VRSK — Verisk Analytics, Inc.

As of 2026-09-13 the tape and the sell side are both negative: the stock is -5.2% over one week, -12.4% over six months, and sits below every major moving average, and on 2026-09-08 Barclays kept its Overweight rating while cutting its price target from $275 to $225. Of 25 analyst ratings, 15 are Hold (consensus rating: Hold), and Q2 2026 GAAP EPS fell 9.8% YoY to $1.75 from $1.81 on divestiture noise, higher interest expense, and legal fees. Importantly, the bear case is not about current results - adjusted EPS of $1.98 beat the ~$1.93-1.94 consensus and organic revenue grew 5.8% - it is about what AI-era competition and higher-for-longer rates could do to a premium-multiple data vendor, so the market is compressing the multiple rather than cutting estimates.

Revenue has compounded steadily but slowly: $677.6M in Q3 2023 to $806.3M in Q2 2026, with YoY growth cooling from ~5.9% in Q3/Q4 2025 to 3.9% in Q1 2026 and 4.3% reported (5.8% organic constant currency) in Q2 2026 - deceleration, but modest and partly a reported-figures effect after divestitures. Margins are structurally better than three years ago: operating margin was 41.5% in Q3 2023 and 37.0% in Q4 2023 (a noisy quarter), versus 45.1% in Q2 2026 and 45.0% in Q1 2026; the latest quarter is roughly flat to slightly down YoY (45.9% in Q2 2025). GAAP earnings are the weak spot: Q2 2026 net income of $228.6M and EPS of $1.75 versus $253.3M and $1.81 a year earlier, and TTM net income of ~$886M against TTM free cash flow of $1,236.5M - cash conversion is excellent (accruals ratio -3.06%), so the GAAP softness is more about interest, legal, and divestiture items than about the operating engine. Capex has ticked up to ~$68M/quarter (8.4% of revenue) from ~$55M in 2023-24. The balance sheet is the real change: long-term debt rose from $3.35B (2025-12-31) to $4.22B (2026-06-30), cash fell from $2.18B to $551M, and stockholders' equity went to -$1.19B. The arithmetic implies roughly $1.8B of buybacks in H1 2026 (equity fell $1.5B against $463M of net income), well beyond the $200M ASR the news flow cites - worth verifying in the 10-Q. Net: stable-to-slowly-growing top line, expanded margins, cash-backed earnings, and a materially more levered, negative-equity capital structure.

Cheap versus its own history, and reasonable in absolute terms. As of 2026-09-13, TTM P/E is 27.0x against a 5-year average of 43.4x, and P/S is 7.31x against a 5-year average of 12.59x - the deterministic screen labels this 'Cheap vs. its own history' at a 98.6 context score. On next-FY EPS of $7.712 the forward P/E is ~22.9x, TTM FCF yield is 5.4% ($1.24B on a $22.9B market cap), and EV/EBITDA is 16.3x. What you are paying for: 45% operating margins, 19.6% ROIC, near-100% subscription/recurring mix, a 0.66 beta, and a 1.05% dividend. What justifies the discount to history: 4-6% organic growth rather than the high-single-digit of the zero-rate era, higher rates, AI-disruption anxiety on data moats, and a balance sheet that is no longer a fortress. Note the reported PEG of 0.27 is a data artifact (it implies ~100% earnings growth) and should not be used. On a growth-adjusted basis, ~23x forward for a mid-single-digit grower with high-20s/30s incremental margins is fair-to-cheap, not expensive.

This is a multiple de-rating, not an earnings-miss story: adjusted EPS beat in Q2 2026 and full-year guidance was reaffirmed, yet the stock fell 12.4% in six months and 5.2% in the week into 2026-09-13. That tells you the market is discounting future risk (AI-driven pricing pressure, higher rates, more leverage) rather than current deterioration. Practical read: if Q3 2026 (late Oct/early Nov) confirms ~5-6% organic growth and stable margins, the ~23x forward multiple has room to re-rate toward the high-20s, which is where my fair-value work lands. If instead organic growth slips below 4%, transactional/consulting revenue rolls over, or the company keeps borrowing to buy back stock, the market is effectively telling you the sustainable multiple is high-teens and the stock is not yet cheap. Watch the gap between GAAP EPS (falling) and adjusted EPS/FCF (rising) - as long as FCF stays above net income, the GAAP softness is noise; if that relationship inverts, believe the GAAP number.

As of 2026-09-13, the technical picture is a confirmed downtrend with no bottoming signal: RSI(14) at 38.52, price 4.88% below the 20-day SMA, 7.35% below the 50-day, and 9.0% below the 200-day, with a -5.16% one-week and -12.43% six-month return. RSI is weak but not oversold, so there is no capitulation washout to buy; the 52-week low of $155.94 sits ~11.5% below the current price and is the logical magnet if the decline continues. Entry-timing implication: this is a scale-in-on-weakness setup, not a chase-the-bounce setup. Accumulating toward the high-$150s/low-$160s with a reclaim of the 50-day SMA as the confirmation signal is a more disciplined approach than buying strength above $180 before the trend flattens.

• Higher-for-longer interest rates: the same rate environment that helped compress the P/E from a 43.4x five-year average to 27.0x also raises Verisk's cost of servicing $4.22B of long-term debt (up from $3.35B at 2025-12-31), directly pressuring GAAP EPS.
• P&C insurance industry health: a $31.7B net underwriting gain for U.S. P&C insurers in H1 2026 means healthy IT/analytics budgets for renewals, though a benign catastrophe period can soften pricing discipline and reduce transactional and consulting spend.
• Rising catastrophe loss trend: Verisk's updated benchmark of $171B in average annual insured cat losses (+$19B) is a structural tailwind for risk-modeling demand even if near-term storm activity is quiet.
• AI disruption narrative across data and analytics vendors: this is the single biggest driver of multiple compression, and it cuts both ways - it pressures the perceived moat of legacy data businesses while creating product opportunity (Verisk Fraud Discovery, U.S. Data Center Exposure Database, XactAI).
• Sector rotation and multiple discipline: even with a 0.66 beta, VRSK has been treated as an expensive tech-adjacent name and sold as capital rotated toward cheaper cyclicals/value; the -9.0% gap to its 200-day SMA is a rotation symptom, not a company-specific one.

• Q3 2026 earnings (typically late October/early November 2026): subscription renewal rates, transactional revenue recovery, and full-year guidance confirmation are the three numbers that decide whether the de-rating stops.
• Commercial traction for newer AI/data products - Verisk Fraud Discovery, the U.S. Data Center Exposure Database, and XactAI integrations - would directly rebut the AI-disruption thesis.
• Continued capital return: the dividend ($0.50/quarter, ~1.05% yield) plus buybacks are shrinking the share count (market cap $22.9B on ~130M shares), which can lift EPS even with 4-5% revenue growth - provided it is not funded with more debt.
• Another sell-side re-rating: after Barclays cut its target to $225 on 2026-09-08, the $225.43 consensus target and the $185 low end of the target range frame how much of the gap between price ($176.20) and fair value can close on a single clean quarter.

• Balance sheet deterioration: long-term debt up to $4.22B and cash down to $551M at 2026-06-30, with stockholders' equity at -$1.19B and a current ratio of 1.005. Net debt/EBITDA is roughly 2.2x - manageable for a subscription business, but the flexibility that historically justified a premium multiple is gone.
• Buyback-funded leverage: the H1 2026 equity decline implies ~$1.8B of repurchases against $463M of net income, far above the $200M ASR cited publicly. Buying back stock at depressed prices may prove accretive, but doing so with new debt at high rates adds risk if growth slows.
• Growth deceleration: organic growth has cooled from ~5.9% YoY (Q3/Q4 2025) to 4.3% reported in Q2 2026; a further slip below 4% would undermine the case for any multiple recovery.
• AI/LLM disruption of proprietary data moats could erode pricing power over time - the market is already pricing some probability of this, and a single major insurer building in-house alternatives would crystallize the fear.
• GAAP-vs-adjusted divergence: Q2 2026 GAAP EPS fell 9.8% YoY to $1.75 while adjusted EPS rose to $1.98, with divestiture-related items, legal fees, and interest expense doing the damage. If those items become permanent rather than one-off, the earnings base is lower than the adjusted numbers imply.

Verisk looks like a genuinely good business trading at a discount to its own history for reasons that are partly structural (higher rates, AI anxiety, slower growth) and partly fixable (GAAP noise from divestitures, interest, and legal costs). At $176.20 as of 2026-09-13 you are paying ~23x forward earnings and 5.4% trailing FCF yield for 45% operating margins, ~20% ROIC, and cash-backed earnings - that is not expensive, and my own fair-value work lands near $200. But the technicals are still pointed down (RSI 38.5, below all major moving averages, no capitulation), leverage has risen materially, and there is no evidence yet of a growth inflection that would justify re-rating to the mid-30s multiples this stock used to command. Net: own it, don't chase it. Accumulate toward the high-$150s/low-$160s and treat the Q3 2026 report as the referendum on whether 4-6% organic growth is durable.

Primary method: 26x forward EPS. Analyst next-FY EPS is $7.712, and 26x is my justified multiple for a capital-light, 45%-operating-margin, ~20%-ROIC, low-beta (0.66) compounder growing revenue 4-6% organically and EPS faster via buybacks - deliberately below the 43.4x five-year average (a zero-rate artifact) and below the 27.0x TTM, but above the current ~22.9x forward. 26 x $7.712 = $200.51. Cross-checks bracket it: a 4.5% target FCF yield on $1,236.5M TTM FCF gives ~$211/share, and 17x EV/EBITDA on ~$1.63B TTM EBITDA net of ~$3.66B net debt gives ~$185/share (midpoint ~$198), so $200.51 fits inside both. It implies ~14% upside from the 2026-09-13 price of $176.20 and sits ~11% below the $225.43 analyst consensus target.