MSCI Just Posted Its Best Sales Quarter Ever and Told You Not to Expect a Nickel of Margin

MSCI sold more new subscription business in the September 2026 quarter than in any quarter in its history — best-ever recurring net new sales in both Analytics and Private Capital Solutions, and the best third quarter ever in Index. Total revenue grew 12.4%, asset-based fees 25%, subscription revenue 8%, adjusted EPS 14%. And the stock has gone nowhere for three months. That divergence, not the record book, is the story.

Records everywhere, and accelerating

The operating detail is genuinely strong. Total run rate growth of 12%, an ABF run rate approaching $1 billion, $76 billion of equity ETF inflows, $23 trillion of AUM benchmarked to MSCI indices of which roughly $15 trillion is active. Retention of 95%, up 30 basis points. Recurring net new sales of nearly $55 million, up 88% year over year in Private Capital Solutions, 63% in Analytics, 40% in Index — driven not by asset managers but by new client segments: hedge funds at 32% subscription run-rate growth inside Index, banks and brokers at 15%, wealth managers and asset owners at 10% each, custom index run-rate up 26%.

The most important line in the call, buried in the Q&A, is that the new-product engine is compounding. Chief Financial Officer Andy Wiechmann: “that contribution to new sales from new products increased more than 50% compared to a year ago” — Andrew Wiechmann · 2026-10-20. Three months ago the same measure was running at roughly 40%. That is the difference between a product cycle and a launch splash.

The refusal

Then Faiza Alwy of Deutsche Bank asked about operating leverage, and the answer was not the one a shareholder wants on a record quarter.

The answer is no margin expansion... we're applying AI to... gradually lower the growth rate of EBITDA expenses from mid- to high single digits to the low single digits and eventually to negative growth. And then we want to put all that money into the growing of the business.

It was startling enough that Ashish Sabadra of RBC asked a second time whether he had heard it right. Henry Fernandez clarified that only 20 to 50 basis points of accidental expansion will leak through, an artefact of conservative budgeting rather than intent: “I'd rather see a company double or triple the size in 10 years compared to what it would otherwise be if we let the margin rise” — Henry Fernandez · 2026-10-20.

Read against his own history, this is not new doctrine, but it has hardened into liturgy. Twelve months ago he was promising the opposite: “AI will dramatically increase our margins, really dramatically” — Henry Fernandez, Chairman and Chief Executive Officer · 2025-10-28 — with the caveat that he did not want investors to bank those gains. In July he described himself as “very aggressive risk takers, but we're prudent financial managers” — Henry Fernandez, Chairman and CEO · 2026-07-21. Now the reinvestment is no longer a choice management might make; it is the stated financial policy. The keyword set has followed: this quarter's ranked terms are organic investment, SaaS industry and agentic capabilities — strategy abstractions, where three months ago the leaders were run-rate growth and recurring net new sales. When the numbers are this good, you would expect the numbers to lead. They did not.

The market has already rendered a verdict of sorts. MSCI trades essentially flat over 90 days, 7% below its July high, at 12.2 times revenue — down 18% year over year — and 30 times net income, down 21%. Nothing broke in the P&L: total revenue keeps compounding and operating margin sits at 56.2%, up 1.2 points.

What has changed is how the company funds its equity story. Effective net cash stands at negative $6.1 billion, 45% worse year over year, with liabilities to assets at 148%, up 31.5 points. Interest coverage has fallen to 6.9 times, down 25%. Management explicitly raised interest expense guidance to reflect higher revolver balances supporting $627 million of repurchases between July 1 and October 19 at an average of $562 a share — part of nearly 8 million shares for $4.4 billion since the start of 2024, alongside the chairman's personal purchase of 57,000 shares for $30.1 million. Leverage up, buybacks up, margin capped. That package only works if the organic reinvestment returns are as extraordinary as claimed; Henry says he is turning away projects with high double-digit IRRs. Shareholders are being asked to take that on faith for a decade.

The AI argument quietly changed shape

For three years MSCI sold AI as an efficiency story and a product enabler. This quarter it became a distribution-war argument. Henry's cable-to-streaming analogy is the boldest thing said on the call.

The SaaS industry is the equivalent of cable... there came a streaming technology that for anybody who has access to an Internet connection, you can have access to all the content in the planet. So we believe... the streaming technology is the agentic technology that could bypass all of that.

Translated: MSCI thinks agentic delivery lets it leapfrog the workflow-software layer it never wanted to own, pushing consumption from hundreds of users per client to hundreds of thousands. Supporting evidence is real — “It used to take us several weeks to build and fully calibrate a new model, we can now do it in the space of a couple of hours” — Alvise Munari · 2026-10-20 — and the factor model factory and new specialized data sets are the products that factory is now shipping. Alvise Munari framed the demand simply: “the unit of conversation in the market has gone from being the single stock to being the portfolio” — Alvise Munari · 2026-10-20.

The timing is awkward. The tape is de-rating exactly this narrative: agentic operating systems are down 31% over twelve months across the three names that carry the label, and AI data center exposure has fallen 5.5% in thirty days with 47 of 53 constituents lower. The market is currently skeptical of whoever pays for the agentic build-out. MSCI's pitch is that it is not the payer — it buys models rather than building them, rents no data centers, and sells the content that flows through someone else's pipes. That is a defensible position, but it is also untested, and the company is asking investors to fund it with the margin line.

What went quiet

Index arbitrage, a top-ten theme last quarter when Henry was explaining the trading and liquidity ecosystem, has vanished from the vocabulary; in its place is market makers, with the admission that “we're not even yet penetrating the market makers” — Henry Fernandez · 2026-10-20. Asset managers — half the subscription base — are now described as a steady 6% grower. And Sustainability and Climate, once a headline growth engine, has compressed to a single paragraph: elevated cancels, one sizable downsell, fourth-quarter recurring sales guided to roughly flat to slightly positive, and a hopeful pivot to physical risk via the First Street acquisition. The protracted cyclical downturn Henry identified in July is now an assumption rather than a debate.

So the bet is explicit: a bigger company later instead of a richer margin now, funded partly by debt and paid for in multiple. The fourth quarter is the decisive sales quarter, management flagged lumpiness, and the expense guide sits at the top of its range. The place to check the thesis is 2027, when the AI-driven expense algorithm is supposed to visibly bend and the reinvestment either shows up in the run rate or does not.