating and delivering third-party information rather than owning data. Their moats are not built on ownership of unique branded data, but on how they organise, validate, and securely integrate vast volumes of information for clients.
For institutional investors who upload millions of sensitive portfolios each day, trust and reliability matter more than novelty. We believe the market’s assumption that AI start-ups will easily replicate such infrastructure ignores the operational and regulatory constraints of handling confidential financial data. Clients have a very low tolerance for errors. Few chief risk officers would be comfortable feeding their firm’s trading positions into an untested AI model hosted by a venture-backed challenger.
Similarly, the “plumbing” of live market data – fibre connections to exchanges, cloud-based transmission, and entitlement systems – is capital-intensive and highly regulated. Even where Advanced AI lowers coding costs, new entrants still face substantial barriers to matching the incumbents’ physical and compliance infrastructure.
While aggregators’ moats may be shallower than those of activators, they are far from non-existent. Their strength lies in integration, distribution and security – attributes that are difficult to automate and essential for client retention.
Advanced AI as opportunity, not existential threat
The irony of the recent broad-based sell-off is that Advanced AI, rather than undermining incumbents, may ultimately strengthen them. By reducing the cost of coding, data cleansing and interface design, Advanced AI enables faster innovation and productivity gains. For companies already operating at scale with established trust and distribution, these tools can deepen competitive advantage rather than erode it, and those with pricing power are able to hold on to revenue gains.
In credit ratings, for instance, Advanced AI can help analysts process new issuers faster, shortening time-to-market and improving margins. In risk analytics, Advanced AI-enabled interfaces may widen product adoption across investment teams. These are incremental yet meaningful enhancements that reinforce the incumbents’ position.
Separating signal from noise
Markets rarely price nuance well. When fear of disruption becomes indiscriminate, quality investors can play offence rather than defence. There remains a case for selection: it’s important to probe every business model for resilience, as well as take account of valuation and position size. But the financial information sector’s comprehensive recent de-rating looks to us like a textbook example of the market throwing out the proverbial baby with the bathwater1.
We continue to view S&P Global, Moody’s and MSCI as among the most resilient franchises in global data. We believe their defences – proprietary data, ecosystem embeddedness, workflow integration and distribution reach – are formidable. The so-called “AI threat” may, in time, become a driver of renewed efficiency and innovation across their businesses.
As ever, our approach remains disciplined and long-term focused: to distinguish genuine structural risk from cyclical narrative, to focus on quality and valuation , and to remember that durable moats – especially those built on trust, regulation and reputation – tend to outlast the headlines.
Featured Insights
This image indicates content designed specifically for Financial Advisors / Investment Professionals. This material is not to be used with the public.