FactSet Pricing: What It Costs in 2026

What does FactSet cost in 2026?

FactSet does not publish a rate card. Its pricing page (factset.com, read 31 August 2026) routes buyers to a sales conversation and a customised solution rather than listing tiers, seat classes or feed charges, so any number you find online is a third party's reconstruction of somebody else's contract.

The most widely circulated third-party figure is Wall Street Prep's estimate of roughly $12,000 per year for the full product, republished in Koyfin's comparison page at the same $12,000 per year (both read 31 August 2026). Treat that as an estimate of one configuration, not a list price. Vendr's 2026 page describes the structure more usefully than the headline does: per-user subscriptions that vary by user type, data feeds and content.

That last sentence is the whole problem with the pages ranking for this query. They quote a seat number and stop, when the seat number is one of four things being billed.

This article separates those four, states which figures are publicly sourceable today and which are not, and then does the arithmetic nobody publishes: what share of a typical FactSet bill is constituent, index and reference data that a single-purpose CSV feed delivers at a price you can read on a public page.

How FactSet's pricing model is actually structured

A FactSet quote is generated by four independent variables, and changing any one of them moves the total: seat class, content set entitlements, delivery mechanism, and contract term. A single headline figure describes a bundle priced on all four, which is why the same figure rarely survives contact with a second buyer.

Seat class is the dimension buyers understand best. Vendr's 2026 page (read 31 August 2026) describes FactSet subscriptions as per-user and varying by user type, data feeds and content. FactSet's own public pages do not publish the names or the prices of any seat class, so anyone quoting you a per-class figure is quoting a contract they saw, not a rate card.

Content set entitlements are the dimension buyers understand worst. Fundamentals, estimates, ownership, index constituents, pricing history, fixed income analytics and third-party content licensed through the platform are separately entitled, and two seats of identical class can differ by a wide margin depending on what is switched on behind them.

Delivery mechanism is the third. Workstation access, a bulk feed and programmatic API access are different products with different licences, and the fact that they surface the same underlying content does not make them one line on the order form.

Why two firms quote wildly different FactSet numbers

Vendr's 2026 page (read 31 August 2026) describes FactSet subscriptions as varying by user type, data feeds and content, which means two firms with the same seat count are not necessarily buying the same thing. How FactSet prices redistribution of data outside the firm relative to internal consumption is not published on its public pages, so treat any redistribution terms you are quoted as specific to your contract rather than as a known market rate.

This is one reason the figures circulating in forum posts spread so widely. They report a total without the seat type, entitlement schedule, delivery mechanism or term that produced it, and none of those details is recoverable from the number afterwards.

The practical consequence for a buyer is that the seat count is the wrong thing to negotiate first. Ask for the entitlement schedule in writing, line by line, with the delivery mechanism and redistribution scope named against each entitlement. A quote without that schedule cannot be compared to anything, including your own renewal from last year.

FactSet API pricing and data feed costs

Programmatic access is licensed and billed separately from workstation seats, and FactSet does not publish a price for it. The FactSet Prices API page (developer.factset.com, read 31 August 2026) documents what the API does in detail and prices none of it.

What that page does document is worth knowing before you ask for a quote. As of 31 August 2026 it describes endpoints for Prices, Reference, Returns, Dividends, Splits, Shares, Market Value, High Low, Database Rollover and Batch Processing, with global coverage across equities and fixed income, control over date ranges, currencies and rolling periods, and connection to other content sets through FactSet's hub-and-spoke symbology.

That symbology point is the substantive argument for the platform, and it is a real one. Joining prices to fundamentals to ownership to index membership without maintaining your own cross-reference layer is expensive to replicate and easy to underestimate until you have tried.

No per-call, per-seat or per-annum API figure is verifiable from FactSet's public pages as of 31 August 2026. What drives the number in practice is the same set of variables as the workstation line: how many users or applications are entitled, which content sets those calls can reach, how often the data refreshes, and whether the output leaves the firm. If you are budgeting, budget the API as its own line item and assume it does not shrink because you already pay for seats.

FactSet fundamentals pricing and the content set problem

Fundamentals, estimates, ownership, index constituents and reference data are separate entitlements, not features of a seat. That is why the fundamentals line is often the fastest-growing part of a renewal even when headcount is flat: the data footprint grew, the seat count did not.

The buyer's real question is not the annual total. It is the cost per content set actually used, measured against what that content set would cost bought alone.

Here is an audit method that survives contact with procurement:

• Pull twelve months of usage broken out by content set and by user, not by seat count.
• Separate entitlements that feed automated processes from entitlements a human opens.
• For each content set with meaningful automated dependency, write down the specific fields consumed, not the product name.
• Price each of those field lists against a single-purpose alternative with a published price.
• Flag every content set with zero recorded use and every one used by a single person for a single recurring report.

The last two categories are where renewals are actually reduced. A content set consumed once a quarter by one analyst is a candidate for a cheaper substitute or a manual process, and it does not become less of a candidate because it is bundled.

There is a trap in this exercise that has nothing to do with price. Fundamentals definitions differ between vendors: what counts as trailing earnings, how negative earners are handled, whether the aggregate is share-weighted or a simple sum, and which restatement vintage is used. We wrote about that in detail in P/E Ratio S&P 500: How the Number Is Built, and the point matters more than the price line, because a substitution that changes the definition silently changes every downstream model that consumes it.

FactSet pricing for individuals, students and academics

FactSet is licensed to institutions. Its public pages do not advertise an individual retail plan or publish an academic rate card, so there is no figure to quote here, and any number presented as "the student price" is unsourced.

University access generally arrives through a campus, library or business-school licence negotiated by the institution, with terms that the institution and the vendor agree privately. If you are a student, the practical answer is that your access exists or does not exist depending on what your school bought, and the price is not yours to see.

For an individual analyst without institutional access, the realistic substitution is a set of single-purpose subscriptions with published prices, each covering one thing well. That is a genuinely worse product in the ways you would expect: no integrated symbology across content sets, no estimates or ownership data, no screening workstation, no portfolio analytics, and no support desk that knows your models.

What you get back is a price you can read before you buy and a data footprint you actually understand. For a solo practitioner running a systematic process on published index and ETF data, that trade is often correct. For anyone doing bottom-up diligence across private comparables, estimates history and ownership, it is not.

What FactSet pricing looks like next to Bloomberg, Capital IQ and Refinitiv

Comparing these four FactSet alternatives on price is mostly comparing other people's guesses. Comparing them on licensing structure would be more useful, but only one of the four rows below can be filled from a primary page read today, and the honest thing is to show which.

Wall Street Prep's comparison page (read 31 August 2026) is the most-cited public source of per-product estimates for this group, and it lists roughly $12,000 per year for FactSet's full product. Those are its estimates, not vendor list prices, and I am not restating its figures for the other three here, including Refinitiv pricing, because I cannot verify what each vendor publishes today from a primary page.

The empty cells are the finding. If a comparison page fills them confidently without naming the vendor page it read and the date it read it, the figures in it are reconstructions, and a reconstruction of one buyer's contract tells you nothing about yours. When you evaluate the other three, ask each vendor directly for the licensing unit, the entitlement schedule and the delivery terms in writing, and compare those documents rather than published summaries of them.

What FactSet genuinely does better is worth naming plainly, because a comparison where one side never wins is an advertisement. The breadth of content joined under a single symbology is the strongest part of the product, as the Prices API documentation (developer.factset.com, read 31 August 2026) sets out in its description of hub-and-spoke symbology linking to other content sets. The Excel and portfolio analytics workflow is built for analysts who live in a spreadsheet and want the platform to feed it. If that describes your team, the substitution arguments below do not apply to your core seats.

When a constituent and index CSV feed replaces part of the bill

If your requirement is daily index and ETF constituent lists, sector weights, identifiers and fundamentals delivered as files, that is a narrow requirement, and a full-platform seat is an expensive instrument for meeting it. The honest test is whether a human needs to explore the data interactively or a process needs to consume it on a schedule.

AmericanETP's price is published and confirmed by customer: $1,500 per year or $150 per month for an individual, and $2,500 per year or $250 per month firm-wide. That covers constituent lists for 3,878 US and global indexes and US-traded ETFs, updated twice daily at 6pm EST and around noon EST, with Bloomberg FIGIs included in the holdings files, a fundamentals file carrying ETF description, leverage and constituent counts, archived constituent datafiles back to July 2009, and FTP access. Sample files are available on the Downloads page and the FIGI and identifier extensions are described on the Bloomberg Extensions page.

Who each of those lines actually fits, in plain terms:

• One quant or one PM running systematic index and ETF work, needing files rather than screens: the individual feed line.
• A team of five to fifty at an RIA, fintech or exchange where several systems consume the same constituent files: the firm-wide line, because the constraint is distribution inside the firm, not features.
• An analyst team doing bottom-up diligence across estimates, ownership and private comparables: a full platform seat, because no file feed substitutes for that content.
• A firm redistributing derived data to clients: a negotiated platform or feed licence with explicit redistribution scope, whatever else you buy.

What the CSV feed does not do should be stated as clearly as what it does. There are no estimates, no ownership data, no screening workstation and no analytics suite. It displaces a content set, not a platform, and a buyer who pitches it internally as a platform replacement will be embarrassed in the first week.

It also solves problems the platform line item tends to bury. Reconciling why two vendors report different index-level fundamentals is a recurring cost, and we walked through the mechanics in PE of S&P 500: Why the Numbers Disagree. Tracking membership changes without diffing full files by hand is another, which is what the Constituent Change Report exists for. And if your firm is moving toward extended sessions, the timestamp and reference-data assumptions that break are covered in 24×5 Trading: What It Breaks in Data.

How to negotiate or reduce a FactSet renewal

The leverage in a renewal comes from evidence, not from a target discount. Arrive with a usage inventory mapped against the entitlement schedule, with each narrow requirement priced independently, and the conversation stops being about percentage off and starts being about what you are actually buying.

Run the sequence in this order:

1. Inventory entitlements against twelve months of usage, by content set and by user, and identify entitlements with no recorded consumption.
2. Separate redistribution rights from consumption rights on every line, so you know which lines carry scope you do not use.
3. Price the narrow requirements independently and bring the published figures. A named alternative with a public price is a fact in the room; an assertion that "this should be cheaper" is not.
4. Bring dated third-party figures, not forum numbers. "Wall Street Prep's comparison page, read 31 August 2026, estimates around $12,000 per year for the full product" is a sentence a vendor can respond to. "I heard someone pays eight" is not.
5. Align the ask with the contract term. Multi-year commitments are where discounts live and where flexibility dies, so decide which you are buying before the term is set.

Treat the seat line, the feed line and the API line as three separate negotiations, because they are three separate licences on the order form. Consolidating feeds mid-term and adding them mid-term produce very different conversations, and it is worth knowing which one you are having before you open it.

One thing not to do: never cancel a content set before confirming which downstream models depend on its identifiers. The identifier dependency is usually deeper than the data dependency, and discovering in January that a nightly job resolved securities through an entitlement you dropped in December is a worse outcome than any renewal figure. Map identifier lineage first, then cut.

FAQ

Can I use FactSet for free?

There is no free tier advertised on FactSet's public pages as of 31 August 2026. The site directs prospective buyers to contact sales, and demonstrations or trials, where offered, are arranged through that route rather than self-service. Access for students generally comes through a university or business-school licence rather than an individual free plan.

How much does FactSet cost per user in 2026?

FactSet does not publish a per-user price. The most-cited third-party estimate is Wall Street Prep's roughly $12,000 per year for the full product, repeated by Koyfin's comparison page at the same figure (both read 31 August 2026), and both are estimates rather than vendor list prices. Vendr's 2026 page describes subscriptions as per-user and varying by user type, data feeds and content, which means a single per-user number cannot describe a firm's bill without the entitlement schedule behind it.

Does JP Morgan use FactSet?

This is not something to answer from memory. No dated public source in front of me confirms which vendors any specific bank licenses, and vendor relationships at that scale are rarely disclosed with enough precision to be useful anyway. It would not help your budgeting even if it were confirmed: what sets a bill is the seat class, entitlement schedule, delivery mechanism and term behind each seat, none of which travels with the vendor's name.

Who is FactSet's biggest competitor?

The commonly compared set is Bloomberg, S&P Capital IQ and LSEG Refinitiv Eikon, which is the grouping Wall Street Prep's comparison page uses (read 31 August 2026). Which of them is closest to FactSet depends on your workflow, and I am not assigning each vendor to a workflow here, because I have no dated primary source for how those vendors currently position themselves. On single content sets, the competition is narrower still, coming from dedicated feed and index data providers rather than from full platforms.

Is FactSet cheaper than Bloomberg?

The two are commonly described as sitting at different points on price, but FactSet does not publish a rate card (factset.com, read 31 August 2026), and I have no dated primary source for Bloomberg's published terms, so a definitive comparison is not available here. The workable approach is to compare the two quotes in front of you rather than published summaries of either. A configuration comparison beats a headline comparison, because a light seat on a narrow content set and a full seat with redistribution rights are not the same purchase at any vendor.

What do the FactSet pricing threads on Reddit actually tell you?

They tell you the shape of the distribution, not the price. Posts of this kind report individual contract outcomes without the entitlement schedule, seat class, delivery mechanism or term that produced them, so the wide spread people notice is mostly variance in what was bought rather than variance in what was charged. Use them to confirm that the number is negotiable and configuration-dependent, then get the entitlement schedule in writing and price your narrow requirements independently.

Prices in this article were checked on 31 August 2026 and change without notice. Every figure attributed to a third party is that third party's estimate, not a vendor list price.

If part of your data bill is index and ETF constituent lists, sector weights, identifiers and ETF fundamentals delivered as files, that portion has a published price you can check before you talk to anyone. See what the feed covers at americanetp.com, or get in touch with the specific fields your models consume and we will tell you plainly whether they are in the files.