Good morning, and welcome to the LSEG Q3 results. At this time, all participants are in listen-only mode. Later, we will conduct a question-and-answer session through the phone lines, and instructions will follow at that time. I would like to remind all participants that this call is being recorded. I will now hand over to David Schwimmer, CEO of LSEG, and Michel-Alain Proch, CFO, to begin the presentation. Please go ahead.
Good morning. Thanks for joining the call. I am here with MAP, our CFO, and Peregrine, Head of Investor Relations. We continue to deliver very strong growth. Revenue is up 8.8% in the third quarter and 8.6% year-to-date. All our businesses are contributing to this growth. Markets continues its outstanding year, with our sustained investment ensuring it is well-positioned to help customers navigate the ongoing economic and political uncertainty. Our subscription business has accelerated, growing at 6.6% in Q3, up from 6.3% at H1, and on track to deliver 6.5% for the full-year. We are raising 2026 guidance again to reflect this continuing strength. We now expect full-year revenue of at least 8%, putting us on track for the strongest year of growth since 2018. We are also raising our guidance for EBITDA margins and cash generation. MAP will walk you through those in a moment.
Taken together, our revised guidance speaks to a business that is growing faster, is more profitable, and is generating more cash than anticipated at the start of the year. Our strong cash generation gives us significant flexibility in capital allocation, combining bolt-on M&A and significant cash returns to shareholders. I have spoken before about the culture of innovation we have embedded at LSEG and the pace of product development this is driving. Last week, we announced a partnership with NVIDIA that will reduce the time for customers to analyze our massive Tick History data set from days to minutes. Last month, we launched World-Check Resolve, a new AI layer which reduces false positives by around 75% and saves our customers a lot of manual checking. Where we have these leading positions, we continue to invest to maintain that gap and drive growth. On AI, we are driving adoption at scale.
95% of our top 200 customers are now using our AI solutions. I will come back to this in a moment. As you will have seen, we have had two really exciting developments this week. On Monday, Tradeweb announced its acquisition of Trumid, a high-growth electronic credit trading marketplace. Billy and Sarah have already taken investors through the logic of this deal, but I will recap this briefly later. This morning, we have confirmed the acquisition of Octus, a leader in the provision of credit solutions in the illiquid and private credit space, backed by highly proprietary data. This deal fits squarely within our criteria in terms of both strategic fit and financial discipline. It is highly complementary to our existing fixed-income product suite and enhancing to growth. Much more on that later.
Finally, on capital allocation, we have made record returns to shareholders this year, returning GBP 3.4 billion in the first nine months. With GBP 800 million remaining on our current buyback program in Q4 and Q1 of next year, we are on track to return around 10% of our market cap in total since January of this year. We also have capacity for an additional GBP 1 billion- GBP 1.5 billion in 2027. We have built a business capable of partnering with customers through the whole trade life cycle and data value chain across multiple asset classes. Trading and data are ever more intertwined as financial markets grow and increase in sophistication, reinforcing our strategy and strengthening our position. It is inevitable that the external environment will evolve, whether that is asset prices, issuance trends, market volatility, or GDP growth.
The balanced and diversified nature of our subscription and transaction revenues and our alignment with multiple long-term growth drivers position us to perform well in almost any environment. You can see the power of this all-weather business model in the steady and accelerating growth on this slide. Before I hand to MAP, let me update you on the significant progress we are making in the delivery of our AI solutions and how they are resonating with our customers. You will recall our focus for this year was rolling out solutions and driving customer adoption. With 95% of our largest customers using at least one of our AI solutions and 80% of our top 50 customers using two or more, we have made real progress in driving adoption in Q3. We had 47,000 users of Workspace AI Search at the end of Q3 and have just passed the 50,000 mark.
User engagement with the product also continues to deepen, with the average queries per user up 30%. Crucial to this success are the ongoing enhancements we are making, including scheduled queries through Microsoft Copilot, a bespoke library of user prompts, and greater export functionality. LSEG Everywhere continues to build. Over 2/3 of our largest customers and over 300 in total have now connected or are onboarding to our MCP. Our customer engagement on enterprise AI goes far beyond just MCP. As highlighted at the first half, customers increasingly want to access our AI-ready data via multiple distribution channels, including APIs, cloud platforms, and traditional bulk feeds. Overall, we saw a 50% increase in AI-related commercial discussions in Q3. 80% of AI license sales involved a customer buying new or additional data sets. With new AI distribution channels, our customers want more of our data.
AI is enhancing the value of our data, showing that our Data & Analytics business is more valuable in an AI world. With that, let me hand over to MAP to take you through our financial performance in more detail.
Thank you, David, and good morning to all of you. As you just heard, we delivered another strong financial performance in the third quarter. Organic revenue up 8.8%. Data & Analytics was up 5.2%, as our ongoing innovation and investment in content drives demand across our solutions. Data and feeds continue to be particularly strong, up 7.5%. FTSE Russell was up 10.2%, with subscription revenues re-accelerating as we said it would. We continue to see sustained demand for our screening and identity verification services in Risk Intelligence, which grew 9.8% in Q3. Taken together, our subscription businesses grew 6.6%, a good acceleration from the 6.3% in the first half. Our market division is sustaining its very strong performance, up over 12%, with strengths across all the businesses. OTC derivatives was the standout performer, with continued strength at SwapClear and almost 30% growth in client volumes in Q3.
We continue to work with the industry to drive adoption of our Post Trade Solutions for uncleared OTC products, signing four new members to TradeAgent in the quarter. Tradeweb also grew strongly, benefiting from sustained innovation and investment against a backdrop of shifting interest rate expectation. That drove 22% growth in volumes in Q3. Finally, cost of sales grew below revenues, benefiting from the change to the SwapClear agreement we made at the end of 2025. As a result, we delivered almost 10% growth in gross profit for the quarter. Turning to our subscription revenues. We are well on track to deliver full-year subscription growth of around 6.5%, an acceleration from the 6% growth in 2025. These revenues are highly recurring, typically around 93% annually, and that brings good visibility. New sales made in previous quarters feed through to revenues once a product is installed and activated.
We have a good line of sight to these timelines too. The ongoing investment we make into content, capabilities, and distribution is reflected in product pricing and annual increase in LSEG data access agreement, representing 18% of D&A revenues. Over time, AI revenues will make a growing contribution to these subscription revenues with a long runway for growth. Turning now to our Markets division, which has seen another outstanding quarter. Volume in interest rate swaps on our trading and clearing platforms were very strong in Q3 as customers turn to our solution amidst economic and political uncertainty. These dynamics also supported strong volumes across fixed income equities and FX, as you can see on this slide. Putting market condition to one side and taking a long-term view, these are leading businesses with strong structural growth drivers.
Our sustained investment is helping drive this growth by developing new solution and expanding reach. Through this strategy, we have delivered a five-year CAGR of over 9%. Turning to our full-year guidance, most of which we are raising once more. On revenues, we expect organic revenue growth in 2026 to be at least 8%. This is a meaningful increase to the guidance we issue at H1. We are raising our margin guidance again, expecting between 100 basis points and 120 basis points of improvement this year. This reflects our continued strong top-line performance and our focus on efficient management and delivery of operating leverage. We reiterate our CapEx guidance at 9.5%. Finally, our guidance for free cash flow. That already reflected the highly cash generative nature of our business model.
But we think we can do even more this year and now expect equity free cash flow to be at least GBP 2.8 billion. As David said earlier, our strong cash generation is allowing us both to execute on M&A when it makes sense and return cash to shareholders. You can see on this slide that we have returned a record amount of capital to shareholders this year, and we are on track to return around 10% of our market capitalization in a little over 12 months. With that, I will hand back to you, David.
Thanks, MAP. I'll now spend a few minutes focusing on the Octus acquisition. We are really excited about this transaction. It's a great business, and the timing is excellent given the work we have done already to transform our data platform and scale up in fixed income and private company data. I'll explore all of these points in more detail, but to summarize here, Octus is a leading provider of credit solutions with a focus on illiquid and private credit. Its data is highly proprietary. Its financial track record is outstanding, with very strong growth and improving margins in a market which will continue to grow at a double-digit rate. Our margin guidance is unaffected by this deal. Octus coverage and customer footprint are highly complementary to LSEG's and will unlock significant product and cross-sell opportunities.
The transaction will be accretive to our top-line growth and accretive to EPS after 2028, and we are retaining a high-class management team led by Kent Collier. As MAP will lay out, this deal fits well in the context of our broader capital allocation strategy. We're completing our current GBP 3 billion buyback, and we believe investing in our shares at this level continues to be a good use of capital. We can see capacity for a further GBP 1 billion-GBP 1.5 billion of buybacks in 2027, taking us to the top end of our leverage range a year from now. Octus looks a lot like our D&A business, but focused on a specialized asset class, illiquid and private credit. For example, distressed debt, leverage loans, and CLOs. Its solutions span the full credit life cycle, idea generation, deal tracking, detailed analysis, underwriting, and portfolio and risk management tools.
As with D&A, it delivers through feeds, APIs, and workflow platforms. These solutions are based on differentiated and proprietary datasets. Private credit is not like the investment-grade corporate bond market. Documentation can be extremely hard to come by. Structures can be highly bespoke and complex, and sometimes you have to spend days sitting in courtrooms following legal processes. Even accurate financials on private companies can be hard to trace, as many of you have found over the last few days. Octus has been building up this data and associated insight and expertise for over 12 years, and it includes thousands of contributory records from customers, issuers, and investors, often detailed documents on specific loans. This is just not possible to replicate. We've done extensive customer due diligence, and it is clear from the feedback that Octus solutions are extremely important for day-to-day research, workflow, and risk management.
Like our existing data assets, this is a business that is becoming more valuable in an AI world. Octus has built strong AI functionality with MCP connectors and its own credit AI platform, bringing proprietary intelligence and analytics directly into customer workflows. To give you a couple of stats, just in 2026, its AI platform has generated more than 4 million automated classifications of complex market data and had 500,000 tool calls to its MCP. Octus is a high-growth business in a high-growth market. Revenue has grown at a strong double-digit CAGR since 2022. In fact, it has compounded at a double-digit rate since its inception. It now has over 750 customers and 40,000 individual users with very high retention rates. The private credit market is growing at around 10%-15% annually.
There are more participants every year, and as I just mentioned, greater deal complexity just drives increasing demand for specialist solutions. The user base is broadening from investors and advisors to underwriters, risk managers, and compliance teams. As we are seeing in our own business, AI is driving a range of different delivery mechanisms, increasingly delivering Octus proprietary credit solutions directly into customer workflows. Revenues are currently weighted around 80% to the buy side and legal profession, and around 70% to the Americas. As I will cover on the next slide, that is highly complementary to LSEG. Let us look at how these businesses will fit together and create opportunities for additional growth. First, the combination of LSEG's more liquid and public credit coverage combines with Octus' illiquid and private focus to give coverage of the full credit spectrum.
Second, we are able to drive a step change in our overall private companies coverage, marrying all the work we have done building our private company dataset with Octus' 95% coverage in the private credit universe. This will give us leading private markets coverage. Third, our business today is weighted to the sell side and global institutions, where we have deep strategic partnerships. We are under-penetrated on the buy side, where Octus excels. This presents a significant cross-selling opportunity. Similarly, for our respective geographic footprints, we will bring global distribution for Octus, which until now has been more focused in North America. MAP will spend a couple of minutes on the financial aspects of Octus and the deal.
Yes, David, and on the financial for LSEG, this deal is also highly attractive. We have talked about the revenue and growth profile. This will be accretive to D&A growth and will, of course, be incremental to our 7% guidance for our subscription businesses in 2027. It is also already generating really strong margins, which is a sign of high quality for a business that has achieved such strong growth, and we only expect those to improve as the top-line growth continues. Octus has a roughly 40% adjusted EBITDA margin and a strong operating leverage, which is allowing us to confirm the group EBITDA improvement of 150 basis points over 2027 to 2029 from a 2026 baseline. Putting this together, it is neutral to EPS in 2028 and accretive thereafter. David, back to you.
Thanks, MAP. As you will have seen a couple of days ago, Tradeweb announced its planned acquisition of Trumid, an electronic credit trading platform. You can see how it fits together with Tradeweb on this slide. Just as Octus is for LSEG, Trumid is highly complementary to what Tradeweb does today. It extends Tradeweb's presence into gray markets and new issuance, has great customer penetration with hedge funds, and additional specialized dealer protocols. We think this is a great deal for Tradeweb and LSEG. When you combine it with what Octus does for us and our ongoing organic investments, we have made huge strategic strides with these two transactions. We are building a global leader in fixed income, in proprietary data, news, and analytics, with AI-enabled workflows and execution tools, and with trading venues operating from public investment grade through high yield to private and distressed.
This is a great growth engine. Yes, it has been a big week for M&A, but we think about capital allocation over multiyear periods. We have looked at many deals over the last five years, and very few made the cut in terms of strategic fit or financial return. These acquisitions do both. At the same time, we have bought back GBP 8 billion of our own shares since 2022, accelerating significantly over the last 18 months with our share price depressed. We will continue to buy back shares because we think that provides attractive returns. The current GBP 3 billion program will complete by February next year, which means another GBP 800 million across Q4 and the first two months of 2027.
Then we have capacity for a further GBP 1 billion- GBP 1.5 billion through 2027, which would take us to around the top of our leverage range by the end of next year. I appreciate that is a little longer than our usual Q3 presentation, but there are a lot of great things happening across LSEG. We are delivering record results and raising full-year guidance. We are driving strong adoption of our AI solutions and expect that to make a growing contribution to revenues. We are driving an unprecedented pace of innovation across the business. I have spoken to some of this already, but this is also a good moment to remind you that our Head of Markets, Daniel Maguire, and his team will host an online investor event in December to give more insight into our initiatives across digital markets.
Finally, we continue to invest, both organically and where it makes strategic and financial sense through acquisition to further enhance our customer proposition. Through that strategy, we are building a business that is better positioned than ever for sustained long-term growth and value creation. Now MAP and I are happy to take your questions. Peregrine, over to you.
Thanks, David. As usual, please restrict yourself to one question, and maybe there will be time for you to come back and ask another, at the end. So operator, I will hand to you. Thanks.
Thank you, Peregrine. If you would like to ask a question, please signal by pressing star one to raise your hand and join the queue. To withdraw your question, press star one again. When called upon to ask your question, please ensure that you use your handset and that you are not on mute. Again, star one to raise your hand. Your first question is from the line of Andrew Lowe of Citi. Please go ahead.
Hi. Thanks for the very detailed presentation. I just wanted to ask, on the Octus deal, it fills a gap in your data coverage with private credit. As you look to explore options to fill this, did you look at building this data organically? Could you just talk through what the challenges of doing this were that led you to go down the organic route? Related to that, if you could maybe talk a little bit more about the moats to Octus's data and provide some tangible examples of why the data cannot be found elsewhere. If Octus can access the data, what stops others from replicating its offering over time, and what is the right to win in the market? Thanks.
Thanks, Andrew. Great questions. We always think about the build versus buy equation. On both private markets data and fixed income data, we have been investing substantially in those areas. We have added a significant amount to our fixed income content, our fixed income evaluated pricing, corporate actions, et cetera. We have also added extensively to our private assets data over the last few years, including bringing in Preqin data, Nasdaq eVestment data, but also a lot of our own. What is really special about Octus is how proprietary it is and how hard to source a lot of the data that Octus provides is. Just to give you a couple of examples, there are 12 + years of deals data covering substantial aspects, 95% market coverage. That is data on thousands of deals.
It's information on thousands, 3,000 + public and private issuers over, again, 12 + years. It has thousands of, I'll say 11,000 + transcripts analyzed for structuring and valuation comps. Just a huge library of content, and deal documents across 28,000 issuers and 65,000 deals going back, again, 12+ years. That in and of itself, huge amount of highly valuable data. The content that Octus has, this is not publicly available. This is often only available through confidentiality agreements in secure data rooms with restricted rights. Octus has aggregated and anonymized rights to certain private data that is protected by those confidentiality provisions. So no one else is getting access to that information. On top of that, Octus has about 280 experts, financial and legal experts who are providing their own analysis. They're getting information through their relationships, through their on the ground research.
They are gathering insight in complex bankruptcy restructurings, often sitting in courtrooms or finding out information right outside of courtrooms. They put out 60,000 + articles every year, and that ranges from sort of real-time news and alerts to legal or covenant analysis, as well as research on private companies and credit. You put all that together, this is highly specialized, highly proprietary, not replicable. No LLM is going to be able to scoop that up anywhere. So very, very high quality, very highly proprietary. They also have contributed data. As you can see, I could go on for a little while on this, but we have done extensive due diligence. I would encourage any of our investors to do extensive due diligence on how proprietary this is. This is a great proprietary data set.
The other thing I should just say as a final point, in addition to what Octus brings, LSEG has great fixed income and credit capabilities, in parts of our business that people typically don't focus on. For example, LPC, which has great syndicated loan data. IFR, which has insight across mostly the public credit sector. Of course, our Yield Book analytics, which are industry standard fixed income analytics. So when you bring this together with what Octus provides, it really is an extraordinary and attractive combination.
Great. Thanks so much.
Your next question comes from the line of Hubert Lam at Bank of America. Your line is open.
Great. Thank you, and thanks for taking my question. For Octus, can you talk about how you get to EPS neutral in year one and accretion thereafter if you include interest costs of, say, like 5%-6% to finance it? So what are the assumptions to get there for revenue growth and EBITDA margin at Octus from 40% today? Are the cost savings at Octus coming through, and how quickly do you expect revenue synergies to come through? Thank you.
Yeah. So, I think to get to this point, you need to take the financial we gave during the presentation. So we think that Octus will land anywhere between GBP 300 million and GBP 320 million in 2026 in term of revenue. Then you take between low and mid, the double digit growth for 2027, and you can take the same for 2028. EBITDA margin, you have an accretion of about 300 basis points in 2027 and the same in 2028. So that's giving you a EBITDA margin for 2028. And as far as the debt is concerned, we think we'll be able to have a compound debt interest rate at 5.5%. And when you make this calculation, the 5.5% on $3.5 billion is about, if you take out the cash generated by Trumid, it's about $180 million of net financial expense for 2028.
And that's the actual profit that will be generated by Octus. I'm sorry. It's too many acquisitions at the same time. So by Octus, yeah. And that's where you'll get the use of 2028 EPS accretion.
Hubert, just to touch on, you asked about synergies as well. There will be modest cost saves here. This is not a cost synergy-driven transaction. We think there are very attractive revenue synergy opportunities over time, initially through cross-sell. LSEG is global platform, strong penetration across the large institutional and sell-side institutions. Octus tends to be more North America-focused and has great presence amongst the credit buy-side firms. We think there is really attractive cross-sell opportunities going both ways there. You should expect to see us building both enhanced and new product over time that will combine a bunch of the great assets I was just talking about before, as we create, build, and monetize new products. So, excited about that opportunity as well over time.
Great. Does that mean that you are saying near term, you expect revenue growth to be low to mid double digit with the revenue synergies, should we expect it to be higher than what it is near term?
No, I think as I was saying, Hubert, I think a good proxy is to take mid teen double digit for the next two year, all included.
Okay, great. Thank you.
Thank you.
We'll take our next question from Michael Sanderson at Barclays. Please go ahead.
Good morning, David. Good morning, MAP. Thank you for taking my question. On the other side of Esther, what he was asking about, just trying to understand a couple of the other elements of the Octus dynamics. First of all, the costs of bringing on and implementing the business. I'm sure we've all seen the legal cases about whether that's been spinning around at the moment around providing the structure to optimize these businesses when they come across. I guess just trying to understand how you bring that in and how much that will cost for us to think about. All added into that, the key individuals at Octus, what sort of terms, how are they kept within the business? Because from my understanding, they are seen as a very powerful and important part of the story over the last decade.
Thank you.
Yeah. I will answer your second question, and then MAP can try to touch on your first one. This is a talent accretive transaction for us. We are excited to have Kent and his team joining us. They have done a great job building this business. I think they are excited about the opportunity to partner with LSEG. I think together we can do a lot more in the space. We have put in place, in addition to what I will call the natural excitement and the alignment that we all have around serving our customers, a retention program. A significant part of the proceeds that the management team will be receiving will be reinvested in LSEG shares. There will be an appropriate incentive structure around that and around the performance. We are, as I said, really excited to have Kent and the team joining.
We think they will add a lot to both the Octus business within LSEG, but also to LSEG more broadly. Very excited about that.
Yeah. In term of cost to bring Octus in, maybe before answering to this question, a bit more color about what we are expecting to do. During the first two years, 2027, 2028 or the 24 months following closing, we will really have two major priorities. One is to make sure that Octus has the relevant means to carry on having this high level of growth that I was commenting to Hubert. At the same time, it ensure that it respects the level of cybersecurity protection that we have for the group, and so forth and so on. It is a relatively light integration, if you want, in this first two years.
At the same time we are doing this, we will spend money into building the combined product that David was referring to, which is going to take us about 24 months in order to get it on the market, 18- 24 months to get it on the market, 18 - 24 months after. In term of full cost to achieve, if you want, you can take a proxy of about GBP 100 million, that will be mostly in 2027, 2028, and maybe one 2029, but I think we will be done by end of 2028. Most of it being fundamentally engineering and the ability to bring all Octus data sets into LSEG structure. I just want to precise that I am talking cash here, so I am talking both OpEx and CapEx together.
Thank you very much.
Your next question is from the line of Russell Quelch of Morgan Stanley. Your line is open.
Yeah, good morning, both. I had a question on AI monetization and when adoption becomes incremental to revenue. You've given a bunch of new disclosures today signaling increased adoption and increased upselling of data. How should we be thinking about the conversion of this adoption into incremental revenue growth? I'm thinking as much detail as you can give around size and timing. I'm also wondering, how you approach pricing discussions, will this be more linked to increased data usage this year than in previous years, where perhaps inflation was more the input into the pricing discussion?
Thanks, Russell. On the first question around AI monetization, look, we've been pretty clear and consistent that we expect to see acceleration of the revenue growth in our Data & Analytics. We're comfortable with the 6.5% for this year. You've seen the growth rate tick up to 6.6% in Q3, and MAP's been very clear about what we expect in 2027 in terms of the 7%. Just to be clear, incremental growth from Octus will be on top of that. I'm not going to break out specifics around anything related to AI. We're seeing great adoption. We are seeing significant signs of monetization at this point, and I expect that to continue.
With respect to pricing, we are taking a slightly different approach to pricing going into next year. In the past, we've had pretty much a sort of single approach across most of our estate to pricing, across almost all the products and all the customers. Going into next year, we're taking a more targeted approach where we will have those who have bigger discounts, i.e., greater discounts to list pricing, their price rise will be a little bit more than those who are closer to list pricing. We expect that overall this should have an impact that is roughly in the same ballpark in terms of what we have seen this past year.
I don't think from your perspective you should expect meaningful or dramatic differences, but just wanted to be transparent that we are approaching pricing with a more nuanced or more multifaceted approach going into 2027.
Just to be clear on that, if you're closing some of the discounts you've given to customers previously and inflation is higher year-on-year, why should we not see a higher price increase next year than we did this year? I think you said they're the same.
I think, as I said, roughly in the same ballpark, but it's going to be a little bit different based on both product and different customer segments and what level of discount each customer has. As we've said in the past, we always look to get an appropriate return on our investment in new capabilities and in new product. We will be doing that during the course of this pricing cycle. At the same time, we are very focused on the long-term relationships that we have with our customers. On top of that, we have, as we talked about at H1, we will have in certain areas some usage-based or consumption-based pricing on top of that. So a few different moving parts going into 2027, which is a little bit different from years past. I wouldn't make a huge deal. You've mentioned inflation a couple of times.
A few years ago, we had significantly higher inflation. Do we take inflation into account when we're going into our pricing? Of course. That's a part of the overall formula.
Got it. Okay. Super. Thank you very much.
Thanks a lot.
Your next question is from the line of Michael Werner of UBS. Your line is open.
Thank you very much. Just wanted to get back to Octus, if that was all right. If you could help us understand, you've highlighted very strong revenue growth expectations and what we have seen in the past at Octus. I was just wondering if you could break that down a little bit, whether you can provide some indication as to what's coming from pricing, what's coming from upsell, what user growth has been, and whether that's been accelerating or decelerating in recent years, just to try to get a little better understanding as to the drivers behind that revenue growth. Thank you.
Thanks for the question. Let me answer the question this way. I can give a little bit of a breakdown in terms of the different parts of Octus's business, and the intelligence and expert analysis part, that's where the Octus business really started. That's sort of 60% + of the revenue. The data and analytics and the workflow platforms are smaller parts of the business. I would say very strong performance across the different parts of the business. There's no one area that I would say is the growth driver. One of the great things that Kent and his team have done has been really adding new product and also expanding product per customer, and they've been very successful in doing that. Because of that approach, they've seen consistent, steady growth among their existing products and then adding new products.
They also have the ability to take price on an annual basis. There's no one area at this point that I would point you to and say, "This is the growth driver and these other parts are slower." It's a very strong business, good growth across the business. Lots of innovation, and really both bringing in new logos, but also selling more products to existing logos.
Thank you.
Yep.
Your next question comes from the line of Arnaud Giblat of BNP Paribas. Please go ahead.
Good morning. My question's on treasury income. You saw a 21% increase year-on-year, and I think some of that came from a revised cash and margin framework. I'm just wondering how sustainable that increase is into the future. I assume that those changes remain and therefore a higher level should remain, but just wanted to get a bit of your take there. Secondly, can you talk a bit more about the potential sensitivity to rate rises or cuts given that new framework? Thank you.
Thanks, Arnaud. The change with respect to the treasury income, modest impact. As you have seen with the overall markets growth, a big part of that was driven by our OTC derivatives business, and a big part of that was driven by SwapClear. So very strong, very robust environment. The NTI is, as you know, paired very closely with that. I don't know off the top of my head, maybe a very marginal incremental impact from the change in our cash policy. But I wouldn't look at that 21% and say, "Oh, gosh, that's because of the change in the cash policy." That's because a very strong underlying performance in the business. Then with respect to interest rate sensitivity, as we've talked about in the past, this business is not driven by directionality.
This business, we see lots of activity when there is uncertainty about interest rates and volatility with respect to interest rates. Whether they're moving up or down is not really a driver of that. I think we'll see how the environment continues to play out, but this goes back to where I started the presentation. This is an all-weather business, and SwapClear and LCH are a great, I'll say, microcosm of that, in terms of their business. We see activity when there's lots of debt issuance. We see activity when there is hedging. We see activity when there's speculation. Whenever there's any uncertainty around inflation or rate moves, that's additive to the business. Does that help?
Great. Thank you.
Thanks, Arnaud.
You have a question from Enrico Bolzoni of JP Morgan. Please go ahead.
Good morning. Thank you for taking my questions and for the presentation. Apologies to go back on the monetization of AI, but I think it's just an important topic. You are very clear and explicit in saying that we should expect an acceleration coming from greater AI adoption and, I guess, agent consumption. If I look at your numbers, and I look at Data & Analytics, the pace of growth has been roughly stable over the last few quarters. I would think that this imply that this acceleration has not materialized yet. I just wanted to hear a bit more from you in terms of the price discovery that I suspect is going on in your conversation with clients. This is in a way a new technology, and I suspect you still have to define the contractual terms that will lead to revenues associated with agentic consumption.
I know you send your price letter typically in September. I would like to know where you stand in this conversation with clients. What sort of proposition you have? Is it going to be quite homogeneous, or you think that for anybody that will consume data through AI to MCP, there will be a similar pricing structure? Or again, it will differ greatly depending on the client. Any sort of information you can provide to help us understand how the monetization will occur, when it will occur, and where you stand at the moment in the conversation with clients on this? Yes.
Thanks, Enrico. Maybe MAP, do you want to kick off on this?
Yeah. Maybe I begin with the growth of D&A, and we've posted 5.2% in Q3 equal to Q2. That was 5.2% too. I just want to give here information that I've shared in several conference and one to one, but not on a call, which is that at the end of 2022, there was an RFP that was launched on a U.S. bank in its wealth division. After one year of RFP in November 2023, there was a decision taken, and we were the incumbent, and we lost this deal. I'm mentioning it because we are talking about GBP 25 million of revenue. I did not mention it in Q2 because it impacted D&A by GBP 2 million, so it's 20 basis points. But in Q3 it's GBP 5 million, so it's 50 basis points.
If you take aside this loss, which is quite exceptional, and we haven't been notified since that for another of this one. Remember, it was an old product. It's T1, so clearly not a Workspace. You should read the 5.2% as a 5.7%, and your Q2 5.2% at 5.4%. So it's kind of a 5.4% going to a 5.7%. I just want to say this, and the last thing I want to say that it's going to affect D&A until Q2 of next year, and then the GBP 25 million will be gone. But it's entirely taken into account in the guidance I'm giving you of the subscription business going from 6% in 2025, to 6.5% in 2026, to 7% in 2027. All right. Just to give some color on D&A.
And then to the latter part of your question, Enrico . We are having these conversations with pretty much all of our customers now. It is a wide range of discussions in terms of the monetization of the AI. In my remarks this morning, I referred to 50% in terms of the number of discussions going up 50% in our AI-related discussions. This comes from what we have put together, which we call an AI deal forum. Because with pretty much all of our customers, we are having these kinds of conversations. We are trying to make sure that it works for us and it works for our customers. There is a wide range of possible commercial agreements, and some of it relates to AI usage rights, some of it relates to AI usage rights plus consumption. Some of it relates to access to our MCP content.
If you go back to our first half results, we laid out the different channels or the different ways that our customers are interacting with our AI, all of that still applies. There is no one answer. I wish I could just tell you expect X percent uptick across all of our customer relationships for AI usage, but it is just not that simple. We are seeing with some of our smaller customers, more flexibility and nimbleness and eagerness just to strike a deal and then move on with the usage. With some of our larger customers, they are still trying to figure it out and trying to understand what it looks like embedded across their systems, how we can help them get it embedded across their systems. What does that look like in the context of our broader LDA agreements?
What is the appropriate surcharge, et cetera? So ongoing discussions, some of them resolved and monetized. We just did not think it was constructive to break it out at this point and then be breaking it out every quarter. But we are very comfortable, as you have heard from MAP, with the acceleration in our growth, going through the end of this year and through 2027.
Sorry, may I go back to what MAP said just to ensure I understood it correctly? So, we should expect from Q3 next year in basically constant currency revenue growth because of the exit from the window of that contract you lost. Is that the right reading into that?
Yeah. That's correct. Q3 2027. That contract is affecting D&A revenue. I repeat, GBP 2 million in Q2, GBP 5 million in Q3, will be GBP 7 million in Q4, 7% in Q1, and finally 5% in Q2 of next year. You have the full thing. Then it will be included into the guidance of subscription revenue of 7% in 2027, and obviously in the 6.5% of this year. Is that clear?
I think we'll take that as a yes.
Yeah. Okay.
Maybe I'll just, to pick up on MAP's comment. When we talk about ongoing discussions with our customers, it just is worth reminding everyone how this industry often works and the pace at which some of our customers sometimes move. MAP is talking about impact from an RFP process that kicked off in 2022 and was decided in 2023, and we're seeing it in 2026 and 2027. We see the opportunity, we see the benefits, we see the engagement, and the usage with AI, and we're starting to see the monetization, but we also see this as a steady trajectory over coming years as opposed to any kind of spike up. Again, because of the nature of the financial sector, how our customers often go through their processes, their risk committees, their new product committees. We know how that works. We work closely with them.
We're deeply embedded. We're viewed as a trusted partner. We will benefit and are benefiting from this, but I just wanted to take this opportunity to remind everyone of the timeframe and how this industry works.
Your next question is from the line of Oliver Carruthers of Goldman Sachs. Please go ahead.
Hi there. Morning, Oliver Carruthers from Goldman Sachs. David, I wanted to pick up on your comment on the OTC derivatives line and your LCH business. It may be a little bit of a difficult question to answer, but I'm just trying to get a sense of the structural versus the cyclical growth here, because I think this is a business that the market thought was kind of overgrowing last year in 2025, and the reality is, it's accelerated again in 2026. If we think about some of the moving parts of this, as you say, SwapClear, very strong volumes, but you're also kind of calling out onboarding of new members at CDSClear. You've got the Post Trade Solutions business kind of scaling. I think consensus has this business kind of meaningfully decelerating over the next couple of years.
I just wanted to get your sense on what is cyclical, what is structural at the moment, and how are you thinking about the kind of medium term growth rate from these levels. Thank you.
Oliver, thanks for the question. Since I joined LSEG in 2018, I have gotten questions about the fact that SwapClear must be mature and its best days of growth must be behind it. Clearly, that has not been the case. Look, the way we manage this business and the way we invest in this business, and the way that Dan Maguire and the team approach this business is to have a very strong, attractive growth rate regardless of market activity. Then when there is market activity, market volatility, that is upside. You can make your own predictions, and this is a little bit of the issue out there. You all can make your own predictions on what market volatility looks like.
The way that we think about it, just to be completely transparent, we have sort of a base case of a couple incidents of market volatility per year. That can obviously be. You can have more volatility than that in a year. You can have less volatility than that in a year. But if you look back over several years as we do, that tends to be a pretty good baseline. So that is how we think about sort of volatility and market activity. On top of that, I should just mention, this business over the last several years has become much more than SwapClear. We are seeing excellent growth in ForexClear. We are seeing excellent growth in CDSClear, the Post Trade Solutions also growing double digits.
What you have is a really diversified offering across multiple asset classes, that is growing very well through lots of different kinds of market conditions. Again, hopefully that helps. Again, I am going to put it back on you. You can make your own judgment on what kind of market volatility you expect in 2027 and beyond.
But yeah, just to complement a couple of numbers. SwapClear this quarter, 18%, CDSClear 15% growth. Post Trade Solutions that David was mentioning, 14%, because you know that in OTC derivative you have closed the cleared and the uncleared. Beyond the volatility, we are extending to new products. PTS is a good example. We just introduced TradeAgent, so we just signed four new members. Revenue is flowing in Q3. We have expanded geographical reach of existing products, and we are accessing new asset classes. There is a volatility, yes, but there is everything we do on top of that, and that materializing to this record performance.
Thanks for the color and the numbers. That's helpful. Thank you.
Sure.
Your next question is from the line of Ben Bathurst of RBC. Your line is open.
Morning, all. I'd like to return to the cross-sell opportunity by Octus. You mentioned in the initial remarks that you've got 750 customers. I just wondered, could you share the overlap with existing LSEG customer base, even if just a rough percentage? And in terms of the growth assumptions you've outlined in EPS accretion, MAP, thanks for the detail there. Does that factor in any cross-sell benefit out to 2028, or is that upside, do you think? Thank you.
Let me touch on the first part of your question and then MAP can answer the second part. In terms of overlap of customers, there's some. Octus has some customers in, for example, the big sell side institutions, and has some customers, of course, outside of the U.S. But it's frankly not a huge amount. And where there is overlap, there's still significant opportunity. So, for example, we may be providing lots of other product to customers, and of course, what Octus and Kent and the team are providing there is completely different outside of our core data set or areas of expertise. I would say the area where there's really very little overlap is in the buy side credit firms. That is not a space where LSEG is well penetrated. So we think that's a great area of potential cross-sell.
The ability to add on the services that we provide in fixed income, in credit, whether that is LPC, Yield Book, IFR, or whether it is the broader offerings that we have, where we can be serving those customers very effectively compared to what they currently have outside of their Octus offering. I would say what is also exciting about this is once we get past the cross-selling, there is a lot that we can do with respect to new products and integrating Octus into our broader data offerings, evaluated pricing, Yield Book, model enhancements, private credit valuations. The list goes on. We are excited about that as well. Do you want to add?
Yeah, no, I just want to go back to the numbers I was mentioning at the beginning. I think for 2027, 2028, the proxy. The Octus market grow somewhere in between 10% and 15%. In order to calculate the accretion and to have an idea of the cost center, and so on, I have taken 2x 13% for 2027 and 2028. Now, call me conservative. I went into the middle. I think there are some benefits of the cross-sell in there. Can we have more? Certainly. But that was the assumption I took in order to commit to the market to a 2028 neutral, and then after accretion in 2029.
Again, at the same time we are doing this in 2027 and 2028, at the same time, our engineering function with Irfan and the team is working with Octus engineering function in order to integrate the data set of Octus into LSEG, and create new product. This part will be commercialized mostly in 2029 and 2030. Hope it is giving more clarity.
That is us. Thank you very much.
Thanks, Ben.
You have a question from Iulian Dobrovolschi of ABN AMRO. Your line is open.
Hi, good morning, gentlemen. I have a question on the margin expansion, please. With the upgraded guidance that you have on EBITDA, for 2026 between 100 basis points and 120 basis points, that basically goes down to about 50 basis points per year between 2027 and 2029, given the 150 margin expansion guide that you put out today. Whilst at the same time, you expect the subscription business to actually accelerate over the medium term for AI and some other means. Just wondering why the incremental margin falls that much after 2026 and what drives that. Also, if you look at some of your peers, this actually flagged exceptional cost savings through AI, potentially making room for operating leverage in the business. So wondering what sort of cost savings do you see coming through already from adopting AI internally.
Yeah. Just to go back on the margin, to set the scene. We began the year with 80 basis points- 100 basis points. In H1, we increased it to 100 basis points and now to 120 basis points. So behind this is the quality of our operating leverage and the quality of the cost control we have in place in two of our major costs, which is fundamentally resource costs, so internal employee and contractors, and cloud costs. In this cloud cost, now you have in this cloud cost or next to it, if you want, you have the AI tokens cost. We make 120 basis points of margin with a record growth. Clearly, there is a part of the 120 basis points, which is coming from the revenue, which is with the quality of the control that I just explained, going down into the bottom line.
Finally, you need to look at this 120 basis points. You know, it's 30 basis points, which is the final impact of the SwapClear deal we've made. Remember, we had 100 basis points in 2025, but that's 30 basis points in 2026 and 90 basis points, which is underlying. You look at this 90 basis points, 90 basis points going to 50 basis points, that's what you say. Maybe in the 90 basis points, you've got about 20, which is coming from the elevated growth at 8.6% year-to-date.
Can we do better than the 150 basis points in three years? We'll see. Maybe. But the assumption I have taken is an acceleration of the AI solution that we have put in place in the group. The overall assumption, I want to be clear on that. The overall assumption, I think I've said it already, but I just want to repeat it. Behind this 150 basis points of margin accretion is pretty much a stable headcount, meaning that we are able, with the resources we have, to sustain 7%-8% growth in the years to come. In a nutshell, the dynamics of this.
Understood. Thanks.
This concludes our Q&A session. I will now hand the presentation back to David Schwimmer, CEO of London Stock Exchange Group.
Thank you very much. Thanks to everyone for your questions. We know a lot going on in today's call and a lot going on in our Q3. Just a quick recap. Really strong performance, raising the guidance across revenue, margin, and cash. Very strong adoption and monetization around our AI capabilities. Octus and Trumid this week. Really excited about what each of those does for LSEG and Tradeweb, as well as what we can do in the fixed income space more broadly. Then, of course, continuing with our current buyback and having capacity to do another GBP 1 billion -GBP 1.5 billion next year. Again, thanks for joining. If you have any further questions, you know where Peregrine and the team are.