Verisk: down 28%, but not yet cheap

Verisk is the kind of business that rarely goes on sale: 83% of its revenue is subscriptions, its adjusted EBITDA margin is 57% and its return on capital is about 32%. It is down 28% from its high of $248.22. But the company also spent $1.7 billion on buybacks in the first half at an average price near $200, using cash and debt, and the stock still trades about where my Base case says it should.

At the October 9, 2026 close of $177.63, the shares trade 6.7% above my triangulated Base value of $166.43. The Bear value is $103.72 and the Bull value is $209.49. The required margin of safety (MOS), the discount to fair value I need before buying, is 45%, so the maximum buy price is $91.54, 48% below the price.

Route: SF-01 standard value. SF-07 does not apply: Verisk is a profitable, cash-generating data and analytics company, not an ARR software story. Economic classification: Quality Growth, 7/12 on the sector scorecard, so the engine is GROWTH_ADJUSTED_MULTIPLE with an explicit multiple build. Primary method: normalized owner free cash flow times a multiple. Secondary method: a six-year DCF at 10%. Values are present fair-value estimates, not 12-month price targets. Owner free cash flow here is cash after interest, taxes, capex (including capitalized software) and stock pay.

What Verisk is

Verisk sells data, models and analytics to insurers. Its underwriting segment (about 70% of revenue) supplies the industry's standard forms, rules, loss costs, catastrophe models and risk data. Its claims segment supplies anti-fraud analytics, estimating tools and property restoration solutions.

• Subscriptions are 83% of revenue and grew 8% organically in the second quarter. Transactional revenue is 17% and fell 4.2%.
• Insurers rely on Verisk's loss data and standard forms. That dependence is the moat, and it is why management sees AI as additive.
• Management retired 8.5 million shares in the first half, so the share count fell 3.6% in a year, and shareholders' equity is now negative because buybacks exceeded retained earnings.

The latest quarter

Organic constant-currency revenue growth was 5.8% in the quarter, up from earlier in the year. The drop in GAAP net income came from a higher tax rate, higher net interest expense and legal fees tied to ongoing litigation. Free cash flow rose 58%, which I read as partly working-capital timing, so I do not extrapolate it. For the half, revenue was $1,589 million (+4.1%) and adjusted EPS $3.80 (+5.0%).

Management reaffirmed guidance, expects gradual improvement through the second half and noted a slower start to the hurricane season, which can reduce transactional revenue. A federal contract work stoppage was lifted at the start of the third quarter for a one-year term.

Over four years revenue grew 5.7% a year and free cash flow per share 11.9% a year, because the share count fell from 161.7 million to 138.4 million by 2025. The per-share growth is therefore partly financial engineering: total debt including leases rose from $3.1 billion in 2023 to $4.6 billion now, and cash fell, as buybacks continued. The 2022 operating margin of 56.4% is out of line with every other year and probably reflects one-time items (I did not verify the cause), so it is not a run-rate.

Owner free cash flow

My Base for 2026E is $1.17 billion, which is 36.5% of the revenue midpoint of $3.215 billion. This is a little below the trailing 37.6%, because cash flow timing helped the last twelve months and guidance includes $260 million to $280 million of capex and $190 million to $200 million of net interest.

Base assumes organic growth of about 5% to 5.5% a year, below the 5.8% reported in the second quarter and below the historical 6.8% to 7.3%, with owner cash flow margin edging up to 37%. I assume no further buybacks, so the unit count stays at 131.0 million, and I assume no acquisitions beyond small tuck-ins.

Debt and shares

Net debt is 2.2 times 2026E adjusted EBITDA, in the middle of management's target of 2 to 3 times. Operating leases are left out because rent is in EBITDA. Net debt rose from $2.6 billion at the end of 2025 to $3.9 billion in six months, as the company used cash on hand to buy back $1.5 billion of stock in the first quarter and a further $200 million through an accelerated repurchase. About $800 million remains authorized. The sources give $1.9 billion of year-to-date repurchases in one place and $1.7 billion in another, and I use $1.7 billion. The shares bought back averaged about $200 (CALCULATED from the $1.7 billion and 8.5 million shares), above today's $177.63 and above my Base value, so the repurchases did not add value at the price paid. The $2.00 annual dividend is a 1.1% yield and 31% of earnings.

Primary valuation: owner free cash flow times a multiple

Verisk is a Quality Growth company under the framework, so the multiple is built from a base and explicit premiums. The base is 12 times, between the average-stable and very-stable tiers. I add premiums for growth per unit, a high return on capital (about 32%) and the share of subscription revenue, and subtract a risk discount for AI, litigation, soft commercial property volumes and the planned departure of the chief information officer.

The Base multiple of 18.5 times is 54% above the 12-times sector base, below the 75% line that triggers a mandatory DCF cross-check, though I run one as the secondary method. It is also well below the multiples Verisk has traded at in recent years (ESTIMATED from public data, not verified here), because those multiples paid for faster growth and a lower-rate environment than I assume.

The primary Base value is $176.65. The market price of $177.63 is within 1% of this number.

Sensitivity: value per unit

Secondary valuation: six-year DCF

The secondary Base value is $156.23. Nearly three quarters of the DCF is terminal value, and the DCF uses the same cash flow path as the primary method, so it is not fully independent. It differs by assuming a 3.5% growth rate forever and a 10% discount rate instead of an 18.5-times multiple.

History and growth diagnostics

Valuation history. The enterprise value is 15.0 times 2026E adjusted EBITDA and the price is 18.8 times trailing free cash flow (19.7 times after stock pay). Verisk has traded at higher multiples in recent years (ESTIMATED from public data, not verified here), so today's price is below its own history, but the history belonged to a period of lower rates and faster per-share growth.

PEG (SM-19, diagnostic only). The price is 21.7 times forward adjusted earnings and adjusted EPS grew 5.0% in the first half, so the PEG is about 4, high on its face. I treat it as a diagnostic only. Growth type: STRUCTURAL for subscriptions, with ACQUISITION_AFFECTED flags from a divestiture at the end of 2025 that the sources do not name.

Free cash flow growth quality (SM-20, diagnostic only). Free cash flow rose from $887 million in 2021 to $1,192 million in 2025, a compound growth of 7.7% a year, and $1,237 million in the last twelve months. Per share it grew 11.9% a year because the share count fell 14% over four years. FCF margin rose from 36.0% to 38.8%. Forward, my Base owner cash flow compounds at 5.5% a year from 2026E to 2031E, from $1.17 billion to $1.53 billion. Forward owner cash flow yield on market value is 5.0% for 2026E and 5.3% for 2027E. Flags: WORKING_CAPITAL_DISTORTION (second-quarter free cash flow rose 58%, which I read as partly timing) and per-share growth that is helped by debt-funded buybacks. Forecast reliability: MEDIUM. Stress test: 2028E owner cash flow 20% lower gives a primary value of $143.50, and 40% lower gives $110.36.

Valuation triangulation

The two methods differ by 12.3% of their midpoint, inside the 20% line, so no divergence penalty applies. I weight them 50/50 because they use one cash flow path and the multiple and terminal growth are judgments. The bias table at the end shows 70/30 and 30/70.

What is the market already pricing in?

Reverse valuation: at $177.63, the market value of $23.3 billion implies a multiple of 18.6 times on my Base 2028E owner cash flow, against my 18.5 times. Holding my multiple, the price needs 2028E cash flow of $1.33 billion, against $1.32 billion in Base. The market is paying for about 5% growth, a stable margin and no AI disruption. It is not pricing in a Bear case, and that is why a 28% drop does not make it cheap.

Bear, base and bull scenarios

Bear: insurers slow spending as soft markets continue, AI-based competitors and clients' own models take some share, pricing power weakens and the owner cash flow margin falls to 32.5%. The multiple falls to 11 times. Base: subscriptions keep growing 7% to 8%, transactional revenue recovers, margins hold. Bull: pricing power and new products lift growth to 7% to 7.5%, margins reach 39% and the multiple rises to 23 times. Bear and Bull use the same discount rate and net debt as Base, so each risk is counted once. The price of $177.63 is 71% above Bear and 18% below Bull, which is the wrong asymmetry for a buyer.

How the buy price is set

Confidence is C, which starts the margin of safety at 40%. I add 5 points for structural disruption (AI and in-house data models) and 5 for terminal value dependence, and subtract 5 for subscription revenue and returns on capital. The result is 45% and the maximum buy price is $166.43 x 55% = $91.54.

Even with the lightest margin of safety in the Confidence C range, the buy price is $108.18, 39% below the price. Verisk would need a much lower price, or a proof of stronger growth, before the framework says buy.

Dividend and capital return

The quarterly dividend is $0.50, up from $0.45, a 1.1% yield with a payout of about 31% of earnings. Buybacks are much larger, $2.2 billion in the last twelve months, funded with cash and new debt. I do not count future buybacks in value per unit.

Reasons to own VRSK

• Subscription base. 83% of revenue is subscriptions that grew 8% organically, with strong price realization in renewals.
• Margins and returns. Adjusted EBITDA margin of 56% to 57% and ROIC of about 32%.
• Growth is re-accelerating. Organic constant-currency growth was 5.8% in the second quarter and management expects gradual improvement.
• Cash conversion. Free cash flow of $1.24 billion over twelve months is 37.6% of revenue after stock pay.
• AI as a product. XactAI licensees rose nearly tenfold since March to about 7,000, though management says AI monetization is early and hard to quantify.

What could go wrong?

• AI and in-house models. Carriers building their own large language models could pay less for third-party data, which is the main long-run risk.
• Soft commercial property market. Clients quote fewer opportunities and spend less on underwriting data, which hurts transactional revenue.
• Leverage and buybacks. Net debt rose by $1.4 billion in six months and equity is negative. Buying back $1.7 billion at about $200 a share has not helped returns at today's price.
• Litigation. Legal fees tied to ongoing litigation reduced net income, and I have no estimate of the outcome.
• Execution. The chief information officer is leaving and the chief technology officer is interim CIO.

Management execution

Management has delivered steady margins and organic growth near 6%, and it keeps leverage inside a stated target of 2 to 3 times. The capital allocation record is mixed: large buybacks at an average price above today's, a leverage increase to fund them, and an unnamed divestiture at the end of 2025. Management credibility scores 14/20.

Stock Analyza scorecard

Business quality is high because of the moat, returns on capital and cash conversion. The balance sheet score is held down by the move to negative equity and higher debt.

The valuation score is low because the price is above Base and 71% above Bear, even though the methods agree.

Confidence score

Confidence is C. Data quality is 3/5 because the sources disagree on buyback totals, intangibles fell by $541 million in the balance sheet without an explanation in my sources, and I did not read the quarterly filing in full. Accounting transparency is 3/5 because of adjusted measures and the negative equity. Valuation robustness is 3/5 because three quarters of the DCF is terminal value.

What I would watch from here

Next review: the third quarter report, expected around late October 2026 (date not confirmed).

The three most important thesis breakers

1. Subscription growth slows below 5% organically, which would show that AI or in-house models are taking demand.
2. The owner cash flow margin falls below 34% without a clear investment case.
3. Leverage rises above 3 times EBITDA to fund more buybacks, or litigation costs prove large.

Adversarial review

I recomputed every table from the listed inputs and the arithmetic reconciles. Revenue, EBITDA, guidance, debt and cash flow come from the second quarter release and call. The margin path, the multiples, the discount rate and the dilutive shares are my estimates. The strongest bullish counter-thesis is that Verisk is a toll road on insurance data with 83% subscriptions, and that its multiple de-rated for AI fears that management has answered with products, so the Bull value of $209.49 is a fair outcome if growth returns to 7%. The strongest bearish counter-thesis is that carriers and AI firms will build alternatives to Verisk's data, which would cut price realization, and that debt-funded buybacks at $200 have already used up the balance sheet's flexibility.

The choices that lean generous are an 18.5 times multiple, 5% growth and a 37% cash margin. The ones that lean conservative are deducting stock pay, normalizing second-quarter cash flow timing and leaving out future buybacks. I did not tune any of them toward a verdict. Publication gate: PASS_WITH_WARNING, because the sources conflict on the buyback total, the divestiture is unnamed and the intangibles decline is unexplained.

Not available or estimated. Unavailable: the name and size of the 2025 divestiture, the reason intangibles fell, litigation exposure, a verified peer table, and the full quarterly filing. Estimated: dilutive awards, the forward paths, the multiples, the discount rate and the historical multiples.

Final verdict: hold at $177.63

Verisk is a very good business at a fair price. The shares are 6.7% above my Base value, within 1% of my primary value, and the price has already absorbed the 28% drop. I would want a price between about $108 and $92, or clear evidence that growth is above 6%, before treating it as a buy.

VERDICT: HOLD at $177.63; Triangulated Base $166.43; downside to Base 6.3%; required MOS 45%; maximum buy price $91.54. The main thesis breaker is subscription growth that slows below 5% as AI and in-house models replace third-party data.

Sources

Primary sources include Verisk's second quarter 2026 results (revenue, adjusted EBITDA, guidance, balance sheet), the second quarter earnings call transcript and call highlights, and the financial statement history and statistics on stockanalysis.com.

The reference market price is the October 9, 2026 close of $177.63. Balance sheet items are as of June 30, 2026. Third quarter results were not yet available. The owner free cash flow path, the multiples, the discount rate and terminal growth are analytical estimates, not company guidance.