Financebotresearch desk研究台

TRI › analyze

TRI · Analyze

Financial Services

Date: 2026-07-23 Price: $85.77 (close, Jul 23 2026) | Day: -1.73% | Volume: 825K 52-Wk Range: $76.28 (Jun 22 2026) – ~$222 (Jul 2025) Market Cap: ~$36.9B (~430M shares post-consolidation) | EV: ~$39.1B Sector: Industrials / Specialty Business Services (classified here under Financial Data & Analytics) Exchange: NASDAQ Global Select + TSX | HQ: Toronto | CEO: Steve Hasker | Employees: 27,100 Controlled by: The Woodbridge Company (Thomson family), ~68%

Prior analysis: analyze-2026-04-26.md — price $89.79, conviction [7.5], fair value $120–150. Since then: Q1 beat + guidance reaffirmed, dividend +10%, special dividend, Global Print JV'd to KKR, and a violent round trip ($76 → $101 → $86). Thesis intact; fair value revised down to $95–120 on sharper workflow risk.


⚠️ Data Integrity Note — Read First

Yahoo Finance MCP was not connected for this run, and roic.ai is on the free plan (2 years of history only). Consequences:

Framework requirement Status
FCF 5–8yr CAGR Could not compute — only FY2024/FY2025 available
Revenue / net income CAGR (5-8yr) ⚠️ Partial — 2021–2025 from third-party (stockanalysis.com), unverified against filings
ROIC trend (multi-year) ❌ Only FY2025 ROIC available (10.47%); FY2024 ROIC not returned
Dividend 8–10yr CAGR ⚠️ Estimated from 3 years + reported 33-yr increase streak
Institutional / insider positioning ❌ Unavailable (no get_holder_info)
Gross margin ⚠️ Two conflicting definitions — roic.ai 75.8%, stockanalysis 38.8%. Different COGS treatment. Trend direction agrees (rising); absolute level is unreliable.

Everything below flagged with ⚠️ carries lower confidence. Reconnect Yahoo Finance MCP and re-run before sizing a large position.


Manager's Verdict

The market has repriced TRI from a compounder to a melting ice cube. The financials say the opposite is happening.

TRI is -61% from its July 2025 peak and -36% YTD, while organic growth accelerated from 7% to 8%, EBITDA margins are guided up 100bps to ~40%, the dividend was raised 10%, and GenAI products now make up 30% of Legal annual contract value. Reported results and the share price are pointing in opposite directions.

At $85.77 the stock offers a 5.7% FCF yield on company-guided FY26 FCF of $2.1B, a 3.05% dividend yield (roughly 2x its own 5-year average) growing ~10%/yr from a ~50% payout, on a balance sheet carrying 0.60x net debt/EBITDA. A reverse-DCF says the market is pricing ~3.3% perpetual FCF growth into a business currently compounding at 8–9%.

But the bear case is not stupid, and it is not resolved. The threat was never "will law firms cancel Westlaw" — it's who owns the interface. TRI's own May 2026 MCP integration puts Claude at the front and CoCounsel at the back. That is excellent for retention and potentially terrible for pricing power. Harvey went from ~$100M to $190M ARR in five months. If TRI's economics shift from $30K/seat platform to per-call data vendor, revenue per customer falls even with usage flat.

The most likely outcome is neither the 2025 multiple nor the current one: TRI keeps the content economics, cedes some interface premium, and compounds FCF at 6–7% instead of 9–10%. That business is worth ~$95–120, not $222 and not $86.

Rating: BUY — accumulate in tranches | Conviction: [7.5] (unchanged) Fair value range: $95–120 (base) | Bear: $68–78 | Bull: $140–160 Add zone: $82–98 | Trim zone: $150+

Conviction held rather than raised: the bull case got confirmed (Q1 organic 8%, guidance reaffirmed, dividend +10%) but the bear case got sharper (Harvey ARR nearly doubled; TRI conceded the front-end). Those roughly cancel.


Phase 1 — Fundamentals Analyst

Income Statement

Metric FY2024 FY2025 Δ
Revenue $7,258M $7,476M +3.0% (reported)
Organic growth 7% 7%
Gross profit $5,467M $5,667M +3.7%
Gross margin ⚠️ 75.32% 75.80% +48bps
Operating income $1,896M $1,989M +4.9%
Operating margin 26.12% 26.61% +49bps
EBITDA $2,718M $2,919M +7.4%
EBITDA margin 37.45% 39.04% +159bps
Net income $2,210M $1,502M -32.0%
Diluted EPS (GAAP) $4.96 $3.38 -31.8%
Effective tax rate negative 21.9% normalizing

On the -32% net income "collapse": it is an artifact, not deterioration. FY2024 carried a -$123M income tax benefit (negative tax expense) that inflated GAAP net income by roughly $0.60–0.70/share. FY2025 paid a normal 21.9% rate. Strip the tax noise and look at the operating line: EBITDA +7.4% with margins up 159bps. The business got better; the tax line got normal.

Reported revenue understates the business. Reported +3.0% vs organic +7% — the gap is divestitures (Regulatory Intelligence, Oden sold 2025) and FX. This is deliberate portfolio pruning, not lost ground.

Free Cash Flow — ⚠️ Definitional Correction

roic.ai reports free_cash_flow = $2,651M for FY2025, which is simply cash from operations with no capex deducted. TRI books capitalized software/capex inside "other investing activities" (-$695M). Corrected:

FY2024 FY2025 FY2026E (guided)
Cash from operations $2,457M $2,651M
Less: capex & cap. software ~$878M ~$695M
Free cash flow (corrected) ~$1,579M ~$1,956M ~$2,100M
FCF/share (~430M sh) ~$3.56 ~$4.41 ~$4.88
FCF growth +23.9% +7.4%

FCF grew 24% in FY2025 and is guided to $2.1B in FY2026. ❌ The framework's 5–8yr FCF CAGR is not computable with the data available — the prior April analysis cited a 15% 3-yr CAGR, which is directionally consistent but I could not independently verify it this run.

Cash conversion is excellent: CFO/net income = 1.76x (FY2025).

Capital Allocation — FY2025

Use Amount % of FCF (~$1,956M)
Dividends $1,039M 53%
Buybacks / return of capital $1,000M 51%
Net M&A (–$843M acq. +$254M divest.) $589M 30%
Debt paydown $1,063M 54%
Total deployed $3,691M 189%

🚩 Flag: FY2025 deployment was ~189% of FCF. Shareholder returns alone ($2,039M) exceeded FCF (~$1,956M). This was funded by drawing down the cash pile from the 2024 Financial & Risk-era divestitures — cash fell from $1,888M to $465M (-75%). The 2026 special dividend ($1.44/sh, ~$620M) continues the pattern.

This is not a red flag on its own — returning proceeds from asset sales is correct capital allocation, and management has been explicit about it. But it is not repeatable indefinitely. The dry powder is now spent. From here, shareholder returns must be funded from the ~$2.1B FCF run-rate, which comfortably covers the ~$1.15B dividend but leaves less room for $1B/yr of buybacks plus M&A. Watch the FY2027 capital return plan.

Balance Sheet — Fortress

Metric FY2024 FY2025 Read
Total debt $3,076M $2,372M ↓ -23%
Cash & equivalents $1,888M $465M ↓ -75% (deployed)
Net debt $967M $1,752M ↑ but trivial
Total debt / EBITDA 1.13x 0.81x 🟢 Excellent
Net debt / EBITDA 0.36x 0.60x 🟢 Excellent
Debt / assets 16.68% 13.22% 🟢 Falling
EBITDA / interest 13.5x 15.9x 🟢
EBIT / interest 9.4x 10.9x 🟢
Equity / assets 63.2% 63.0% 🟢
Current ratio 1.02 0.64 🟡 See note

Current ratio 0.64 is not a liquidity problem. $1,251M of current liabilities is deferred revenue — cash already collected for services not yet delivered. For a subscription business that is a source of funding, not an obligation requiring cash. Excluding deferred revenue, current ratio is ~1.00.

🚩 Real flag — tangible book value is negative: -$1.94/share. Goodwill ($7,913M) + intangibles ($4,747M) = $12,660M, or 71% of total assets ($17,940M). This is structurally normal for a 150-year serial acquirer of information assets, but it has two consequences: (1) Graham's formula is unusable here, and (2) any impairment of the acquired content/technology assets would hit book equity hard. There is no impairment signal today — margins are expanding — but it's the balance-sheet expression of the AI risk.

Share Count — Consistent Shrinkage 🟢

Year Diluted shares Δ
2022 477M
2023 457M -4.2%
2024 444M -2.8%
2025 443M -0.3%
2026E ~430M -2.9% (share consolidation)

-9.8% over four years, ~-2.5%/yr. Real per-share tailwind. The unusual 2026 mechanic — a $1.44 special dividend paired with a 0.98456 share consolidation — is a tax-efficient Canadian return-of-capital structure, not financial engineering to flatter EPS. It confused some retail commentators; it shouldn't concern us.

Per-Share Metrics — The Owner's View

Metric FY2024 FY2025 Δ
Revenue/share $16.36 $16.91 +3.4%
EBITDA/share $6.13 $6.60 +7.8%
Operating income/share $4.27 $4.50 +5.3%
FCF/share (corrected) ~$3.56 ~$4.41 +23.9%
Book value/share $26.27 $25.56 -2.7%
Tangible BVPS $0.11 -$1.94 🚩 negative
Dividend/share $2.14 $2.35 +9.9%

Every operating per-share metric is rising faster than the headline revenue line — buybacks doing their job on top of margin expansion.

Fundamentals Scorecard

Dimension Grade Note
FCF generation & growth 🟢 A +24% FY25, $2.1B guided FY26, 1.76x conversion
Margins & operating leverage 🟢 A EBITDA margin +159bps, guided to ~40%
Balance sheet 🟢 A 0.60x net leverage, 15.9x coverage
Share count 🟢 A- -2.5%/yr consistent
Capital allocation 🟡 B Correct, but 189% of FCF in FY25 — cash cushion now spent
Revenue growth (reported) 🟡 B- +3.0% reported vs +7% organic; divestiture drag
Return on capital 🟡 B- ROIC 10.5% — modest for a "wide moat" (see Moat)
Book quality 🟠 C Negative tangible book, 71% goodwill+intangibles
Data completeness 🔴 D Multi-year CAGRs uncomputable this run

Phase 1 — Sentiment & Intelligence Analyst

The Price Action Tells the Story

Date Level Event
Jul 2025 ~$222 All-time high
Dec 31 2025 $134.03 -40% from peak
May 11–12 2026 $89.22 → $87.30 -6.1% on Anthropic MCP announcement
Jun 01 2026 $93.96 +8.6% relief bounce
Jun 02 2026 $86.33 -8.1% — whipsaw
Jun 22 2026 $76.28 52-week low
Jul 17 2026 $100.85 intraday +32% off the low in 3 weeks
Jul 23 2026 $85.77 Round trip complete

-61.4% from peak. -36.0% YTD. Currently +12.4% off the June low.

This is not orderly de-rating — it is a narrative fight playing out in 8% daily swings on a $37B Canadian blue chip with 33 years of dividend increases. Realized volatility here is behaving like a small-cap. That itself is information: the market has no settled view, which means the range of outcomes being priced is genuinely wide.

Q1 2026 Results (reported May 5, 2026)

Metric Q1 2026 vs PY
Revenue $2,087M +10%
Organic growth 8% vs 7% FY2025
"Big Three" organic (Legal, Corporates, Tax) 9% accelerating
Legal Professionals organic ~9% upper end of 8–9% framework
Adjusted EBITDA $881M +9%
Adjusted EBITDA margin 42.2% expanding
Adjusted EPS $1.23 +10%
Free cash flow $332M +19%
Net income $459M

FY2026 guidance — reaffirmed, not cut: - Total organic revenue growth 7.5–8% - Big Three organic ~9.5% with 100bps margin expansion - Adjusted EBITDA margin ~40% (+100bps YoY) - Free cash flow ~$2.1B - Q2 2026 organic 7–8%, EBITDA margin ~38%

The single most important disclosure: GenAI products are now 30% of Legal annual contract value. AI is showing up in TRI's results as revenue, not attrition. That is the opposite of what a disruption victim looks like at this stage.

Recent Events — Last 90 Days

Date Event Read
Jul 14 Global Print 51% JV with KKR 🟢 Portfolio pruning — sheds a structurally declining, low-multiple segment
Jul 13 Cutting ~500 engineering roles 🟢 AI-driven efficiency; funds the 100bps margin expansion. Stock rose on it.
Jul 8 Q2 earnings set for Aug 5, 2026 📅 Near-term binary catalyst — 13 days out
Jun 22 Future of Professionals report: $143B US client revenue at risk 🟡 Double-edged — TRI's own research quantifying the disruption it's exposed to
Jun 10 Shareholder vote on ICE contracts 🟠 ESG/reputational overhang; all 14 directors re-elected
May 27 "Fiduciary-Grade AI" standard launched 🟢 Explicit moat positioning: accuracy + accountability + auditability
May 13 Sterne Kessler co-developed patent AI tool in CoCounsel 🟢 First attorney-built workflow — ecosystem forming
May 12 Anthropic MCP integration — Claude ↔ CoCounsel 🔴🟢 The pivotal event. Stock -6.1%. See debate below.
Apr 20 CoCounsel Legal Reimagined beta — agentic workflows 🟢 Product response is real, not vaporware

Analyst Positioning — Wide and Confused

Source Consensus Target
22 analysts (Investing.com) $90.37
18 analysts (stockanalysis) 13 Buy / 5 Hold / 0 Sell median $120.00
15 analysts (other) $169.33
Canaccord (Galappatthige) Upgraded Hold → Buy $174 (from $182)
Zacks consensus +29.8% upside
GuruFocus GF Value $196.54

The dispersion is the signal. Targets span $90 to $196 — a 2.2x spread on a mature, transparent, guidance-giving blue chip. That does not happen when analysts disagree about next quarter's margin. It happens when they disagree about whether the business model survives the decade. Zero sell ratings despite a 61% drawdown, with average targets cut 8.3% over three months: the sell-side is walking targets down without abandoning the thesis.

Sentiment Read

Cohort Stance
Sell-side 🟢 Bullish but retreating — 72%+ Buy, targets grinding lower
Market / price action 🔴 Deeply bearish — -61% peak-to-trough
Management 🟢 Confident — reaffirmed guidance, +10% dividend, special dividend, aggressive buyback
Commentary (Seeking Alpha) 🟡 Split — "Buy TRI, AI won't replace info services" vs "Key AI bear case is not resolved"
Insider / institutional Unavailable this run — data gap

Management is voting with cash. Raising the dividend 10%, paying a special dividend, and buying back 2% of shares into a -61% drawdown is not the behavior of a team that sees revenue cliff ahead. Controlled 68% by the Thomson family via Woodbridge — a long-horizon owner with no incentive to prop up a quarter. That alignment is worth something.


Phase 2 — Moat Analyst

Quantitative Base

Metric FY2024 FY2025 Trend
ROIC n/a ⚠️ 10.47% ❌ no trend
ROE 19.93% 13.05% ↓ (tax artifact)
ROA 11.91% 8.26% ↓ (tax artifact)
Return on capital n/a ⚠️ 11.49%
Gross margin ⚠️ 75.32% 75.80%
Operating margin 26.12% 26.61%
EBITDA margin 37.45% 39.04% ↑↑

The margin trend is the moat evidence, and it is unambiguously positive. Gross, operating, and EBITDA margins all expanded in FY2025, and management guides EBITDA margin to ~40% in FY2026. A company losing a pricing war does not expand gross margin. This is the single strongest rebuttal to the disruption thesis available in the financials — margin compression is the first thing you'd see, and it isn't there.

🚩 But ROIC at 10.47% is a genuine problem for the "wide moat" label. Best-in-class information-services compounders run 15–25%. TRI's is dragged down by $12.7B of goodwill and intangibles from decades of acquisitions. Ex-goodwill, returns on tangible operating capital would be far higher — but I could not compute that cleanly with the data available, and I won't fake it. What this says honestly: TRI has a strong product moat but a mediocre capital-deployment record. It has repeatedly bought growth rather than generating it organically at high incremental returns. That matters for how much you'll pay.

Moat Sources — Named and Graded

Source Strength Assessment
Intangibles (proprietary content) 🟢 Strong Westlaw headnotes, KeyCite citator, Practical Law templates. Court opinions are public domain; TRI's editorial layer is copyrighted. 150 years of accumulated curation.
Switching costs 🟢 Strong Westlaw + Practical Law + ONESOURCE are embedded in firm workflow, billing, and associate training. Multi-year enterprise contracts, ~90%+ retention historically.
Regulatory / liability lock-in 🟢 Strong Lawyers cannot file briefs on hallucinated citations. Courts have issued $145K+ in AI-hallucination sanctions (2026). Citation-grounded, auditable output is a compliance requirement, not a preference.
Efficient scale 🟡 Moderate Legal research is effectively a duopoly (TRI/Westlaw vs RELX/LexisNexis) — but AI entrants are attacking exactly this.
Cost advantage 🟡 Moderate Content costs amortize across a huge base; AI is now cutting the marginal cost of curation itself (see: 500 engineers).
Network effects 🔴 Weak Minimal. Westlaw gets no better because your rival uses it. This is the moat TRI does not have — and it's precisely the one Harvey is building via firm-level workflow data.

The Legal Wall — ROSS Intelligence

Thomson Reuters v. ROSS Intelligence (D. Del., Judge Bibas, Feb 11 2025): the court held on summary judgment that ROSS's use of Westlaw headnotes as AI training data was copyright infringement and NOT fair use — explicitly rejecting "transformative use" even at the intermediate training stage.

This is the most underappreciated asset TRI owns. It means a general-purpose model cannot legally bootstrap Westlaw-equivalent capability by training on Westlaw's editorial layer. Competitors face three options: 1. License from TRI → TRI captures the economics regardless of who owns the interface 2. Build a competing editorial corpus from scratch → expensive, slow, and precisely the 150-year moat 3. Infringe → now with adverse precedent on the books

🚩 Binary risk: ROSS has appealed to the Third Circuit — the first appellate court to rule on fair use for AI training. A reversal materially weakens this wall. No decision date known. This is a genuine, datable, uncontrollable risk to the thesis and should be sized for.

Adversarial Stress-Test

"I'm a well-funded rival — $500M and a frontier model. How do I kill Westlaw?"

The attack that works: Don't attack the content. Attack the interface. Build the agentic layer lawyers actually live in — matter management, drafting, review, client comms — and relegate TRI to a citation-verification API called in the background. Price the platform at $500/seat/month against Westlaw's effective $2–3K. Let TRI keep the content revenue while I take the workflow, the user relationship, and the expansion revenue. Then I renegotiate the content deal from a position of strength, because I own the customer.

This is exactly what Harvey ($11B valuation, $190M ARR, 100K lawyers, 1,300 firms) and Legora ($5.55B, explicitly positioned as the Westlaw/Lexis alternative) are executing. It is also, uncomfortably, what TRI's own May 2026 Anthropic MCP integration enables — Claude at the front, CoCounsel at the back.

Where the attack breaks down: - Liability. A malpractice-exposed partner will not sign off on unverified output. The verification layer is where TRI lives, and it is the part that cannot be hallucinated into existence. - Copyright. ROSS says you can't train your way to the editorial layer. You must license it. From TRI. - Tax & Accounting is a harder target than Legal. ONESOURCE, UltraTax, Checkpoint are wired into filing obligations, ERP systems, and jurisdictional rule engines with statutory deadlines. Displacing that is a multi-year enterprise migration with regulatory penalties for failure. The market is pricing Legal disruption across the whole company — but Legal is only one of the Big Three. - Distribution. TRI already sits inside virtually every AmLaw 200 firm and Big Four practice with existing procurement, security review, and budget lines.

Verdict: the moat holds on content and compliance. It is genuinely contested on interface and pricing power.

Revenue Stream Map

Segment Durability Risk / Upside
Legal Professionals 🟡 Contested Highest AI exposure and highest AI monetization (GenAI = 30% of ACV, organic ~9%). This is where the fight is.
Corporates 🟢 Durable Legal/tax/compliance/IT for in-house teams. Growing complexity is a structural tailwind.
Tax & Accounting 🟢 Most durable Statutory deadlines, jurisdictional rule engines, ERP integration. Hardest to displace. Underrated by the market.
Reuters News 🟠 Declining Structurally challenged; strategically valuable as a trust/brand asset and content source.
Global Print 🔴 Melting Being resolved — 51% sold to KKR (Jul 2026). Correct decision.

Disruption Forecast (5–10 years)

Now: Generative AI is simultaneously TRI's largest threat and its largest product cycle. Evidence to date favors monetization over cannibalization — but it is early, and contract renewals lag technology by 2–3 years. We have not yet seen a full renewal cycle post-agentic-AI.

5–10 years: The plausible endgame is that TRI becomes the trusted, liability-grade data and verification layer underneath a competitive interface market. That's a good business — but it is a lower-multiple business than owning both layers. Volume may rise (AI drives more queries, not fewer) while price per seat compresses. Net effect: 6–7% FCF growth, not 9–10%. That is the base case, and it is already close to what today's price implies.

Evergreen rating: 6.5/10 (down from 7/10 in April) The need — professionals requiring authoritative, liability-safe answers on law and tax — is permanent and growing with regulatory complexity. The delivery mechanism is in the most violent transition of the company's 150-year history. TRI owns the need. It no longer owns the mechanism uncontested. Downgraded a half-point because the MCP integration was TRI itself acknowledging the interface layer is contestable.

Moat Defense — What Management Is Doing

Action Read
$650M+ invested in AI (2025) — largest corporate investor in legal AI 🟢 Serious, not defensive theater
"Fiduciary-Grade AI" standard 🟢 Reframes the battle onto TRI's turf — accuracy, accountability, auditability
Anthropic MCP integration 🟡 Pragmatic, and a real concession. Meets lawyers where they work; also legitimizes Claude as the front-end.
CoCounsel Legal Reimagined (agentic) 🟢 Building the interface rather than ceding it
Sterne Kessler co-development 🟢 Ecosystem/network-effect seeding — the moat TRI lacks
500 engineering cuts 🟢 Using AI on its own cost base
Print → KKR, Reg. Intelligence & Oden sold 🟢 Focusing the portfolio on the Big Three

Risk/reward on the defense: good. Management is engaging rather than denying, and monetizing rather than protecting. The MCP move is the one genuinely debatable decision — it trades pricing power for retention. In a world where retention was the thing at risk, that is probably the right trade, but it caps the bull case.


Phase 2 — Valuation Analyst

Company type: mature, cash-generative dividend grower with contested growth. → Weight FCF-based and dividend models. Graham is structurally unusable.

Model 1 — Graham's Intrinsic Value ❌ Not Applicable

√(22.5 × EPS × BVPS) = √(22.5 × $5.00 × $25.56) = $53.62

Discard this. Graham's formula assumes book value approximates replacement/liquidation value. TRI's book is 71% goodwill and intangibles, with tangible BVPS of -$1.94. Applied to an asset-light IP business, Graham produces a systematically meaningless number — it would value a company whose entire worth is a 150-year copyrighted content archive at less than its two-year cash generation. Weight: 0%. Noted for completeness only.

Model 2 — Reverse DCF (What Is the Market Pricing?) — Primary

Company-guided FY2026 FCF = $2.1B; ~430M shares → FCF/share ≈ $4.88 At $85.77 → P/FCF 17.6x | FCF yield 5.69%

Solving perpetuity P = FCF / (r − g) at r = 9%:

$$g_{implied} = 0.09 - \frac{4.88}{85.77} = 0.09 - 0.0569 = \mathbf{3.3\%}$$

The market is pricing 3.3% perpetual FCF growth into a business that grew FCF 24% last year, is guided to 7.4% this year, and is compounding organic revenue at 8% with margins expanding.

Sensitivity (r = 9%):

Perpetual FCF growth Fair value Implied view
2.0% $70 Bear — real decline, pricing compression
3.3% $86 ← today's price
4.0% $98 Modest disruption
5.0% $122 Base — interface ceded, content holds
6.0% $163 Bull — moat holds fully
7.0% $244 2025's multiple

Model 3 — Dividend Yield Theory 🟢 Strong Signal

FY2026 dividend $2.62/share (+10%). At $85.77 → yield 3.055%.

TRI traded at $150–222 in 2024–25 against $2.14–2.35 dividends → historical yield band ~1.1–1.5%.

Yield assumption Implied fair value
1.5% (historical bull) $175
2.0% (historical mid) $131
2.5% (conservative re-rate) $105
3.0% (permanent de-rate) $87

DYT fair value: $105–131. The current yield is roughly 2x its own multi-year average — the highest in over a decade. ⚠️ Caveat that matters: DYT assumes the multiple compression is cyclical/sentiment-driven and mean-reverts. Here it is thesis-driven. If the market is right that TRI's terminal growth is permanently lower, the yield band has legitimately reset and DYT is anchoring to a world that no longer exists. Weight: 20%, not 40%.

Model 4 — Dividend Discount Model (Gordon Growth)

D₁ = $2.62 × 1.09 = $2.86

Discount rate (r) Growth (g) Fair value
9.5% 6.0% $82
9.0% 6.0% $95
9.0% 6.5% $114
9.5% 6.5% $95
9.0% 7.0% $143

DDM range: $82–114, midpoint ~$95. Highly sensitive to the r−g spread. Supporting facts: dividend CAGR has run ~10% (2024 $2.14 → 2025 $2.35 → 2026 $2.62), payout is only ~53% of FCF, 33-year increase streak, and net leverage is 0.60x. The dividend has substantial room to keep growing even if revenue growth halves.

Model 5 — Bogle Expected Return

Expected return = dividend yield + earnings growth ± P/E change

Scenario Yield Growth P/E change Annualized
Bear — further de-rate to 13x 3.1% 4% -5.2% ~1.9%
Base — multiple flat at ~17.6x 3.1% 8% 0% ~11.1%
Re-rate to 22x over 5 yrs 3.1% 8% +4.6% ~15.7%
Re-rate to 26x over 5 yrs 3.1% 9% +8.1% ~20.2%

Base case ~11% annualized with the multiple never recovering. That's the key asymmetry: you don't need a re-rating to earn a good return — you get paid double digits just for the business doing what it's already doing.

Supporting Multiples

Metric FY2024 FY2025 YE Today Compression
P/FCF (corrected) ~23x ~28x 17.6x -37%
EV/EBITDA 27.1x 21.0x ~13.0x -52%
P/Sales 9.96x 7.92x ~4.9x -51%
P/Book 6.20x 5.24x ~3.4x -45%
Dividend yield ~1.3% 1.75% 3.06% +135%

EV/EBITDA has been cut in half in 18 months — from 27x to ~13x — on a business whose EBITDA margin rose over the same period. That is the entire debate in one row.

Weighted Fair Value

Model Fair value Weight Rationale
Reverse DCF / FCF $98–122 50% Hard guided FCF; best fit for a cash compounder
DDM $82–114 25% Genuine dividend grower; sensitive to g
DYT $105–131 20% Strong signal, discounted for thesis-driven reset
Bogle (return, not value) 5% Cross-check — confirms ~11% base return
Graham $54 0% Structurally invalid for asset-light IP
Scenario Fair value What must be true
🔴 Bear $68–78 Interface and content ceded. ROSS reversed on appeal. Renewal cycle brings real pricing compression. FCF growth →2%.
🟡 Base $95–120 TRI keeps content economics, loses interface premium. FCF compounds 5–6%. Multiple settles ~20x FCF.
🟢 Bull $140–160 Moat holds; Fiduciary-Grade AI becomes the standard; Tax & Accounting durability recognized; partial re-rate.

Weighted fair value: $95–120 (midpoint ~$107). At $85.77 the stock trades ~20% below the low end of base case and ~25% below midpoint — with a live bear tail to $68.

Revised down from April's $120–150. Not because fundamentals disappointed — they beat — but because the May MCP integration was TRI conceding the interface layer is contestable, which lowers the defensible terminal growth rate. Honest reassessment, not a data change.


Phase 3 — Debate Round

Tension identified: Fundamentals (🟢 accelerating), Valuation (🟢 cheap on every applicable model), and Sentiment (🔴 -61%) are in direct conflict, with Moat split down the middle.

Round 1 — Moat (bear) challenges Fundamentals

"Your 8% organic growth is a lagging indicator. Enterprise legal contracts run 3–5 years. Every number you cite was locked in before agentic AI was deployable. Harvey went $100M → $190M ARR in five months — that's net-new budget being created, and at renewal it comes out of TRI's line. You're reading a rear-view mirror and calling it a windshield."

Fundamentals responds:

"Partially conceded — the lag is real and I flagged that no full renewal cycle has completed post-agentic-AI. But three things are not lagging indicators. (1) Gross margin expanded 48bps and EBITDA margin 159bps. Pricing pressure shows up in margin before it shows up in revenue, and it isn't there. (2) GenAI is 30% of Legal ACV — that's current-period contracted value on new products, not legacy renewals. (3) Legal organic accelerated to ~9%, the segment supposedly being disrupted. Harvey's $190M ARR is real, but the legal AI market grew from $4.59B to $5.59B — Harvey is taking new budget, not obviously TRI's."

Round 2 — Valuation challenges Moat (bear)

"Grant your entire thesis. At $85.77 the market prices 3.3% perpetual FCF growth. For you to be right about the price, not just the direction, TRI must compound FCF below 3.3% forever — while holding a 3% dividend at a 53% payout with 0.60x leverage. Even in your world where TRI is demoted to a licensed data layer, query volume rises with AI adoption. Name the mechanism that gets FCF growth below 3.3% in perpetuity."

Moat (bear) responds:

"Price-per-seat compression at scale. If interface owners capture the customer and renegotiate content licensing from strength, TRI's revenue per firm can fall 30–40% even as query volume triples. And you're assuming static competition — RELX/LexisNexis is fighting the same war and may price destructively to hold share. But I'll concede the honest point: sub-3.3% perpetual growth requires more than disruption, it requires permanent decline. My fair value is $70–85, not $50. I think the stock is roughly fairly valued, not a short."

Round 3 — Manager challenges Valuation

"Every model you ran anchors to a historical multiple or a growth rate the market rejects. EV/EBITDA halved from 27x to 13x while margins rose — markets aren't usually that wrong about a $37B liquid blue chip with 22 analysts covering it. What if the 27x was the error and 13x is correct?"

Valuation responds:

"Fair, and it's why I discarded Graham, weighted DYT down to 20%, and made reverse-DCF primary — that's the one model that doesn't assume mean reversion. And note: I'm not arguing for a re-rate. Bogle base case says you earn ~11%/yr with the multiple flat at 17.6x forever. The 27x was almost certainly too rich. I'm not underwriting a return to it — my base case tops out at ~20x FCF, well below history."

Conflict Resolution

Weighting by company type (per §5): TRI is a mature dividend grower with contested growth — not high-growth SaaS (where Moat + Sentiment dominate) and not a stable REIT (where Fundamentals + Valuation dominate). The moat question here is the valuation question, so neither can be subordinated.

Final weighting: Fundamentals 30% / Valuation 30% / Moat 30% / Sentiment 10%.

The debate genuinely narrowed the range. The bear conceded fair value of $70–85 (not a collapse); the bull conceded no re-rating is needed for a good return; both agree renewal-cycle evidence is the missing variable and won't arrive for 2–3 years.

Where that leaves us: $85.77 sits at the bear's fair value and ~20% below the base case. You are being paid a 3% growing dividend to wait for evidence, with downside roughly bounded by where the pessimist already thinks it belongs. That is a favorable asymmetry — not a screaming bargain, but a genuinely good one.


Type Overlay — Dividend Grower 🟢 Strongest Leg of the Thesis

Metric Value Assessment
Current yield 3.055% 🟢 ~2x its 5-yr average; highest in a decade+
FY2026 dividend $2.62 (+10%) 🟢 Raised into a 61% drawdown
3-yr dividend CAGR ~10% 🟢 $2.14 → $2.35 → $2.62
⚠️ 8–10yr dividend CAGR not computable ❌ Data gap — framework requirement unmet
Increase streak 33 years 🟢 Dividend Aristocrat-class consistency
FCF payout ratio ~53% 🟢 Well covered, room to grow
Earnings payout (GAAP) 69.98% 🟡 Elevated — but on tax-depressed GAAP EPS
2026 special dividend $1.44/share 🟢 Total 2026 cash yield ~4.7%
Net debt/EBITDA 0.60x 🟢 Dividend is not leverage-dependent

A 3.06% yield growing ~10%/yr from a 53% FCF payout on a 0.60x-levered balance sheet with a 33-year streak. On a 5-year hold, yield-on-cost reaches ~4.9% at 10% growth, or ~4.0% even if growth halves to 6%. This leg of the thesis survives the bear case entirely — even the bear's $70–85 fair value assumes a covered, growing dividend. For a user explicitly transitioning from pure growth toward growth + income, this is the single most attractive attribute of the position.

⚠️ Caveat on the 3.06% yield: it is elevated because the price collapsed, not because the payout ratio rose. That's the right reason for a high yield — but it means the yield is only attractive if the price decline was an overreaction. The yield is a symptom of the debate, not an independent argument.


Risks — Named and Ranked

# Risk Severity Probability Notes
1 Interface/workflow ownership loss 🔴 High Medium-High The thesis risk. TRI becomes a licensed data layer; revenue/customer compresses. Partly self-inflicted via MCP.
2 ROSS reversed on 3rd Circuit appeal 🔴 High Low-Medium Binary, uncontrollable, no known date. Removes the training-data wall.
3 Renewal-cycle evidence gap 🟠 Medium High 3–5yr contracts mean disruption won't appear in results for 2–3 years. We are flying on lagging data.
4 Multiple stays compressed 🟠 Medium Medium-High Sector-wide info-services de-rate. May be a permanent regime, not a dislocation.
5 ROIC only 10.5% 🟠 Medium Certain Mediocre capital deployment; growth historically bought via M&A, not generated.
6 Capital returns exceed FCF 🟡 Low-Med Certain FY25 returns 189% of FCF, cash -75%. Cushion spent; FY27 plan must fit inside $2.1B.
7 Negative tangible book 🟡 Low-Med Certain 71% goodwill/intangibles. Impairment risk if AI thesis proves right.
8 Q2 earnings Aug 5 🟡 Low-Med Certain — 13 days Near-term binary. Any organic deceleration below 7% validates the bear.
9 ICE contract / ESG overhang 🟡 Low Medium Reputational; shareholder vote failed but pressure persists.
10 FX / Canadian domicile 🟢 Low Certain USD reporting, ~75% US revenue. Minor.
11 Woodbridge 68% control 🟢 Low Certain Cuts both ways — long-horizon alignment vs. minority-holder powerlessness.

A portfolio-specific passage was removed from the public build.

Monitoring — What Would Change the Verdict

🔴 Downgrade triggers (→ [6.0] or lower): - Q2 2026 (Aug 5) organic growth <7% or FY guidance cut - Big Three organic drops below 8% - Gross or EBITDA margin contraction — the first true sign of pricing pressure - Legal Professionals organic falls below 6% - ROSS reversed by the Third Circuit - GenAI share of ACV plateaus or declines - Net debt/EBITDA above 2.0x, or a dividend increase below 5%

🟢 Upgrade triggers (→ [8.0]+): - Q2 organic ≥8% with margin expansion on track to ~40% - Renewal-cohort disclosure showing retention holding at 90%+ post-AI - GenAI ACV share rising past 40% - Third Circuit affirms ROSS - Tax & Accounting organic accelerating (proof the durable segment is being ignored) - Insider or Woodbridge buying at these levels

📅 Calendar: Q2 2026 earnings — Wednesday August 5, 2026, 8:30am EDT. This is the single most important near-term event. Re-run this analysis after.


Final Verdict

Rating: BUY — accumulate in tranches Conviction: [7.5] / 10 (unchanged from 2026-04-26) Fair value: $95–120 base (revised down from $120–150) | Bear $68–78 | Bull $140–160 Add zone: $82–98 | Trim zone: $150+

TRI is a 150-year-old, family-controlled, 33-year dividend grower generating $2.1B of free cash flow at a 0.60x-levered balance sheet, whose operating results are accelerating while its stock sits 61% below its peak. Every applicable valuation model puts fair value above the current price, and the market's implied assumption — 3.3% perpetual FCF growth — is difficult to reconcile with 8% organic growth and expanding margins.

The bear case is real and I will not dismiss it: the interface layer is genuinely contestable, TRI itself conceded as much with the Anthropic MCP integration, and no full renewal cycle has yet tested the thesis. But the bear's own fair value is $70–85 — meaning at $85.77 you are buying at the pessimist's estimate, collecting a 3% growing dividend to wait, with base-case upside of 20–40%.

This is a close call on quality-of-business and a clear call on price. The asymmetry is favorable but not overwhelming, which is exactly why the answer is accumulate deliberately rather than back up the truck. The larger error in this position has not been the analysis — it has been three months of inaction while the stock sat in the add zone. Size it or drop it.


Sources


Analysis run 2026-07-23. ⚠️ Yahoo Finance MCP unavailable; roic.ai limited to 2 years of history. Multi-year CAGRs (FCF, dividend, ROIC trend) and insider/institutional data could not be computed. Reconnect Yahoo Finance MCP and re-run before committing capital beyond the tranche-1 size.