Imagine a game studio that sells a majority of its disc-pressing plant to outside capital, keeps every line of copyright, and attaches one clause: if disc revenue falls below a certain line, the studio makes up the difference. The studio’s own digital launcher is what is killing disc demand. The party eroding the demand is the party insuring against its disappearance.
Translate that into financial language and you have the transaction Thomson Reuters, the Toronto-based legal and tax information company, announced in July 2026 and closed on October 1: the sale of 51% of its Global Print business to capital accounts advised by KKR for USD 500 million in cash. The copyright is the legal and tax content; the launcher is Westlaw and CoCounsel; the disc plant is Global Print, now operating as Westbridge Print. The clause appears in the disclosure as one sentence — financial support “designed to guarantee minimum returns in certain circumstances.”
The structural question this deal answers is not what a declining print business is worth. It is who pays for a decline the seller itself is engineering, and what a buyer that needs a floor rather than an upside will pay once that cost has been contractually reassigned.
The Deal
Thomson Reuters is listed in Toronto and on Nasdaq, with Woodbridge, the Thomson family holding company, owning roughly 71%. In 2025 it reported revenue of USD 7.48 billion and adjusted EBITDA of USD 2.94 billion, a 39.2% margin; its three core segments grow organically at 9–10% a year, and Global Print sat at the end of the table. Credit ratings are A- (S&P) and Baa1 (Moody’s), both on positive outlook, with net debt at 0.9x EBITDA.
- Buyer: capital accounts advised by KKR, not a named KKR fund. The KKR representative quoted in the announcement was the co-chief investment officer of Global Atlantic, KKR’s insurance platform.
- Consideration: USD 500 million, all cash, no financing condition, for 51%. Thomson Reuters retains 49%.
- Asset: print facilities in 13 countries plus exclusive rights to distribute Thomson Reuters content in print and on ProView, its e-reader platform. 2025 segment adjusted EBITDA: USD 185 million.
- Flows back to the seller: a royalty for the exclusive distribution rights and fees under a multi-year transition services agreement (TSA — the seller’s temporary provision of back-office functions). Per the Q2 2026 management report, these payments “largely offset” the seller’s stranded costs.
- Seller support: financial support “designed to guarantee minimum returns in certain circumstances.” Trigger, cap, and duration are not disclosed.
Three forces shaped the structure. Generative AI is absorbing legal and accounting research workflows, and a listed publisher’s multiple now depends on how cleanly it sits on that narrative. Managers running insurance balance sheets have become the most aggressive buyers of stable cash flow, with return requirements several points below a buyout fund’s. And the legacy asset is squeezed between the print-materials cost pressure Thomson Reuters lists as a risk factor and the demand erosion it is itself accelerating.
Structure Breakdown
USD 500 million for 51% implies roughly USD 980 million for 100%, about 5.3x 2025 segment adjusted EBITDA of USD 185 million. Five times for a shrinking asset looks cheap, but the 185 is a Thomson Reuters segment figure, pre-royalty, not a joint-venture figure. If royalty and TSA fees were calibrated to leave the seller’s margin neutral, as “largely offset” suggests, the royalty rate was reverse-engineered from stranded costs rather than set at a market price. Assume stranded costs of USD 20–40 million a year and the JV’s standalone EBITDA falls to an estimated USD 145–165 million, or 6–7x. The first number an investment committee should want is the unpublished royalty rate, not the published 185.
The second weight is the direction of margin. Segment margin rose from 33.1% in Q3 2024 to 37.7% in Q2 2026 while revenue fell 3–7% in almost every quarter, the stated reasons being lower shipments and lower costs. A business whose margin expands as revenue shrinks looks most attractive at the moment it is most dangerous: if the savings came from deferring maintenance, headcount, and sales investment, those costs return to the standalone P&L after separation. Organic decline had slowed to 3% by Q2 2026, but assuming that linearity holds for ten years through a generational turnover in the legal profession is optimism, not a base case.
The third weight is the identity of the capital. An insurance executive spoke, the vehicle was “capital accounts” rather than a fund, and the consideration was all cash with no financing — clues pointing to a buyer satisfied with 8–9% rather than one requiring 15% or more. That is the desk’s inference, not a disclosure, and the calculations below rest on it. Take post-royalty EBITDA of roughly USD 155 million and unlevered free cash flow near USD 108 million in a perpetual-decline model: a private equity discount rate puts 100% of the business in the USD 600 million range, an insurance discount rate at USD 800–900 million. The implied USD 980 million is reachable only through the insurer’s lens, and even there it sits at the top of the range.
Operations alone do not get there. On the same assumptions, KKR’s first-year distribution is about USD 55 million — an 11% initial cash yield that shrinks every year. At a 6% annual decline, ten years of distributions total about USD 420 million, less than the USD 500 million invested; at 4%, about USD 460 million; even with a terminal value at 3x EBITDA, the unlevered return sits in the low single digits. The gap is filled by one line of contract, and that line exists because Thomson Reuters wears four hats: content supplier, seller of the digital substitute that displaces print, 49% co-owner, and guarantor. When a licensor cannibalizes its licensee’s demand, the only alignment device a buyer can demand is to charge the cost back to the seller — so if the guarantee’s trigger is tied to the rate of demand decline, the support is less a credit enhancement than a put option on self-cannibalization, and if it is not, the buyer has failed to transfer the single most important risk. The deal’s center of gravity is five undisclosed terms: the trigger, the cap, the duration, the royalty rate, and whether the support is an external guarantee by Thomson Reuters or a subordination of its 49% in the JV’s distribution waterfall.
The Governing Logic
This structure is weighted almost entirely toward protecting principal, and the reason is the buyer. A buyout fund must double or triple capital inside ten years to raise its next fund. An insurance account must meet liabilities promised to policyholders over decades at a cost in the mid single digits; a spread of two or three points above that completes the mission. Insurance capital asks not how much an asset can rise but what it pays in every case, and every clause here is drafted to answer that question — on the standing caveat that the buyer’s insurance identity is inferred, not disclosed.
Downside Protection: Four Lines
The first line is the seller’s minimum-return guarantee, in one of two forms. In the external form, Thomson Reuters pays the shortfall from its own balance sheet, and within the guarantee’s scope the stake’s credit risk stops being print risk and becomes Thomson Reuters credit — A-/Baa1, 0.9x levered. The asset is then an investment-grade floor with 51% control layered on top: an 11% initial yield on near-investment-grade downside, with declining print as the packaging. In the internal form — the 49% ranking behind the 51% in the waterfall — the floor is only as thick as the distributions the 49% can forgo, and those exist only if the JV earns them. Which form it takes decides whether this is a credit instrument or print equity.
The second line is the exclusive license: the seller cannot switch printers or resume printing itself, so horizontal competition is closed. The hole is vertical — a customer cancelling print for Westlaw is not a breach but the seller’s ordinary course of business, which is why the first line was needed. The third line is cash conversion: depreciated plants and subscription prepayments let 70–80% of EBITDA convert to distributable cash, returning 11% of principal in year one, and the faster cash is extracted before any guarantee is tested, the less the guarantee must carry. The fourth, the multi-year TSA, keeps part of the separation risk with the seller so a first-hundred-days failure does not consume the first year’s distributions.
On these lines the designer’s base case is a run-off, not a sale — declining distributions absorbed against the maturity profile of insurance liabilities — with JV-level leverage against a guarantee-backed cash flow as the second path and consolidation of competitors’ print volumes a third the designer is unlikely to have priced. From the seller’s side, conceding the guarantee is cheap: Thomson Reuters receives USD 500 million, books a pre-tax gain on disposal, reclassifies Global Print as discontinued operations, lifts group organic growth by 60–70 basis points, keeps every piece of IP and editorial control, and retains 49% of the run-off. For a listed company managing an AI narrative, a contingent liability costs far less than a lower headline price. The guarantee is not a cost; it is part of the price.
Readers of this library will recognize a cousin in The Synthetic Debt Mirage — equity on the press release, debt in the economics. The mechanism sits elsewhere: the Mirage is a constrained parent raising capital against a unit it keeps, while here an unconstrained parent disposes of a unit it erodes, and the floor comes from the seller’s conduct-linked undertaking rather than from the instrument’s terms.
What Must Be True
Proposition one: the support is a parent-level guarantee, not a waterfall preference inside the JV. Falsifier: Thomson Reuters’ next annual filing records no guarantee liability at the parent and the support appears only as a subordination of the 49%’s distributions — the floor is then print equity and the credit-product reading fails.
Proposition two: the margin expansion from 33.1% to 37.7% is structural, not deferred cost. Falsifier: a reversal in Westbridge Print’s first standalone year, with maintenance capex and staffing rising while the revenue decline holds near 3%. Three layers already separate the published 185 from what the JV will earn — excluded corporate costs, a royalty set at the seller’s neutrality, and possibly deferred costs. If the third is real, the multiple paid approaches double digits on a cash basis.
Proposition three: the decline stays linear at 3–4% long enough for distributions to recover principal within the guarantee’s term. Falsifier: the rate moving to double digits as law-firm AI adoption accelerates, tariff or materials shocks raise costs, and the federal budget pressure Thomson Reuters cited at group level reaches print subscriptions concentrated in government and law schools. With fixed costs near 45% of the base, a 10% revenue fall becomes an 18% EBITDA fall. The chain then runs: EBITDA halves within three years, the guarantee triggers, the payment dents the seller’s P&L, the dent shows up as slower content updates and cooler renewal talks, churn accelerates, the guarantee triggers again, the cap is exhausted, and the buyer stands exposed with no strategic acquirer and every financial buyer discounting cash flows that no longer carry a floor.
The tail risk, then, is not print’s decline, which is priced. It is reflexive: drawing the insurance damages the relationship with the insurer, and that damage hits the business the insurance was meant to protect. The guarantee is the floor and the detonator beneath it, and the clause worth reading most closely is the one that governs how the two parties must treat each other the day after it pays.
The Architect’s Takeaway
If you were building your own deal room with none of KKR’s capital, the first thing to take from this structure is not the USD 500 million but the principle of separation: keep the source, sell the distribution. Thomson Reuters held the layer where value stays and sold only the layer where value is leaving — to the capital that discounts it least. The test for which layer to shed is one question: are you the one reducing demand for it? A creator’s source is the fan relationship and the plant is physical merchandise; a developer’s source is the core engine and the plant is legacy on-premise support; an architect’s source is a design language and the plant is permit filing.
Second, use a guarantee as a substitute for a price cut. “Sell it to me cheaper” is easy to refuse; “you are the one killing the demand, so you cover what you kill” is hard to refuse. Wherever a counterparty controls the upstream of your revenue — a platform’s algorithm, an anchor customer, a sponsor’s deal flow — ask for a floor tied to its own conduct: compensation on a policy change, a termination fee on a decision to insource, a tail clause that pays if a deal you introduced closes by another route. Pricing the other side’s self-cannibalization is the essence of a guarantee clause.
Third, the price of an asset is set by the buyer’s discount rate, not by the asset. The same plants were worth about USD 600 million to a buyout fund and USD 900 million to an insurer. Show the asset to the capital with the lowest required return — professionals approaching retirement, family offices, community members who value predictability — and the same thing sells for 1.5x, because to them a declining asset reads as a bond with a visible maturity. Choosing the buyer is choosing the price.
Fourth, accept run-off as a design principle: sales depend on the market’s permission, distributions only on contract. For a fading newsletter or a product with slipping renewals, ask whether the investment comes back in distributions within a few years on a conservative decline rate; if that period is shorter than the guarantee’s term and the guarantor stays solvent, the structure does not lose. A guarantee from a counterparty without the capacity to pay is a drawing, not a floor. Fifth, read whose neutrality the royalty was set to. Platform take rates, franchise royalties, and JV revenue splits all deserve the question: whose P&L was this number reverse-engineered to protect?
And once a guarantee is won, write the day after it triggers on the same page — for instance by converting part of any compensation into credit against future business, so the relationship survives the payment. Whoever designs the floor also designs the step after the floor is reached. That is the difference between a structure that does not lose and one that wins once.
Syndicate’s Note
House Verdict: WATCH. The read flips if Thomson Reuters’ next annual filing shows the minimum-return support as a distribution preference inside the JV rather than a guarantee liability at the parent — at that point the floor is print equity, not Thomson Reuters credit. The verdict classifies structural learning value; it is not an investment judgment.
The desk is adding this mechanism to the declassification queue as The Cannibalization Put: a disposal in which the seller, itself eroding the asset’s demand, guarantees the buyer’s minimum return, so that the floor under the price is a put on the seller’s own conduct. One case does not make a pattern; a second disposal with a seller-funded floor on a seller-cannibalized asset would.
What was dissected here is not KKR buying a print business for USD 500 million. It is a buyer that appears to be insurance capital acquiring a yield asset whose floor is supplied by the seller, and a seller that walked away with the cash, the gain, the growth-rate uplift, and all of the IP, having conceded one clause. That clause is both insurance and detonator. Until it is read, the USD 980 million has not been valued.
FAQ
What is a minimum-return guarantee in a carve-out?
A seller undertaking to make up the shortfall if the carved-out business delivers less than an agreed return to the buyer, usually within a defined trigger, cap, and period. In the Thomson Reuters Global Print deal it is disclosed only as support “designed to guarantee minimum returns in certain circumstances.” Its form — parent guarantee or distribution preference inside the JV — determines whether the buyer holds seller credit or business equity.
Why did Thomson Reuters keep 49% of Global Print?
The retained stake keeps nearly half of the run-off cash flow while removing the segment from reported growth, and the JV’s royalty and TSA fees largely offset stranded costs. The disclosure does not say whether the 49% is also the vehicle for the minimum-return support.
Why would insurance capital pay more than private equity for a declining business?
Because price is a function of the buyer’s discount rate. Insurance accounts can capitalize the same cash flows at 8–9% where a buyout fund needs 15% or more; on the desk’s estimates, that difference moves 100% of Global Print from the USD 600 million range to USD 800–900 million before any seller guarantee is counted.
This article is for informational purposes only and does not constitute investment advice, an offer, or a solicitation. Figures are based on public reports and may be estimates.
For more structural insights and deep-dive video breakdowns, visit Structure Syndicate on YouTube.