FACTSET RESEARCH SYSTEMS INC (FDS) stock intrinsic value fair value estimate & owner earnings analysis

FactSet sells subscriptions to financial data, analytics and workflow software. About 82% of its subscription value comes from asset managers, hedge funds and other buy-side firms, and about 18% from banks and other dealmakers. Its main competitors are Bloomberg, S&P Global Market Intelligence and LSEG, with MSCI, Morningstar and BlackRock's Aladdin competing in specific niches. The model is unusually sticky. At August 31, 2026 annual subscription value (ASV) was $2.57 billion, up 7.0% organically. More than 95% of ASV renewed in each of fiscal 2024 and 2025, and retention was also above 95% in fiscal 2010, just after the financial crisis. No client is more than 3.5% of revenue, and revenue has risen every year since at least fiscal 2016. About 66% of ASV is in the Americas, 24% in Europe, the Middle East and Africa, and 10% in Asia Pacific. We value the business on the cash it generates before paying lenders. That means operating cash flow with working-capital swings removed, stock compensation charged as a real cost, and maintenance capital spending deducted, with interest added back. We then subtract net debt. For FactSet we made several judgements. Operating-lease payments are treated as an ordinary running cost, not a working-capital item. Capital spending ($109 million in fiscal 2025 and $115 million in fiscal 2026, mostly capitalized software) is treated as mostly maintenance, because the platform has to keep being rebuilt to stay competitive. The one-time fiscal 2026 charges are not carried forward: $46 million of restructuring, $19.5 million of make-whole awards for the new CEO, and a $12 million equity-investment impairment. The gains from divestitures are also excluded. We measure the business from fiscal 2023, the first full year after the March 2022 purchase of CUSIP Global Services changed its size. Cash flow before interest in fiscal 2023-2025 was $598 million, $621 million and $681 million, so our starting level is the median, about $621 million a year. That is deliberately below fiscal 2025's $681 million and below our rough read of fiscal 2026 (about $700 million, on figures from the company's year-end release; the audited 10-K is not yet filed). We credit growth only from the median level, not from the best recent year. Our central case assumes cash flow grows about 4% a year for the first two years, then tapers to 2% by year ten. The first step is already contracted: year-end ASV of $2.57 billion sits about 3.7% above fiscal 2026 revenue of $2.48 billion. After that we rely on the way FactSet has grown for many years, selling more to existing clients. We taper it because the company already serves 95 of the 100 largest asset managers, so there is less new ground to cover. We credit no margin expansion, even though management guides to higher margins next year. We assume cash flow then stays flat for ten more years, which the retention record and the long client tenures support, and after year twenty we credit only existing contracts eroding at the 5% a year that retention implies. We leave out growth from new AI products, competitor weakness, and management targets because none is secured. On this basis the business is worth about $11.1 billion and the company owes about $1.16 billion net, leaving $9.9 billion, or about $280 per share on roughly 35.5 million diluted shares. Our low case replays the last financial-sector downturn with today's cost pressures. In 2008-2010 FactSet's revenue growth slowed from 8.1% to 3.1% while margins held. We assume growth falls by about that much for two years (to roughly 1%) while margins slip a point a year, as they did in fiscal 2026 when technology and compensation costs rose faster than revenue. Cash flow therefore falls about 2.5% a year for two years, then grows 2% a year, stays flat in years eleven to twenty, and erodes afterwards. That gives a business worth $9.3 billion, net debt of $1.16 billion, and about $231 per share. Our high case assumes the fiscal 2026 pace holds. Organic ASV growth of 7.0%, with AI products already above 10% of Q4 ASV additions, fades from 7% to 4% over ten years, again with no margin expansion. That gives a business worth $13.5 billion, net debt of $1.16 billion, and about $347 per share. FactSet owes $1.375 billion: $500 million of 2.9% notes due March 2027, $500 million of 3.45% notes due March 2032, and a $375 million term loan due April 2028. Against that it holds about $265 million of cash and investments, of which we treat $100 million as needed to run the business. Because the 2032 and 2027 notes pay fixed interest well below what money would cost for the same period, we reduce net debt by about $63 million, and we add about $10 million of acquisition earn-outs. The March 2027 maturity is covered by cash, a $1.5 billion undrawn revolving credit line, and roughly $700 million a year of cash flow, and management has said it expects to refinance at a higher rate than 2.9%. Even in our low case, cash flow covers interest about nine times. Since our last look, the business has been mixed. Organic ASV growth accelerated to 7.0% from 5.7%, and $644 million of buybacks cut the share count by about 7%. Margins fell, however: GAAP operating margin dropped to 28.3% from 32.2% on restructuring, technology and compensation costs and a CEO transition, and management guides fiscal 2027 ASV growth to 5.0%-6.5%. Our earlier figures were built on a different basis, so they are not comparable. The main thing that could make this wrong is AI. FactSet's own filing warns that third parties could use AI to build tools that reduce demand for its products, and competitors are launching AI data services of their own. So far there is no sign of it in the numbers: retention is above 95% and AI products are adding ASV. But no case models it, because we have no evidence of its size. If cash flow were permanently 30% lower than our starting level, the shares would be worth about $186; if growth stopped after year ten, about $233. We could not reconcile the unpaid portion of the fiscal 2026 severance accrual or the price of the BCC Group acquisition; both look small. The margin-of-safety price is $180 per share. We are fairly sure of the earnings level: cash flow has risen every year for a decade, more than 95% of subscription value renews, and fiscal 2026 operating cash flow rose 13%. The worst plausible outcome is a lasting AI-driven loss of earning power that no case models. The balance sheet would survive it: net debt is about 1.7 times annual cash flow and interest is covered many times over. The business is mostly intangible, so its value rests entirely on clients continuing to renew. If we are wrong about AI by 30%, or growth stops after year ten, the shares are worth $186 and $233. Both of those values are above $180, and dividends are about $4.64 a year. The price assumes about a three-year holding period for the price to reflect value. In that time the business pays out its cash through dividends and buybacks. If the shares then reflect the roughly $280 central value, dividends and buybacks made below that value imply a return of about 20% a year from $180, and the weaker outcomes we consider imply values above that price.