ICE’s Recurring Growth Holds Up as MarketAxess Raises the Stakes

Results vs. Consensus

The quarter ended June 30 tests the case we established in May: ICE should retain customers and grow subscriptions even when exceptional exchange turnover subsides. Q2 passes the retention test more convincingly than the earnings-growth test. Recurring revenue reached $1.353 billion, but adjusted EPS fell from $2.35 in Q1 to $1.90 as energy activity cooled. Higher annual spending and the proposed MarketAxess acquisition now require more capital discipline to translate franchise strength into shareholder returns.

Year-over-year comparison

Sequential comparison

Revenue assessment: Recurring businesses supplied 78.9% of year-over-year net revenue growth, a substantial improvement in the composition of growth from Q1. Transaction revenue added only $26 million as lower energy revenue offset stronger rates and other products. Even after removing the $11 million of disclosed temporary benefits, net revenue would have risen approximately 4.4%; the growth is real, but its quarterly pace is far below the first-quarter surge.

Margin assessment: Adjusted operating expenses rose 5.6%, slightly faster than revenue, leaving margin at 61.1% versus 61.3% a year earlier. The sharper sequential change was mix: revenue fell $311 million while adjusted costs increased $3 million, reducing adjusted profit by $314 million. Q1’s 65.2% margin was never a durable group-wide run rate. The annual cost increase makes growth in subscriptions more valuable, but also means incremental demand must arrive before investment earns an adequate return.

EPS assessment: Adjusted net income grew 3.0%, with fewer diluted shares lifting EPS growth to 5.0%. GAAP EPS grew 14.2%, helped by $63 million of investment fair-value gains compared with $2 million a year earlier. Acquired-intangible amortization of $237 million remains the largest adjustment; the $958 million GAAP profit reconciles to $1.074 billion adjusted after integration costs, the regulatory credit, investee income, investment marks and taxes. Neither the larger GAAP growth rate nor the temporary revenue items justify annualizing this quarter’s improvement.

First-half cash generation

Q2 operating cash flow was $1.998 billion and adjusted free cash flow was $1.450 billion, derived by subtracting Q1 from the six-month totals. The $286 million adjustment for net Section 31 fees removes cash collected for regulatory remittance; that cash is not available for permanent shareholder returns. First-half adjusted free cash flow of $2.600 billion exceeded adjusted net income of $2.412 billion and funded $1.793 billion of dividends and repurchases, leaving $807 million before other investing and financing uses.

Segment Performance

Growth and margins in this table are calculated from the reported dollar amounts; revenue is net of transaction-based expenses.

Exchanges: the network retains demand, while turnover determines profit

Energy revenue fell $77 million year over year and $296 million from Q1, yet total exchange revenue still grew. Financials, cash equities, agriculture and recurring fees supplied the offset. Adjusted exchange profit fell $325 million sequentially, more than the group’s decline, showing how concentrated the operating leverage remains in this franchise.

The prior thesis distinguished customer retention from transaction frequency. Total futures and options open interest rose 20% year over year, while exchange recurring revenue rose 10% to $416 million. Those measures support an intact network. They do not reverse the current energy decline: the same customer can hold a hedge for months while generating fewer new transaction fees.

Assessment: We retain the benchmark pillar at ON TRACK but reduce annual exchange net revenue to $6.30 billion from $6.42 billion. That requires $1.528 billion per quarter in H2, 4.3% above Q2 and well below Q1’s $1.781 billion. Stronger recurring fees and rates activity can support that modest recovery without another energy shock.

Fixed Income and Data Services: the prior growth test improves

Recurring revenue reached $531 million from $483 million, while transaction revenue was flat at $114 million. The $48 million segment revenue increase therefore came entirely from recurring businesses. Adjusted operating profit rose $34 million to $295 million, demonstrating useful leverage even as ICE invests in data infrastructure.

Management identified approximately $8 million of one-time items in the segment. Removing that benefit from total FIDS revenue gives about $637 million, up 6.7%, rather than the reported 8.0%. The remaining growth and higher annual recurring guidance strengthen the data pillar; the temporary contribution does not need to recur for the segment to compound.

Assessment: We raise our FY2026 FIDS revenue estimate to $2.62 billion from $2.60 billion. The resulting $659 million quarterly H2 average assumes continued data demand, but no early revenue from Hall 6 and no MarketAxess contribution. That preserves the capacity constraint built into the prior forecast.

Mortgage Technology: a small but meaningful recurring improvement

Recurring revenue of $406 million exceeded the approximately $401 million expectation from the last call. After removing $3 million of one-time items, the $403 million base is 2.0% above the year-ago $395 million and $6 million above Q1 excluding its $4 million temporary benefit. That is better evidence of new implementations and normalized renewals than last quarter’s flat underlying comparison.

Transaction revenue rose 11% to $151 million, with closed-loan monetization and closing activity still growing faster than subscriptions. GAAP operating income improved to $45 million from $11 million a year ago and a $13 million Q1 loss. Of the $34 million year-over-year GAAP improvement, $16 million came from lower acquired-intangible amortization; adjusted profit improved $18 million. The recovery has an operating component, but accounting relief also helps the headline.

Assessment: We raise annual mortgage revenue to $2.23 billion from $2.18 billion while retaining AT RISK. Better underlying recurring revenue and a 42.9% adjusted margin are encouraging, but one quarter of modest subscription growth is insufficient to establish a balanced recovery from the acquisition cycle.

Key KPIs

Key Topics & Management Commentary

Overall Management Tone: Management remained confident in the existing franchises, but the call shifted from defending durability to defending a major acquisition. Repeated questions about MarketAxess’s share and pricing elicited a plausible integration strategy, with fewer measurable commitments on revenue than on costs and leverage.

1. MarketAxess makes capital allocation the central new test

The acquisition would connect ICE’s retail and wealth bond franchise with approximately 2,100 institutional firms. A broader network can improve the probability of matching an infrequently traded bond, while institutional distribution expands the reach of ICE’s evaluated pricing and analytics. This is a tangible extension of the existing data pillar, but it also exposes ICE to the target’s competition, fee pressure and implementation burden.

“We expect to achieve approximately $100 million of annualized expense synergies with 1/3 realized in year 1, 2/3 by year 2 and the full run rate achieved by year 3. These savings will be driven by the consolidation of corporate functions, real estate rationalization, vendor and technology overlap and more efficient use of shared infrastructure across the combined platform.”
— Warren Gardiner, CFO

The purchase price is $167 per MarketAxess share, approximately $6.0 billion of equity value and $5.7 billion of enterprise value, financed through bonds, a term loan and commercial paper. The quoted 10.6-times acquisition multiple includes the full synergy run rate. Dividing $5.7 billion by 10.6 implies about $538 million of EBITDA including savings, or approximately $438 million before the $100 million target: roughly 13.0 times before synergies. The lower advertised multiple depends on execution over three years.

Assessment: The standing capital-risk pillar moves from EMERGING to MATERIALIZING because a large debt-funded commitment now exists, although the acquisition remains pending. We assign no deal premium to fair value. Expected first-full-year adjusted EPS accretion is useful, but is a weaker test than earning an attractive return after financing, integration and competitive reinvestment.

2. Stronger data guidance still carries a capacity timetable

Exchange recurring growth guidance rises from mid-single digits to high-single digits, and FIDS recurring guidance rises to 7%–8%. The raised outlook increases confidence in demand, while delivery through the harder second-half comparisons remains the next test. It does not eliminate the timing constraint: the full-year range includes stronger early growth, temporary items and index-linked assets that can fluctuate with markets.

“As a reminder, second-half growth will likely trend towards the lower end of that 7% to 8% range, driven largely by timing, and specifically the comparison period in the third quarter and fourth quarter of 2025, which benefited from the sell-through of Hall 5 within our MAWA data center. It is worth noting that we have already begun selling Hall 6 and anticipate that revenue will begin to be recognized in early '27, with several additional halls providing further capacity behind it.”
— Warren Gardiner, CFO

Assessment: The raised guide warrants a higher data estimate, with H2 growth closer to the lower end of management’s range. Hall 6 selling ahead of recognition supports the demand case for 2027, while accelerating construction and hardware spending consumes cash in 2026. That timing gap argues against raising the valuation multiple today.

3. Energy options deepen the network without fixing quarterly turnover

Energy’s weaker revenue does not contradict the longer-term hedging opportunity. Options now represent 40% of energy open interest, while more than half of energy positions sit beyond six months and only about 12% in the front month. Those are different exposures from short-lived speculative turnover.

“This options growth matters because it is another sign of how deeply customers rely on us. Options are how they manage complex, longer-dated risk. And once that positioning is on our books, it tends to stay. We have studied the durability of options positions versus futures, and the result was clear that options positions tend to be held for a longer term, often are held to expiry and many clients hedge their delta risk with futures, providing a net benefit to the underlying futures market at the same time.”
— Benjamin Jackson, President

ICE also described risk shifting from physically settled Murban toward its cash-settled Dubai contract as regional delivery became less certain. Its range of related benchmarks can retain business when customers change the instrument they use. The claim that this shift is permanent remains management’s expectation, rather than a requirement in our forecast.

Assessment: Longer-dated options and related futures hedges reinforce the benchmark network. We give that mechanism weight in the normalized exchange forecast, while Q2’s energy decline keeps the turnover-cost risk EMERGING. Positions on the books support future activity; they do not provide a quarterly revenue floor.

4. AI data products move from access trials toward commercial workflows

Last quarter’s MCP initiative tested nonproprietary data under existing licenses. The expanded offering now includes proprietary data with permissions and an audit trail. ICE Compass adds pre-trade estimates of counterparty prices and rankings, with T. Rowe Price as anchor client. The economic opportunity is to improve execution decisions using data ICE already controls, then earn more distribution and analytics revenue.

“Our first MCP release opened a new channel for expanding distribution of our nonproprietary data. Our newly expanded ICE MCP offering now offers ICE's proprietary data into our clients' AI workflows. And we didn't simply build an open data pipe. We built a client engagement channel that runs both ways. The MCP server connection is the easy part. What matters is what sits behind it, organized data that arrives with its own meanings attached and which represents and respects our proprietary rights so that each client model are not left to guess what permissions govern who can see what. We now offer a complete audit trail so that every output can be traced and trusted.”
— Jeffrey Sprecher, Chair and CEO

Assessment: Proprietary content and a concrete workflow product strengthen the defense against data commoditization. They support the existing recurring-growth assumption, while the absence of disclosed incremental contract economics keeps a separate AI valuation premium unwarranted. MarketAxess could improve distribution if the deal closes; that opportunity is not already revenue.

5. Mortgage automation must improve both adoption and monetization

Aurora is embedded in Encompass and MSP, using governed data and role-based permissions. Servicing API activity rose 39%, and agents are handling borrower inquiries and back-office workflows. This advances last quarter’s product-efficiency case; the more important new financial evidence is the small increase in recurring revenue after temporary benefits.

“ICE Aurora embeds this agentic AI directly in Encompass and MSP with governance, audit logs and human approvals built in. AI assists the human in high-risk decisions such as underwriting, pricing and cash movement, escrow and remittance and doesn't autonomously make a call.”
— Benjamin Jackson, President

Assessment: Automation can defend the system of record and create an additional chargeable service, but cost savings accrue first to the lender or servicer. Our modest mortgage estimate increase reflects implementations and early monetization, not a proportional translation of API growth into revenue. Repeated underlying subscription growth remains the condition for upgrading the pillar.

6. Repurchases and accelerated investment compete for the same cash

ICE repurchased $651 million in Q2 and raised baseline repurchases from $350 million to $400 million per quarter. Meanwhile, full-year capital expenditure and software investment rises to approximately $850 million, $85 million above the prior midpoint. Management is bringing some planned 2027 spending forward, rather than identifying an immediate increase in 2026 revenue of the same size.

“Gross leverage is expected to peak temporarily around 3.4x pro forma EBITDA, and we are targeting return to 3x or below within 18 to 24 months, fully consistent with the pace of deleveraging we have demonstrated following prior debt finance transactions. Our commitment to maintaining a strong investment-grade credit rating is unchanged. On capital return, alongside our deleveraging program, we expect to increase baseline share repurchases from $350 million to $400 million per quarter. Our Board has recently authorized up to $4 billion of share repurchases, and we intend to deploy that capital in a manner that is disciplined, opportunistic and consistent with our obligations to creditors and our investment-grade rating.”
— Warren Gardiner, CFO

Unrestricted cash was $1.067 billion and debt $19.846 billion at June 30. The much larger clearing margin deposits are matched client obligations, not a corporate acquisition fund. The proposed return to leverage of 3.0 times or less within 18–24 months must therefore be funded by operating cash generation and disciplined uses of it.

Assessment: Current cash conversion makes continued baseline buybacks credible, but the combination of higher repurchases, construction and acquisition debt reduces flexibility if trading disappoints. Our share-count estimate falls only modestly to 566 million for the full year; we do not extrapolate the extra $300 million of opportunistic Q2 purchases.

7. New distribution can preserve the franchise while changing its economics

Tokenization remains an open issue from the prior thesis, while perpetual futures adds a current question about distribution and speculative flow. ICE is treating perpetual contracts as another distribution channel for licensed benchmarks and tokenization as an evolution in settlement. Economic-indicator and GPU-compute futures extend its product set, while private-credit data could eventually share the institutional distribution network proposed in the acquisition.

“The fixed income network that we're designing will not stop at public credit. We have a plan to use the same rails to connect private credit clients who we're going to bring in via our initiative with Apollo. So public and private credit will increasingly be accessible on one platform.”
— Jeffrey Sprecher, Chair and CEO

Assessment: A common data and distribution system is more credible than a collection of unrelated ventures, but regulatory approval, customer adoption and retained fee economics still determine the return. We preserve the earlier commitment to track clearing, private credit and tokenization without adding standalone project value. The unresolved collateral-income question remains especially relevant as settlement accelerates.

Guidance & Outlook

Q2 adjusted costs of $1.038 billion met the prior $1.030 billion–$1.040 billion commitment. The annual increase reflects compensation, data-center capacity and product development, while Q3’s midpoint is $30 million above Q2. At our tax and share assumptions, the $40 million annual cost increase absorbs approximately $0.05 of EPS. A higher recurring-revenue outlook partly offsets that burden, but does not warrant a blanket earnings upgrade.

Implied H2 path: Our $11.15 billion net-revenue estimate requires $5.507 billion in the second half, or $2.754 billion per quarter, 3.3% above Q2. With $2.137 billion of remaining adjusted operating expenses, the H2 cost run rate is close to Q3 guidance. Approximately $3.85 of remaining EPS, or $1.93 per quarter, is needed for our rounded $8.10 annual forecast. The forecast assumes modest sequential recovery, not a return to Q1’s peak.

Expectations and guidance style: The $1.90 result exceeded the $1.84–$1.88 quarterly consensus range. Management gives component guidance rather than a consolidated EPS target; our earnings forecast therefore depends on transaction activity as well as the new recurring ranges. The explicit H2 capacity comparison and ongoing temporary revenue benefits make the stronger recurring guide credible without assuming habitual conservatism. The prior low-to-mid-single-digit mortgage growth outlook remains context for our 6.1% annual estimate, which now assumes a somewhat stronger outcome.

Analyst Q&A Highlights

Can the acquisition repair market share and fee pressure?

The first challenge asked why ICE could improve a business facing years of share and fee pressure. Management’s answer relied on combining complementary liquidity, data and infrastructure to lower customer friction.

Q: “So I wanted to start with the acquisition. Curious why you are the best owner of this business? And what gives you confidence that you can improve what has been a declining market share and fee per million trends for MarketAxess for the last several years?”
— Daniel Fannon, Jefferies

A: “What I would say is if you look at what's causing the pressure in the market across the entire segment is increasing number of friction points along the way. And as we said in the prepared remarks, this gives us an opportunity to consolidate some of those friction points and create greater economies of scale, which we believe will generate more opportunities to capture greater share over time as those economies of scale are realized and the operational cost and efficiency at the client side become a better shot for them to take advantage of.”
— Christopher Edmonds, President, Fixed Income and Data Services

Assessment: The mechanism is plausible: clients may route more business through a platform that reduces their total operating burden. The answer does not establish a recovered fee rate or measured share benefit. Our capital-risk assessment therefore worsens because ICE is underwriting the repair, not because the strategic rationale lacks merit.

What is actually assumed in the accretion claim?

The question separated the initial accretion calculation from later revenue acceleration and asked how much of the target’s existing platform would be retained.

Q: “I want to come back to MarketAxess. Just more on the deal accretion assumptions in the first year and then the plan over time. I think you said, Warren, you're assuming mid-single-digit growth. Just wanted to confirm, are you looking at consensus expectations for revenue and expenses in your deal accretion analysis? Or do you guys have your own model? And then over the longer term in terms of accelerating that, is that more on just the expense synergy side? Or are you contemplating material revenue synergies to do that? And as you integrate the firms, as you plan to integrate firms, is it -- is the plan to mostly retain what MarketAxess has built and some of the senior management team? Or do you plan on thinking about rearchitecting some of what they've built to try to tackle the market share issue a little bit more aggressively?”
— Brian Bedell, Deutsche Bank

A: “So we use consensus EPS as the base for that calculation. So you can think about it that way. We -- for the synergy side in terms of accelerating them or sort of the accelerated growth that I spoke to, that was in reference really to the top line. I think right now, what I was trying to say there was that in terms of what we paid for MarketAxess, the value that we underwrote, we assumed mid-single-digit growth, which is where they've kind of been a little bit recently. But the target here will be to accelerate that revenue. It may take a little bit of time on that front, but the target here will be to accelerate that revenue for -- and we outlined many of the reasons why we think we can do that, but that's the way to think about that. So those will be sort of the revenue synergies you want to come. They're tough to quantify, obviously, given a transaction business. But that's the opportunity, I think, for us to come in and really reinforce the plan that MarketAxess has laid out to you guys. I think that they do have a solid plan in terms of getting to what they've talked to do in terms of the high single digits. But I think a lot of the assets we bring to them will just really help reinforce that and help that growth profile.”
— Warren Gardiner, CFO

Assessment: Management anchored accretion to consensus EPS and mid-single-digit revenue growth, while treating faster growth as an opportunity. That is more disciplined than requiring an immediate growth turnaround to make the deal accretive. The response left the integration architecture and revenue-synergy magnitude open, so the expense ramp is the firmer commitment for investors to track.

Will mortgage AI produce revenue as well as customer savings?

The question tested whether Aurora could increase revenue, gain share and reduce ICE’s own costs. Management confirmed early monetization, then explained why customer value and human supervision still shape pricing and implementation.

Q: “We wanted to see if you could go a little deeper into how ICE Aurora is embedding Agentic AI in both Encompass with originations and MSP with servicing. So how does agentic AI improve your ability to grow revenues and take share longer term? And is there a benefit on the cost side, too, as ICE Mortgage Tech can potentially run more efficiently with less people?”
— Craig Siegenthaler, Bank of America

A: “And to your revenue question, as we look to drive more and more efficiencies for our clients and how they use our platforms, we will look for areas where we can monetize that. We have started to, as clients have started to engage with some of the AI tools that we have embedded into both Encompass as well as MSP, we are starting to monetize those. And as clients are engaging with them, onboarding them, we're going to crystallize more and more just what is the actual value that's being driven for the end client, and that will inform going forward how much we can charge for them. But throughout Encompass, we're automating things like fee or automated service ordering, fee calculations, generating disclosures, engagement with settlement providers. But the magic is knowing when does human need to be in the loop. And when is there a potential for errors or hallucination in the model where a human needs to be in the loop. And then on MSP, I talked about customer service things that we're automating in a number of different calls, but we've also been automating back-office workflows such as escrow, investor reconciliation, HELOC processes, et cetera. So we're very confident going forward on our position here and being able to drive efficiency for our clients.”
— Benjamin Jackson, President

Assessment: This is progress from a product-efficiency story to initial commercial adoption. It does not quantify a material reduction in ICE staffing or an AI revenue run rate. We give early monetization some weight in the mortgage forecast, with recurring revenue and segment profit providing the observable tests of success.

Are perpetual futures a substitute for the hedging franchise?

Management was asked about the growth opportunity from perpetual contracts after licensing oil benchmarks to OKX. Its central distinction concerned the customer’s purpose and the absence of a forward curve.

Q: “A lot of good ones are on the deal so far. So maybe I'll pivot. Jeff, there's been a lot of attention on perpetual futures recently. You've seemingly been much more open to the idea of perps relative to your largest competitor in the U.S., you, licensed Brent and WTI to OKX during the quarter. So I would love to get your high-level view on perps as an asset class, the CFTC's push to bring them onshore and how meaningful of a growth opportunity do you think that it could be for ICE on both the retail and institutional side going forward?”
— Patrick Moley, Piper Sandler

A: “So first of all, it's a bit of a misnomer in my mind that they're called perpetual futures. The reality is we're looking at these as if they're really a competitor to leveraged ETFs. And as we've mentioned in the prepared remarks and as Warren talked about, we continue to license our data to those ETFs, and we see an opportunity with perpetual futures to continue that because we honestly think they're a very similar product with a different distribution vehicle. In other words, at least in the U.S., ETFs are distributed through FINRA broker-dealers and perps tend to be distributed through crypto blockchain-oriented companies and particularly widely distributed outside the U.S. The reason we don't -- we think it's kind of a misnomer that they're called futures is because they don't produce a forward pricing curve. And so there are very little use for hedgers. So they tend to be a match of a speculator to a speculator, which tends to mean somebody wins and somebody loses. So the long-term success of a speculator to speculator market has to be that people are either enjoying it for entertainment purposes or something else other than our traditional markets where we really lean into commercial hedging.”
— Jeffrey Sprecher, Chair and CEO

Management also described uncertainty around the treatment of real-world assets and securities, with industry concerns about continuous small-size trading affecting traditional price discovery. That regulatory uncertainty limits the near-term opportunity even if benchmark licensing finds more customers.

Assessment: The leveraged-ETF analogy and lack of forward pricing explain why perpetual contracts need not replace commercial hedges. They can still compete for speculative flow that contributes liquidity. Licensing offers participation in the new channel, while the 20% open-interest increase supports the current franchise more directly than predictions about regulation.

Does faster collateral movement improve the whole profit pool?

The question pressed on defensible revenue, collateral efficiency and the interest ICE currently earns on collateral. Management separated matching from settlement, emphasizing that faster transfers introduce difficult title and insolvency questions.

Q: “So you've spoken about tokenization, Jeff, as an evolution of the market infrastructure rather than a replacement of today's exchanges. So as more securities move on chain, just curious over time, how you see industry profit pools migrating? What's most defensible? What areas might need to be defended more? Where might there be scope for new revenue opportunities for the industry, but also for ICE? And then if collateral can just move instantly on chain, how much incremental trading activity or capital efficiency do you think that unlocks? And where might there be givebacks around that? Maybe you can remind us how much do you generate today from collecting interest on collateral?”
— Michael Cyprys, Morgan Stanley

A: “But in terms of settlement and the way collateral can move, we've been limited -- the industry has been limited by U.S. banking hours, really where the main security markets are. And as we go follow the sun around the world, we're having to figure out how we move collateral to these various banking jurisdictions. And that's what on-chain collateral movement can do. It's a bit scary to regulators and to market participants. Retail has embraced it, obviously, as you've seen. But for our traditional infrastructure, we have to deal with what happens if there's a financial crisis, what happens if there's a bankruptcy? What happens if a Silicon Valley Bank collapses and what is in flight and who has title to it and who -- what regulator can raise the walls to keep that collateral in the ecosystem that against the trades that are in the same ecosystem. And so those are yet to be worked out. But as we've been doing, we've been working very closely with the Securities and Exchange Commission in the U.S. to try to move the New York Stock Exchange listed securities on chain. Obviously, others are doing somewhat look-alike securities around the world, and we think that there's obviously a market for the true securities. And I also think that it will open once securities can be on chain, not only will collateral movement be easier, but I think for those that are buyers and holders of securities or on-chain assets, they'll be able to be pledged and lended in ways that the crypto community is already doing with stablecoins and other things that will give better underpinnings to people that loan money. And therefore, I think, unlock more of the economy because there'll be more certainty in the ability to provide capital. So we're working on it. We're trying to do it within the regulated businesses that we run. We have -- we've mentioned a number of major institutions that we have existing agreements with that are all working together to try to solve some of the institutional problems that I just mentioned. But there's real work going on. And so I think later this year and early next year, you'll start to see some significant entities moving on chain.”
— Jeffrey Sprecher, Chair and CEO

Assessment: The answer preserves the prior opportunity in more efficient settlement while adding concrete legal and operational constraints. It did not quantify collateral interest or demonstrate that additional transactions would offset any lost income. The expectation of significant entities moving on chain later this year and early next year is a new implementation signpost, not an earnings contribution in our valuation.

What They’re NOT Saying

1. How much MarketAxess revenue improves after integration. Cost savings have a three-year schedule; share gains, fee stabilization and revenue synergies do not. Lower client costs could increase volume while still limiting revenue per trade.
2. A complete mortgage recurring-revenue bridge. Implementations, renewals and temporary items explain the direction, but pricing, churn and new AI fees are not separately measured. Underlying growth must repeat before the recovery earns greater valuation weight.
3. Whether new data usage raises revenue per customer. Proprietary MCP access and Compass increase relevance; contract values and paid utilization would show how much of that relevance becomes incremental profit.
4. The net effect of tokenization on collateral income. More efficient collateral transfer can help clients while reducing parts of the incumbent revenue pool. The unanswered numerical question prevents a confident net-benefit forecast.

Market Reaction

• Pre-print setup: ICE closed July 29 at $154.28, down 4.7% year to date and 16.9% over twelve months, after a 25.3% rise over thirty days. The S&P 500 was up 6.9% year to date.
• July 30 reaction session: The stock opened at $155.03, traded from $151.06 to $158.55, and closed at $156.27, up 1.3%.
• Participation and benchmark: Volume was 8.0 million shares versus a 4.7 million thirty-day average, or 1.7 times normal. The S&P 500 gained 1.7%.

ICE’s positive close followed both an earnings beat and the acquisition announcement, so the session cannot isolate the value investors assigned to either. The stock lagged the index by approximately 0.4 percentage points and had already rebounded sharply during July. The response is consistent with approval of the recurring franchises tempered by spending and deal risk, rather than an unambiguous rerating of the business.

Contemporaneous coverage emphasized the strategic fit in fixed income and the prospect of moving attention away from perpetual-futures, mortgage-rate and AI concerns. In our view, the acquisition changes the debate more quickly than it changes earnings: investors now have a concrete integration and leverage schedule to judge. The existing business still supplies the near-term valuation support.

Street Perspective

Debate: MarketAxess platform logic versus competitive economics

Bull view: Institutional distribution complements ICE’s retail bonds, data and clearing. More connected liquidity could improve execution and create cross-selling opportunities.

Bear view: An expanded platform does not automatically reverse fee compression or competition. The synergy-adjusted purchase multiple can obscure the time and investment required.

Our take: The strategic fit is convincing; the return is still to be earned. We credit no revenue synergies in the current estimate and retain the existing multiple while tracking the $100 million cost target.

Debate: Recurring growth versus rising capacity costs

Bull view: Two higher recurring-growth guides and new data products indicate that demand outlasts a trading spike. Mortgage subscriptions are beginning to improve without relying entirely on temporary items.

Bear view: Temporary revenue and market-sensitive index assets support the quarter, while capacity and compensation increase costs ahead of revenue. The earnings benefit may arrive more slowly than the investment.

Our take: The higher data and mortgage forecasts partly offset lower exchange revenue. Total EPS still moves slightly down because the stronger recurring trajectory does not fully cover normalization and spending.

Debate: Technology disruption versus licensed distribution

Bull view: Governed proprietary data and embedded workflows make ICE a supplier to AI and new trading channels. Benchmark licensing can monetize new distribution.

Bear view: Perpetual contracts can divert speculative activity, while tokenization can change collateral economics. Customer efficiency does not guarantee ICE captures the saving.

Our take: Current open interest and recurring growth support the existing franchise. We treat new channels as opportunities to protect and extend that franchise, with no separate premium until commercial returns become visible.

Our Estimates & Valuation

Our revised FY2026 estimates value ICE’s existing operations. The acquisition is expected to close in the first half of 2027, subject to shareholder and regulatory approvals, so no MarketAxess operating earnings or acquisition debt interest is included in the 2026 forecast. The twelve-month target nevertheless carries execution risk as the deal may close within that horizon.

The revised calculation is (($11.15 billion − $4.21 billion − $0.720 billion) × 75% − $0.072 billion) ÷ 566 million shares = $8.115, rounded to approximately $8.10. Against the prior unrounded $8.19, the $90 million reduction in operating income costs about $0.12 per share. Lower non-operating expense, modestly lower noncontrolling interests and fewer shares recover about $0.04. The change is an earnings revision, not a valuation-method change.

At $156.27, ICE trades at 19.3 times our rounded FY2026 EPS estimate. We retain 22 times as an analyst valuation assumption reflecting durable benchmark liquidity, recurring data and cash conversion, tempered by transaction cyclicality and acquisition exposure. It is not a peer-derived multiple. We reduce conviction from 6/10 to 5/10 because the operating evidence is stronger in recurring businesses while the capital commitments are larger.

The horizon is twelve months from this report, with FY2026 adjusted EPS retained as the common valuation base. We assume four dividends at the current $0.52 quarterly rate, subject to future board approval. The $178.20 base value requires delivery of earnings and a rerating from 19.3 times to 22 times. If the market maintains the current implied multiple on unchanged earnings, the return is approximately the 1.3% dividend yield.

The bull case assumes $11.60 billion of revenue, $4.18 billion of adjusted costs, $700 million of non-operating expense, 25% tax, $72 million of NCI and 565 million shares, producing approximately $8.80 EPS. It needs stronger trading and recurring growth with expense discipline; a 24-times multiple recognizes greater durability. The bear case assumes $10.60 billion of revenue, $4.25 billion of costs, $780 million of non-operating expense, 25% tax, $80 million of NCI and 568 million shares, producing approximately $7.20 EPS. At 18 times, that represents a meaningful loss without assuming the core franchise collapses.

An annual $100 million revenue shortfall would cost up to $0.13 EPS before expense offsets. Once acquisition financing is in place, a one-percentage-point higher interest rate on an illustrative $6.0 billion of borrowing would cost about $0.08 of annual EPS after 25% tax. This financing sensitivity is outside the 2026 base calculation and helps explain why the deal reduces conviction even though management expects adjusted accretion.

Assessment: The 15.4% base total return remains sufficient, in our judgment, to outperform the S&P 500 over twelve months. The approximately 15.7% bear loss is similar in scale, and the base return depends materially on multiple expansion. Maintaining Outperform rests on the existing business at this price; the acquisition has not earned a higher target.

Thesis Scorecard Post-Earnings

Commitments to watch: At the October 29 Q3 call, assess recurring exchange growth against the new high-single-digit annual outlook, FIDS against 7%–8%, mortgage recurring revenue around $406 million and adjusted costs of $1.063 billion–$1.073 billion. Hall 6 is now expected to begin recognizing revenue in early 2027; adoption and economic returns from private-credit data and Treasury clearing remain open tests. New checks are $100 million of MarketAxess annual savings by year three, one-third in year one and two-thirds by year two, and leverage at or below 3.0 times within 18–24 months of closing. Tokenization’s later-2026/early-2027 implementation expectation also becomes a dated signpost.

What would change our view: A recurring-growth downgrade, renewed mortgage subscription weakness or energy normalization beyond the reduced forecast would weaken earnings support. Deal financing or integration assumptions that undermine deleveraging and cash returns would reduce conviction further. Repeated underlying mortgage growth, data demand through the harder comparisons and evidenced acquisition returns would improve it. A price near our base value without a higher earnings outlook would remove much of the prospective excess return.

Overall: The recurring franchises strengthen the standing operating thesis, while Q2 confirms the turnover sensitivity already anticipated. The new acquisition is the material change in risk, not a reason to discard the benchmark and data pillars or assume their economics automatically transfer to the target.

Action: Maintain Outperform at $156.27 with a $178.20 twelve-month value and lower conviction. The target values the existing earnings base; disciplined delivery on the enlarged capital program is now essential to preserving the expected return.