
Last 4 earnings calls · through Q4 FY2026
The bull case is that AI adoption expands how much data clients consume from FactSet rather than shrinking the number of workstation seats it sells.
Management has moved more renewals into multiyear enterprise agreements, which lengthens contracts and makes the revenue base less dependent on seat counts.
Acceleration across the Americas and Asia Pacific and competitive displacements in banking and wealth suggest the commercial changes are producing new business.
Productivity programs and AI-assisted engineering are expected to translate into margin expansion and a smaller share count, supporting per-share results.
The bear case is that cheaper AI tooling lets clients and model providers assemble data and workflows themselves, eroding seat demand and pricing power.
Growth is expected to slow next year, and management warns the sales metric it guides on may not track reported revenue as more contracts become consumption-based.
Cloud, token and cybersecurity spending is rising faster than the modest margin improvement targeted, so productivity savings may be spent rather than kept.
In short: Sales are accelerating across regions and client types, while next year's growth guide assumes a slowdown.
Revenue in the largest region grew faster than the group as wealth and banking demand held up, which is where management redirected sales incentives and marketing spend.
AI went from a line management declined to break out to more than a tenth of new sales value, which supports the claim that AI adds revenue rather than replacing existing products.
Management guides organic sales growth of 5% to 6.5% for the year ahead after a faster year, saying renewals are back-half weighted and it has assumed no AI acceleration.
As more contracts carry usage-based components tied to AI workloads, the sales value FactSet reports may convert into reported revenue less predictably than fixed subscriptions did.
In short: Margins fell this year on deliberate spending, and management expects only modest expansion next year.
A year ago management declined to discuss medium-term targets; it now calls this year's margin the floor and plans to lay out a multiyear plan at an investor day.
The guided improvement for the year ahead is small against the decline taken this year, so a fuller recovery depends on productivity programs management describes as early.
AI coding tools were first described as freeing engineers for product work; management has since cut about a tenth of the technology workforce, which points to cost pressure behind the productivity story.
In short: FactSet is displacing incumbents with data and AI connections, but cheaper AI tooling threatens its pricing.
Several hundred clients now consume FactSet data through AI connectors, and management says those deals usually expand into data feeds and workstation agreements as well.
Renewals are being extended by roughly 30% while pricing is broadly preserved, which indicates clients accept the value of the platform and gives revenue more durability.
Cheaper models and agent frameworks let clients and new entrants assemble data and workflows themselves, which is the main route by which seat-based pricing could erode.
A court certified a class of subscribers to the municipal data joint venture in August 2026, creating potential damages and scrutiny of how that data is priced.