The page you were trying to view is not available for your role and region.
Stand-out AI story of the year so far?
How the big four hyperscalers—Alphabet, Amazon, Meta, Microsoft—plan to spend US$800bn in 2026 on building data centres for AI initiatives such as Anthropic and OpenAI’s large language models (LLMs)? Or how the Magnificent 7, which have been at the vanguard of the AI boom (think Nvidia) were collectively down 2% in the first half of 2026? Or how semiconductor manufacturers assumed the AI leadership mantle (the US semiconductor index doubled during the first half)? No lack of contenders, but so far it appears the AI story of the year could be the great de-rating of perceived losers to the technology.
As well as on this page, you can listen to this podcast on:
Shooting first
Back in January and February, stocks in sectors such as information services and software suffered steep price falls. RELX and Wolters Kluwer are down 37% and 39% respectively, while the equivalent numbers for LSEG and S&P Global are down 18% and 24%. And yet, there was no spate of profit or revenue warnings. Instead, the declines were triggered by the release of a series of AI productivity tools. A classic case of shoot first, ask questions later. So, with the shots fired, it’s now time to ask the questions.
It was always going to be this way
If you look back through history, innovation typically leads to disruption. Think railways and canals. Before railways, canals were the primary method of transporting goods around the UK. Then along came railways. Today canals are largely the preserve of boating enthusiasts. Fast forward to 2026 and, just as the internet disrupted the distribution of information, AI is bringing down the cost of intelligence and improving the speed we gain access to knowledge.
The problem is, many businesses which seemingly had moats protecting them from competition based around intellectual property and data could now be at risk. The new coding plugins prompted markets to ask how many customers will use AI to do the data exploration and analytics themselves?
It’s never that simple
Judging by share price moves and the indiscriminate selling seen, the market believes the answer could be a lot. The trouble with indiscriminate selling is, well, that it is indiscriminate. And this appears harsh. After all, no two companies are the same. For some companies, any AI risk to revenues appears some way off.
Strong moat
Take S&P Global, the credit ratings and index business. It’s effectively a regulatory utility—bond issuers (companies and governments) require credit ratings to enable investors to buy their debt. Trust in the brand and the methodology used is crucial. This reinforces the moat by making it harder for AI to replace a globally accepted ratings agency. On the index side, trillions of dollars of passive funds track S&P’s benchmarks, and switching benchmarks is extremely rare as it is a complex process with plenty of governance and operational risks.
The nature of its activities helps explain why S&P has avoided meaningful disruption from emerging technologies in the past. There’s no reason to suggest this should not hold with AI. And yet the 12-month forward rating of its shares has fallen to sub-20 times earnings, having previously been in the high-20s. S&P Global, an example of a de-rated stock with a strong moat seemingly intact.
Leaky moats?
Not all companies are in such a relatively strong position. Some have clear AI vulnerabilities. LSEG’s workspace terminals, for example, are widely used by traders and analysts. The concern is AI agents using LLM interfaces could replicate these terminals at a fraction of the cost. So, is the market right to be concerned? Yes and no.
LSEG is much more than a workspace terminals’ provider. Like S&P, LSEG has an index business. It also owns London Clearing House, a middleman between buyers and sellers in financial markets. And it operates trading venues such as the London Stock Exchange and has large proprietary datasets. All assets that could become more valuable were AI to drive more data consumption. Markets though have chosen to focus on the AI threat. LSEG shares currently trade on 16 times 12-month forward earnings compared to a mid-20s rating a year ago.
It’s a similar story in business services. Historically, the sector’s valuation proposition rested on companies owning, creating and curating datasets. They extract insights from vast unstructured information for end-customers. On top of that they provide analytics tools that help professionals, such as lawyers, compliance professionals and accountants, become more efficient and better informed. Then along comes AI, promising to spark new competition (in the form of start-ups), disintermediation (where companies become passive databases) and end-customer disruption (budget cuts and headcount reductions). Easy to see why shares across the sector have been on the backfoot.
Equally, it’s easy to argue how the de-ratings are overdone. True, RELX’s legal division appears most at risk from AI. But legal is just one of four segments, alongside exhibitions, risk, and science, technological and medical publishing. Legal represents only about 12% of group earnings. But the scale of share price weakness suggests real concerns over contagion across the business.
Soft software?
Software has also been at the vanguard of the AI-disruption story. And arguably for good reason. More companies are looking to implement AI initiatives but the likes of Anthropic and OpenAI have to be paid for the use of their LLMs. Typically, corporate AI funding comes from IT budgets. These though are not being materially increased, meaning money has to be diverted from elsewhere. One area where companies can cut back is software because they now have more options. Products used to be bought from big software vendors but now there is the do-it-yourself option using LLMs or there are start-ups providing alternatives.
Software budgets could be vulnerable then. But beware of generalisations. The key is to look at what is the value proposition offered by the software company. If it is Adobe and they offer image/document creation, then yes that’s a potentially very disrupted space as plenty of AI offerings create images. A company with a core ERP (Enterprise Resource Planning) system by contrast offers more than just software—regulatory compliance, business logic, and security are all bundled in too. The value proposition is notably different from a one-off product. It pays to discriminate.
Not standing still
And we haven’t touched on the opportunities presented by AI. Credit-checker Experian estimates its AI-enabled total addressable market at around £15bn. That’s on top of its existing markets. The company is also seeing AI-generated efficiency gains—the productivity of its coders for example is up 15% while labour costs as a percentage of sales have come down significantly over the last couple of years. LSEG too is actively looking to harness AI opportunities via its MCP (Model Context Protocol) Connector—the software that connects AI to external data sources.
What will it take?
Despite the above, share prices have barely budged since the great de-rating. It begs the question what can companies do? Share buybacks? These send a positive message (management believing the shares are undervalued) but they are no panacea. Salesforce has been executing an aggressive US$50bn buyback but the share price has barely budged.
Companies can also start to innovate and invest more to drive AI revenues. This is what Wolters Kluwer is doing. The information services business is increasing product development spend this year to 12%-13% of sales, up from 11% previously. For technology companies, innovation has always been important, but this is particularly true in periods of disruption.
The bottom line is that share prices are being kept at current levels because of opportunity cost. Investors see revenue acceleration at semiconductor manufacturers but not in software or information services companies. If follows that tangible revenue acceleration from AI initiatives will help catch the market’s attention.
Another AI story?
Unlike the great de-rating where stocks were sold indiscriminately, when it comes to potential re-ratings, the market will likely discriminate positively on a stock-by-stock basis. Companies may need to be able to show they are not AI losers but winners (or at least that their moats can stay intact). The makings of yet another AI story perhaps? A classic comeback story made up of lots of AI-generated ones…
Important information
This material is a marketing communication provided for information purposes only and does not constitute independent investment research. References to financial instruments are for general information purposes and are not subject to requirements applicable to independent investment research.
Any references to securities or financial instruments should not be regarded as a personal recommendation, or as an offer, solicitation or invitation to buy or sell any financial instruments. The views expressed are those of the authors at the time of publication and are subject to change. Past performance is not a reliable indicator of future results.
This material does not constitute tax, legal or accounting advice. You should seek independent professional advice appropriate to your individual circumstances before making any financial decision or engaging in any transaction.
