> ## Content Index
> Fetch the complete content index at: https://slippage.ghost.io/llms.txt
> Use this file to discover other available public pages before exploring further.

# ICE buys MarketAxess
- URL: https://slippage.ghost.io/ice-buys-marketaxess/
- Published: 2026-07-31T22:57:42.000Z
- Updated: 2026-07-31T23:35:48.000Z
- Description: ICE is paying $6bn for MarketAxess just as venue fee capture erodes and the public data on who trades what goes dark — a week in which the price of measurement kept rising.
- Author: Vincent Leung
- Tags: Venues & Exchanges, Public Credit, Data & Technology, Private Credit, Treasury, Asset Management

## ICE buys MarketAxess as venue fee capture erodes

ICE agreed on 30 July to acquire MarketAxess for $167 a share in cash — roughly $6.0bn of equity value, $5.7bn enterprise value, a 33% premium to the 29 July close and about 10.6x LTM EBITDA pro forma for $100m of annual run-rate expense synergies. The whole thing is funded with newly issued debt, taking ICE's gross leverage to approximately 3.4x at closing, which is expected in the first half of 2027\. Tradeweb reported second-quarter revenue of $558.9m the same day, up 9.0%, and the shares closed down 9.9%; cash credit fees per million fell 11.4% on mix shift toward European credit and portfolio trading, and Billy Hult told the call the deal was "a validation of the fee pool" before adding, "We are market share takers, period." MarketAxess reported revenue of $218.4m, flat year on year, with commission revenue down 3% and US credit commissions down 9% against record portfolio trading ADV of $2.0bn, up 33% — and simultaneously suspended its earnings call, its 2026 guidance, its medium-term targets and its monthly volume releases. For scale on where share actually sits: Tradeweb's June fully electronic US high-grade share was 20.2% against MarketAxess's estimated 17.9%, while Trumid's second-quarter ADV reached $10.0bn, up 50% year on year.

*Two things are happening at once, and only one of them is being reported: the venue layer is consolidating at the exact moment per-unit fee capture is falling, and the public series that let a desk check its own venue mix against the market is being switched off — June's 17.9% and 14.9% estimated shares may be the last MarketAxess prints anyone sees. Nothing in the explicit cost stack changes before 2027, so the number worth chasing is the total one: ICE already licenses evaluated pricing, reference data and indices alongside ICE Bonds and would own CP+ as well, no published analysis of that overlap exists anywhere, and the renewal conversation where it matters is the one happening now rather than after close.*

[ICE press release](https://ir.theice.com/press/news-details/2026/Intercontinental-Exchange-to-Acquire-MarketAxess-Creating-a-Premier-Fixed-Income-Marketplace/default.aspx?ref=slippage.ghost.io) · [The DESK](https://www.fi-desk.com/ice-buys-marketaxess/?ref=slippage.ghost.io) · [Tradeweb Q2](https://www.tradeweb.com/newsroom/media-center/news-releases/tradeweb-reports-second-quarter-2026-financial-results/?ref=slippage.ghost.io) · [MarketAxess Q2](https://www.businesswire.com/news/home/20260729765121/en/MarketAxess-Reports-Second-Quarter-2026-Financial-Results?ref=slippage.ghost.io) · [Trumid Q2](https://www.prnewswire.com/news-releases/trumid-reports-june-and-q2-2026-trading-highlights-302821454.html?ref=slippage.ghost.io)

## Investment grade is becoming an AI capex trade

Four of the large hyperscalers issued roughly $194bn of bonds through 7 July, against about $108bn in all of 2025, with Goldman Sachs projecting some $250bn across five issuers this year and $400bn in 2027\. The market is charging for it: new-issue concessions have gone to 12bp from 2.25bp in 2025, Apollo data show cover ratios falling from nearly 5x in February to below 2x in July, and 78 of the 91 hyperscaler bonds issued in 2026 with comparable pricing data traded at higher yields on 28 July than at issue, a median 22bp wider. Brown Advisory's Colby Stilson: "We're already seeing fatigue within credit markets in supporting this massive debt issuance." The hedging response has moved to single names — Reuters put Oracle five-year protection around 200bp on 29 July against Meta at 93bp, Nvidia at 78bp and an investment-grade index near 53bp, with technology-sector CDS trading close to $650m in the second quarter, up 20% sequentially and nearly 600% year on year, inside a whole-market daily average that rose to $16bn from $13bn. ISDA's study published the same day put 2025 CDS notional traded at a record $41.8trn, 93.3% of it index.

*The connection is the reason single names are moving rather than the index: a benchmark whose spread is increasingly a function of one capital expenditure cycle no longer isolates the exposure anyone actually wants to hedge, so the demand migrates to instruments thin enough that the hedge itself repositions the cash bond. That turns a passive index allocation into an unpriced bet on data-centre returns and puts a concrete requirement on the construction process — issuer-level look-through on AI capex exposure across index, ETF and separate-account sleeves, which most attribution stacks were never asked to carry and cannot currently produce without hand-assembled mappings.*

[Hyperscaler issuance](https://finance.yahoo.com/markets/stocks/articles/hyperscaler-debt-binge-pushes-yields-134825104.html?ref=slippage.ghost.io) · [Reuters on AI credit CDS](https://finance.yahoo.com/markets/options/articles/explainer-credit-default-swaps-why-110342195.html?ref=slippage.ghost.io) · [ISDA CDS study](https://www.isda.org/2026/07/29/cds-market-dynamics-five-years-of-activity-and-a-record-2025?ref=slippage.ghost.io)

## LSEG starts metering the data agents consume

At its H1 results on 30 July, LSEG said the AI search tool in Workspace reached 17,000 active users, included in subscriptions under a fair-use policy with Deep Research sold as a premium add-on, and confirmed it has introduced an additional use-case licence for AI applications of its data plus usage-based charges for API and MCP access. MCP tool calls rose nearly fivefold from May to June across more than 200 engaged clients, now about a third of the firm's AI-related commercial conversations, and David Schwimmer put the reason plainly: "Agents tend to consume roughly 10 times the amount of data that humans do through the MCP channel." The CFO was equally plain that this year is about usage rather than revenue — "It's minimal… We will see more in 2027." Two surveys landed in the same week from the other side of that transaction: Clearwater Analytics, polling 178 senior fund-manager executives, found 95% had raised AI budgets in the past year and 85% plan to raise them by 50% or more over the next twelve months, while only 56% rate their own data as accurate against 79% who rate it complete; F2 Strategy's survey of 40 wealth firms and broker-dealers found 64% lack a unified data infrastructure for AI, rising to 83% among bank and trust respondents.

*Per-seat licensing was always a proxy for human consumption, and agents break the proxy — which is why the pricing model moves to metered consumption precisely as budgets scale, and why the explicit cost of AI is quietly relocating from model spend to the data bill underneath it. Comparable agentic-access terms from Bloomberg, FactSet or S&P have not been published, so the like-for-like comparison is unavailable; what is available is the shape of the total cost, which is a variable and largely unforecastable line item whose only real control is a normalised internal data layer, and the surveys say roughly two thirds of firms do not have one to control it with.*

[Waters Technology](https://www.waterstechnology.com/emerging-technologies/7953190/lseg-caps-free-prompts-as-it-moves-to-increase-ai-revenues?ref=slippage.ghost.io) · [LSEG H1 results](https://www.lseg.com/en/investor-relations/financial-results/2026-h1-interim-results?ref=slippage.ghost.io) · [Clearwater study](https://www.sfnet.com/home/industry-data-publications/the-secured-lender/tsl-express-daily-articles-news/tsl-express-daily-articles-news/2026/07/28/global-study--fund-managers-say-ai-will-transform-asset-management-operations-within-12-months?ref=slippage.ghost.io) · [F2 Strategy survey](https://www.investmentnews.com/transformation/ai-in-wealth-management-budgets-surge-but-roi-remains-elusive/267540?ref=slippage.ghost.io)

## Public market vendors buy private credit's plumbing

Three transactions in four days, all aimed at the same missing layer. ICE launched ICE IDs on 28 July, persistent identifiers for private credit instruments that run from origination through payoff and survive credit events, with Apollo as anchor partner applying them to its originated assets; Chris Edmonds described them as "the cornerstone that we build everything else off of, and which link together key data points," and ICE already manages reference data across more than 35m instruments in 210-plus markets. On 29 July Bloomberg agreed to acquire Canoe Intelligence, which automates alternative-investment document ingestion for over 500 institutional clients with more than $11tn of assets under service, on top of Bloomberg's existing coverage of 50,000 funds and 16,000 direct loans. The same day TA Associates took a growth stake in Oxane Partners, whose Panorama platform serves more than 100 clients representing $1.4trn of aggregate client AUM. No pricing or licensing terms were disclosed for any of the three. On ICE's earnings call two days later, Jeff Sprecher tied the identifier work directly to the MarketAxess deal, saying that involving MarketAxess "would really accelerate where Apollo and Marc Rowan wanted to take their business."

*What links these is not that they are all private credit deals but that each buys a piece of the same join key — an identifier, a document-to-structured-data pipeline, a facility-level book of record — and the buyers are the firms whose existing business is licensing exactly that kind of infrastructure in public markets. Standards get priced before they get adopted, so the practical consequence is that consolidated look-through across private credit exposures becomes technically achievable and contractually expensive at roughly the same moment; anyone who will eventually need issuer-level valuation transparency across managers should be pinning identifier and redistribution terms down now, because the leverage in that negotiation is at its maximum while three vendors are still competing to set the standard.*

[ICE IDs](https://www.marketsmedia.com/ice-offers-ids-for-private-credit/?ref=slippage.ghost.io) · [Bloomberg–Canoe](https://www.thetradenews.com/bloomberg-to-enhance-private-markets-offering-through-canoe-intelligence-acquisition/?ref=slippage.ghost.io) · [TA–Oxane](https://www.ta.com/news/oxane-partners-announces-strategic-growth-investment-from-ta/?ref=slippage.ghost.io) · [ICE Q2 call](https://www.fool.com/earnings/call-transcripts/2026/07/30/ice-ice-q2-2026-earnings-call-transcript/?ref=slippage.ghost.io)

## Clearinghouses cut prices with five months to go

DTCC published a readiness survey on 27 July covering FICC Government Securities Division full-service netting members, with a 92% response rate: about $1.2trn of daily Treasury cash activity already clears at FICC, with an estimated $300–400bn of average daily par value still to migrate before the 31 December cash-clearing date, and repo following on 30 June 2027\. Seventy-nine percent of respondents already have the required FICC account structures, but only around one third of dealers expect to offer client clearing for Treasury cash. Laura Klimpel said firms "are making meaningful progress toward implementation." Two days later the SEC approved ICE Clear Credit's Treasury clearing rule change, which halves the minimum Treasury guaranty fund contribution from $20m to $10m, caps custodial loss resources at $20m and investment loss resources at $10m, permits borrowed or non-cash assets during settlement liquidity events, expands eligible Treasury Participant types to include FCMs and clearing agencies, and separates the Treasury rulebook from the CDS clearing business.

*That residual $300–400bn a day is the contested flow, and the timing of ICE's approval is not incidental — halving the cost of showing up is how a challenger clearinghouse bids for volume that has not yet chosen a home, five months before it has to. The explicit cost is legible enough to model against FICC on guaranty fund contribution and clearing fees; the cost that will not appear in any comparison is that only a third of dealers intend to offer client clearing at all, which means access, not price, is the binding constraint, and a firm still holding a single clearing relationship in August is negotiating from the weakest position it will have all year.*

[DTCC survey](https://www.dtcc.com/news/2026/july/27/dtcc-survey-firms-progress-toward-us-treasury-clearing-deadline?ref=slippage.ghost.io) · [SEC order on ICE Clear Credit](https://www.federalregister.gov/documents/2026/07/29/2026-15249/self-regulatory-organizations-ice-clear-credit-llc-order-approving-proposed-rule-change-relating-to?ref=slippage.ghost.io) · [Markets Media](https://www.marketsmedia.com/market-makes-substantial-progress-in-implementing-treasury-clearing/?ref=slippage.ghost.io)

## A record bond outflow that never happened

LSEG's Lipper removed a fund from its weekly flow report on 25 July after identifying an outsized flow within the Short-Intermediate Investment Grade Debt classification. The headline series had shown US investment-grade bond funds with $7.1bn of net outflows for the week to 22 July — reported as a record weekly withdrawal — including a record single-day $8.2bn outflow on 20 July. Excluding the fund, the same week showed $1.54bn of net inflows. A Lipper spokesperson said the firm had "identified a sizeable flow within the Short-Intermediate Investment Grade Debt classification that has negatively impacted the fixed income universe," and was determining whether the movement reflected a transfer or another operational event.

*A record outflow and a modest inflow are two entirely different trades, and the gap between them was an unvalidated operational event sitting inside a vendor series that a great many people treat as an observation rather than an estimate. Flow data feeding a signal deserves the same revision history, outlier screen and vendor SLA as any pricing feed — and the cheap version of that discipline, checking whether a print survives the exclusion of its largest single contributor, costs nothing and would have caught this one on the day.*

[Reuters via GV Wire](https://gvwire.com/2026/07/25/lsegs-lipper-reviews-transaction-affecting-fixed-income-fund-flow-data/?ref=slippage.ghost.io)