Thanks for reading! Subscribe for free to receive new posts and support my work.

Subscribe now

WHAT IT DOES

TRI has five reportable segments. Legal Professionals serves law firms and governments, anchored by Westlaw, Practical Law and CoCounsel. Corporate serves companies from small businesses, in house legal teams, and even seven of the largest global accounting firms. Tax, Audit & Accounting Professionals serves accounting firms and other Tax & Audit professionals . Reuters supplies news and data, including to LSEG’s Data & Analytics business under the 30-year agreement running to 2048. Global Print is the legacy print business, which as of this quarter, is on its way out to KKR (more below).

Legal Professionals, Corporates and Tax, Audit & Accounting Professionals together are the “Big 3,” about 83% of total revenue in Q2 2026. Recurring revenue was 82% of the quarter.

The company’s own description of itself changed slightly this quarter. The 2025 annual report said it powers professions with “AI they can trust.” The Q2 2026 MD&A upgrades that to “Fiduciary-Grade AI™,” defined in the glossary as:

“AI designed for professionals with duties of care and regulatory oversight, drawing on our authoritative, domain-specific content; protected by rigorous privacy and security safeguards; shaped by subject-matter experts; and designed to produce transparent outputs that can be verified.”

As someone who uses their product everyday in my profession, there is truth in that. Harvey produces stuff I cannot trust and have to constantly vet, TRI’s legal research uses the case law from Westlaw and cites back to it, so it is very easy to double check and I have not found an instance of it being wrong yet.

OWNERSHIP AND GOVERNANCE

The Woodbridge Company Limited (“Woodbridge”), (Woodbridge is the primary investment vehicle for members of the Thomson family of Canada and is the principal shareholder of TRI with a majority stake). Woodbridge’s stake keeps climbing as the share count has been shrinking, since 2015.

Woodbridge’s position has been increasing due to buybacks and reorganizations by TRI, other than the DRIP on their dividends, they generally are sellers in the market versus buyers.

on August 7, 2026, Woodbridge filed a Form 45-102FI, the standard notice you file before reselling control-block shares in Canada. As of August 5, 2026, Woodbridge and its affiliates held 307,593,126 common shares, a 71% voting interest, and the notice discloses intent to sell up to 1,500,000 shares through the TSX, Nasdaq, or private agreement, with BMO as the broker on both exchanges. That’s about half a percent (0.5%) of the position. This seems like selling in the ordinary course, but flagging it as the shares may come under pressure on days larger blocks are being sold.

I still like CSU and its subs more, due to higher growth prospects and even though it is minor, there is stock-based compensation (“SBC”) at TRI.

TRI has a variety of options and restricted share units on various vesting schedules. There are 69,150,969 shares authorized for issuance, with 6.79M still available to grant as of year-end 2025. This is much different from CSU’s zero-SBC-since-2007 approach that I respect and value much more.

SBC expenses were $37M in Q2 2026 and $70M for H1, up from $29M / $57M a year earlier. One thing I do respect is TRI doesn’t add it back SBC to adjusted EBITDA. Plenty of companies strip stock comp out of their “adjusted” numbers so the margin number looks better than it actually is. TRI just leaves it in.

The realized dilution from it is small. The gap between basic and diluted shares in the EPS calculation, which is the actual effect of SBC added 110,735 shares in Q2 2026 and 193,994 for the first half, against the now 433.2-435.6 million shares outstanding. Roughly ~300,000 shares issued through SBC, compared to the ~11.5 million shares that were bought back / eliminated by the reorg. So, unlike a Tesla that continually dilutes shareholders, or a Meta that spends billions to offset the dilution, TRI’s share count is genuinely decreasing each year and SBC does not seem to be on the same toxic scale as other companies.

REVENUE, ORGANIC GROWTH AND FREE CASH FLOW

TRI’s own definition of free cash flow, from the Q2 2026 MD&A:

“Net cash provided by operating activities and other investing activities, less capital expenditures, payments of lease principal and dividends paid on our preference shares.”

Chart below summarizes the annual figures from 2014 – 2025:

2026 has been going strong thus far. Q1 revenue was $2,087 million (helped by the usual tax-season seasonality), Q2 was $1,954 million, and for the first six months total revenue was $4,041 million, up 10% and organic growth of 8%. The Big 3 grew 11% total and 10% organic in just Q2. Adjusted EBITDA margin was 38.1% in Q2 versus 37.8% a year earlier. Free cash flow was $727 million in the quarter and $1,059 million for the half, up from $843 million in the first half of 2025. Diluted EPS was $1.02 in Q2 versus $0.69, up 48%.

Management raised the full-year outlook twice this year. The original February 2026 guide called for 7.5% to 8.0% total and organic growth, Big 3 around 9.5%. In August, after Q2 came in, they raised it to approximately 8.0% total and organic, and 9.5% to 10.0% for the Big 3. Everything else in the outlook (EBITDA margin, free cash flow, capex, tax rate) was left unchanged.

REFINITIV

Still the single biggest thing to understand about the balance sheet history, and it explains most of the jumps in the table above.

Going into 2018 TRI ran four units: Financial & Risk, Legal, Tax & Accounting, and Intellectual Property & Science. IP&Science had already been sold in 2016. On January 30, 2018, TRI signed to sell 55% of Financial & Risk to funds managed by Blackstone (CPP Investment Board and GIC alongside) for about $17 billion, keeping a 45% stake. Closed October 1, 2018. The business became Refinitiv.

From the 2018 MD&A on what happened with the money:

“$10 billion was returned to our shareholders through a $6.5 billion substantial issuer bid/tender offer, a $2.3 billion return of capital transaction and $1.2 billion of share repurchases... $4 billion of debt was repaid... $2 billion was set aside for future reinvestment in our business through strategic acquisitions in key growth segments.”

The tender offer bought back about 138 million shares at $47 each. Woodbridge tendered proportionately, about 88.9 million shares, to hold its ownership percentage flat, through what the filings call a “Qualifying Holdco Alternative.” Shares outstanding went from 709.6 million in March 2018 to 501.2 million in March 2019, the single biggest step down in the whole share count series.

Then in August 2019, TRI and Blackstone agreed to sell Refinitiv again, this time to the London Stock Exchange Group, an all-share deal at roughly $27 billion enterprise value. Closed January 29, 2021. TRI came out owning about 82.5 million LSEG shares through a jointly-owned entity called YPL, worth about $9.8 billion at closing, with a staged lock-up.

From there TRI steadily sold off the position:

• 2022: sold about 21.2 million co-owned LSEG shares to Microsoft, roughly $1.0 billion of proceeds to TRI.

• 2023: sold about 56 million LSEG shares for $5.4 billion gross, funded a $2.0 billion return of capital with part of it.

• 2024: sold the last roughly 16.0 million shares for $1,854 million, fully out.

The equity stake is gone. What’s left is the Reuters news agreement with LSEG’s Data & Analytics business, running to 2048, $398 million of revenue in 2025 at a contractual minimum that adjusts for CPI and FX.

THE CHANGE PROGRAM

Announced February 23, 2021, alongside a dividend increase: a two-year restructuring “to transition our company from a holding company to an operating company and from a content provider to a content-driven technology company.” Stated 2023 targets: organic growth 5% to 6%, EBITDA margin 38% to 40%, free cash flow $1.8 to $2.0 billion, $600 million of annual opex savings, capex down to 6.0% to 6.5% of revenue.

Completed December 2022, $540 million of annualized run-rate savings against the $600 million target. Actual 2023 results landed inside the ranges set two years earlier: organic growth 6%, EBITDA margin 39.3%, free cash flow $1.871 billion.

THE CAPITAL RETURNS CONTINUE

Dividends increased for the 33rd consecutive year with the February 2026 announcement, 10% to $2.62 USD per share, the fifth straight 10% increase.

Stated targets from the MD&A: 2.5x net debt to adjusted EBITDA, 50% to 60% of free cash flow paid out as dividends, at least 75% of annual free cash flow returned via dividends and buybacks combined, and ROIC double or more the weighted average cost of capital.

TRI has now paired a big one-time cash distribution with a share consolidation, a reverse stock split, three separate times:

• 2018: $2.3 billion, part of the $10 billion post-Refinitiv return.

• June 2023: $2.0 billion, a $4.67 per share cash distribution and a consolidation at 1 pre-consolidated share for 0.963957 post-consolidated shares.

• May 4, 2026: $605 million, a $1.435518 per share cash distribution and a consolidation at 1 pre-consolidated share for 0.984560 post-consolidated shares, funded out of the same pool of LSEG sale proceeds from May 2024. When this was first announced in February 2026 the company estimated it at roughly $1.36 per share; it closed higher, at $1.435518. Woodbridge participated. Outstanding shares fell by about 6.5 million.

The NCIB (normal course issuer bid) has concluded this year, with 1.6B spent on buybacks in the last 1.5 years. Average price paid per share chart below:

In hindsight, if they had not repurchased the 6M shares in the first half of 2025, which they spent $1B on as they could have bought a lot more back this year, since it is now roughly 40% below where they were buying at a year earlier. However, to be buying aggressively in the same six-month stretch that revenue hit a record, organic growth in the Big 3 hit 10%, free cash flow grew 26%, and management raised guidance twice, is nice to see.

GLOBAL PRINT AND THE KKR JOINT VENTURE

On July 14, 2026, TRI signed a definitive agreement to sell 51% of Global Print (a laggard as revenue has been declining for years, as professionals switch to online materials) to capital accounts advised by KKR, keeping a 49% interest in the new joint venture. About $500 million of gross proceeds expected at closing, targeted for Q4 2026, subject to regulatory approval.

TRI keeps the intellectual property and full editorial control over the content. The joint venture gets an exclusive license to distribute it in print and on ProView, Global Print’s eBook platform, and pays TRI a royalty for it. TRI also provides operational services under a multi-year transition services agreement, which per the MD&A will “largely offset stranded costs from the separation.” TRI has also agreed to give KKR “certain financial support designed to give KKR a minimum return on its equity investment in the joint venture under certain circumstances,” which the filings don’t spell out further.

Global Print becomes a discontinued operation starting in Q3 2026 and stops being a reportable segment. Management expects the deal to add 60 to 70 basis points to total company organic growth once it closes, since Global Print has been shrinking (organic revenue down 3% to 4% in recent quarters) and dragging on the average.

Similar to the Financial & Risk sale but on a smaller scale ($500 million against the $17 billion Refinitiv deal), TRI’s selling the majority stake, keeping a minority interest and the content rights, and collecting a royalty instead of holding onto a losing segment.

COCOUNSEL, WESTLAW, AND THE AI STORY

From what I have heard and know, there are no Canadian law firms that have plans to drop Westlaw or Practical Law and if anything, firms have been adding CoCounsel licenses. Westlaw is the case law and treatise database. Practical Law is the precedent and checklist library, written by Bay Street / Wall Street lawyers. CoCounsel is the AI layer trained on, and built on top of both, which is why it gets added rather than substituted. Firms aren’t switching to Claude legal or any other junk for legal research when Westlaw carries cases going back over a century that aren’t available anywhere else, and the subscription cost is a rounding error against our billing rates that rise every year.

When a lawyer wants to draft a document, they will be able to use CoCounsel which will pull a practical law template and be able to draft it using their prompts and additional deal information, which will be way more dependable than Harvey or Legora.

Q2 2026 backs this up in the segment numbers. Legal Professionals organic revenue grew 10% for the quarter, recurring revenue (97% of the segment) up 9% organically “primarily driven by Westlaw and CoCounsel,” per the MD&A, and transactions revenue up 18% “primarily driven by CLEAR,” a product that aids law enforcement and other fraud and risk professionals with searches.

TRI’s marketing for it states:

Accelerate investigations confidently through a vast collection of public and proprietary records with CLEAR’s online investigation software. Access exclusive data feeds, quality domestic data, and incarceration records with source transparency and frequent updates, all in a single platform.

Corporates organic recurring revenue grew 9%, also led by Westlaw, CoCounsel, Indirect Tax, Pagero and CLEAR. Tax, Audit & Accounting Professionals organic recurring growth of 9% was driven by GoSystem and CoCounsel alongside Cloud Audit Suite.

Big 3 organic growth was 10% in Q2, extending a run that’s been 6% or better every year since 2022.

BALANCE SHEET AND DEBT

As of June 30, 2026: net debt was $2,628 million, up from $1,896 million at year end, leverage ratio of net debt to adjusted EBITDA at 0.9x, up from 0.6x six months earlier. Still well under the 2.5x target and the 4.5x covenant max (0.8x under the credit facility’s own calculation). The jump is mostly commercial paper, $1,618 million outstanding versus $295 million at year end, borrowed to fund the return of capital, the buyback, and dividends while operating cash flow alone wasn’t enough to cover it in the first half.

TRI repaid its $500 million 3.35% notes at maturity in May 2026 using cash on hand and commercial paper. Total debt position was about $3.0 billion, average maturity around 14 years at a weighted average rate just under 6%, up from roughly 10 years and just over 5% a year earlier, because the remaining book now skews to the longer-dated notes maturing in 2035, 2040 and 2043. Credit ratings unchanged: Baa1/A-/A-, Moody’s outlook moved to Positive in November 2025 and has stayed there.

THINGS TO WATCH

A few items straight from the filings.

Net interest expense keeps running ahead of where guidance started the year. It was $47 million in Q2 2026 versus $35 million a year earlier, up 36%, and $86 million for the half versus $65 million, up 32%, on the higher commercial paper balance and lower interest income. This is the second year running this line has drifted higher than the original outlook: 2025’s actual came in at $143 million against an original ~$130 million target, and 2026’s outlook has already moved from $150 to $160 million in February to $180 to $190 million by May, unchanged in the August update.

The six-month period of 2026 included “losses recognized from the resolution of a tax dispute on a portion of the receivable balance from LSEG relating to the tax indemnity,” per note 8 of the interim statements. That is in addition to the underlying HMRC Diverted Profits Tax matter, the $430 million paid under protest through 2023 relating to 2015 to 2018 UK tax years, which TRI still believes it will substantially recover. Worth watching how that resolves.

Reuters segment margin is still soft. Adjusted EBITDA margin was flat at 20.8% in Q2 but down 180 basis points to 18.6% for the half, on higher editorial costs.

Revenue, organic growth, and free cash flow all set records through the first half of 2026. Big 3 organic growth hit 10% in the second quarter. Management raised guidance twice this year. The company kept buying back its own stock the entire time, at a price roughly 40% below what it was paying a year earlier, will be curious to see if they renew the NCIB (it expires August 18, 2026).

-CCC

Thanks for reading! Subscribe for free to receive new posts and support my work.

Subscribe now

Subscribe now

Like CSU and TOI, Lumine does VMS, but unlike the broad coverage of CSU and TOI, LMN is focused on the communications and media industry, from their MD&A:

“We acquire, strengthen and grow businesses in the Communications and Media industry, with businesses located worldwide. Generally, our businesses provide mission critical software solutions that address the specific needs of customers in particular segments of the Communications and Media industry.”

At the time of the spin-out prospectus, the business was organized into two Communications operating groups spanning 22 independently managed software business units across Canada, the US, UK, Switzerland, Sweden, Portugal, Singapore, Australia, and Israel, serving cable, fixed-line, fixed-wireless and fiber ISPs, mobile operators, MVNOs, mobile tower operators, and mobile financial service operators. Alongside that sits a Media operating group, anchored by WideOrbit and headquartered in San Francisco, which sells the software that television, radio, and cable networks use to manage advertising sales from order to invoice.

Revenue breaks into four buckets: software license fees, maintenance and other recurring fees, professional services, and hardware resale. In Q2 2026, maintenance and other recurring revenue was $178.5 million of $235.1 million total revenue, or 76%. A year earlier, in Q2 2025, that mix was 72%. The recurring base keeps getting a larger as the company acquires and integrates more businesses, which is reassuring since it’s the same trend we have watched play out at CSU for a long time.

Since finalizing the spinout, Lumine has reported 256,620,389 shares outstanding, basic and diluted, in every quarter through Q2 2026. Like CSU, no new shares issued. For a company that has spent well over a billion dollars on acquisitions since spin-out, that is pretty impressive and I like because it helps reduce the Mark Leonard being gone concern.

Subsidiaries and spinouts have the same discipline CSU has had since 2007 (21,191,530 CSU shares outstanding - always).

The bonus formula is also worth quoting because it replicates the CSU model and also shows how ingrained the culture is, even in the spinouts.

“Our corporate bonus plan... is based upon return on invested capital (’ROIC’)... and net revenue growth.” Bonuses are calculated by netting a risk-free rate (currently 5%) against ROIC, and there is explicitly no employee stock option program, “nor do we currently expect to implement one in the near future.”

Leadership Scare

On January 2, 2025, Lumine announced that David Nyland was taking “a temporary leave of absence, effective immediately,” with Group President Tony Garcia stepping in as interim CEO. No reason was given in the release. Six weeks later, on February 14, 2025, the company announced Nyland’s return effective February 17, with Garcia remaining as Group President. Again, no explanation was offered publicly.

That same month, on January 10, 2025, director Robin van Poelje resigned from the board and was replaced by Laurie Schultz, a long-time software executive (former CEO of Galvanize) who also happens to sit on Constellation’s own board since 2021. In May 2025, Beattie handed the CFO role to Mary Anne Lavallee, previously CFO of Gateway Services and before that EVP/CFO/Chief Transformation Officer at Postmedia, with Beattie staying on as a director. So, in the space of about five months, Lumine cycled through a CEO absence and return, a board seat change, and a CFO transition.

That being said everything seemed fine at the annual meeting this summer, but just flagging so that I also expose potential issues, I am obviously biased and think the company will do well, but don’t want to be too one sided.

Revenue & FCFA2S

Lumine also uses the Free Cash Flow Available to Shareholders, defined the same way as TOI and CSU.

“FCFA2S refers to net cash flows from operating activities less interest paid on lease obligations, interest paid on bank indebtedness, transaction costs on bank indebtedness, repayments of lease obligations... and property and equipment purchased net of proceeds from disposal... our objective is to invest all of our FCFA2S in acquisitions which meet our hurdle rate.”

Below is Q1 2023 – Q2 2026 (in US dollars, with year-over-year revenue growth and organic growth (foreign-exchange adjusted, where the company discloses it):

Revenue has been compounding at a double digit rate since 2023.

Organic growth has been rougher than TOI and CSU, for some stretches. The back half of 2024 ran negative 9% for two straight quarters, the organic maintenance growth has swung from -12% to +6% within twelve months. Some of that is currency (management discloses both the headline and FX-adjusted figures, and they frequently differ by several points).

The company just spent over $500 million integrating three businesses in six months, so some variance in the 2025-2026 numbers is expecting while LMN “digests” its latest targets.

M&A Going Strong in 2025 and 2026:

• 2023: WideOrbit Inc. acquired as the anchor deal at spin-out, the media advertising software business that WideOrbit’s founder, Eric Mathewson, rolled into Lumine in exchange for special shares.

• 2023: Synchronoss Messaging and NetworkX businesses acquired (referenced in the later Skyfall release as the first of a multi-year relationship with Synchronoss).

• 2023: WizTivi SAS, a French subsidiary, financed with a €10 million HSBC term loan (repaid in full in June 2026, ahead of its 2028 maturity).

• December 2023, closed April 2024: The Device Management and Service Management Platform businesses purchased from Nokia for up to €185 million, including up to €35 million contingent on first-year performance, relaunched under the restored “Motive” brand.

• December 2025, closed February 13, 2026: Synchronoss Technologies Inc. acquired outright for $309.3 million in aggregate cash consideration, Lumine’s first acquisition of a full public company. The deal was originally announced at an implied equity value of about $116.4 million and enterprise value of $258.4 million; the final consideration came in higher.

• July 1, 2026: Quortex (Synamedia Vividtec Holdings plus select Synamedia Video Network assets) and Imagine Communications Holdings Inc. acquired together for $211.3 million cash plus $22.4 million of deferred payments, $233.7 million total, deepening the media supply chain presence.

Total capital deployed in 2026 is already over $500 million. That is more than the company’s entire cumulative FCFA2S generated since spin-out, so there is heavy use of debt and the debt facilities are summarized below.

How the balance sheet is built

Every acquisition financing is ring-fenced to the specific subsidiary it funds, collateralized by that subsidiary’s assets and governed by covenants tied to that subsidiary’s own leverage and coverage ratios, not the parent company’s (Lumine is the parent in this scenario). As of June 30, 2026:

That is up from $212.0 million at the end of 2025, almost entirely the Synchronoss financing plus a new draw on the Lumine Facility (which itself was upsized from $310 million to $360 million in June 2025 and had its maturity pushed to 2029 in April 2026, presumably in preparation for these deals). Cash on hand was $421.7 million at quarter end, up from $352.4 million at year-end 2025.

Total assets grew from $1,386.6 million to $1,739.5 million in a single quarter, almost entirely intangible assets and goodwill jumping from $727.7 million to $970.0 million. Total equity moved from $819.1 million to $848.0 million, so the acquisitions were funded mostly with debt and cash rather than diluting shareholders, consistent with the no-new-shares track record noted above.

Organic growth was flat to slightly negative for 2025, the CEO took an unexplained leave and returned within six weeks, and the CFO changed mid-year. Those are legitimate reasons for concern, but a 43% drawdown against a business that grew FCFA2S by 153% in the same twelve months is exactly the kind of inefficiency we seek to capitalize on. The market prices software names on AI anxiety without looking at the actual cash flows.

Lumine is the CSU model but on a single vertical, with the same no-dilution discipline, the same ROIC-gated bonus structure, and the same “buy and hold forever” acquisition model and capital allocation methods, like ring-fencing each deal’s debt to the subsidiary it financed. The 2026 acquisition wave (Synchronoss, Quortex, Imagine Communications) has pushed leverage up and cash conversion down in the short term, but that is par for the course, so long as they keep successfully integrating and cutting costs at the acquired businesses they will be in great shape.

Revenue has compounded at double digits per year since spin-out, maintenance revenue keeps taking a larger share of the mix, and management has now deployed over half a billion dollars in a single year without issuing a single new share to pay for it.

-CCC

Subscribe now

Subscribe now

Topicus.com Inc. (“TOI” or “Topicus”) was the first spin-out of a CSU company. Based on the success of Topicus as an initial guinea pig, Lumine was second (and as others on here have suggested at some point a 3^rd or 4^th spinout may inevitably occur as CSU continues to get bigger).

The newly formed Topicus Coop closed on the acquisition of Topicus.com B.V. for €217.4 million. That’s how the company got its name, from an acquisition it made on its second day of operations.

WHAT IT DOES

Topicus describes itself plainly in its own MD&A:

“We acquire, manage and build vertical market software (”VMS“) businesses, primarily located in Europe. Generally, these businesses provide mission critical software solutions that address the specific needs of our customers in particular vertical markets.”

This is the Constellation model transplanted onto European soil. Although they seem to be coming for North America and Australia now too!

Topicus buys small, mission critical software businesses that dominate obscure niches (a payroll system for a specific type of Dutch employer, a case management system for a regional government agency, that sort of thing) and lets them run with a long leash. Revenue is split into four buckets: licenses, professional services, hardware and other, and maintenance and other recurring. The maintenance bucket is the one that matters most. In Q1 2026 it made up €316.1 million of the €435.7 million total, over 72% of revenue.

The company aggregates its operations into three operating segments for reporting purposes, but describes each one internally as a “mini Topicus company,” essentially a conglomerate of small VMS businesses with its own operating segment CEO who is judged on returns above a hurdle rate set by head office and the board. It’s decentralization stacked on top of decentralization.

THE OWNERSHIP STRUCTURE

Constellation Software controls TOI the same way it controls Lumine. From the most recent Management Information Circular:

Constellation Software Inc. owns 40,406,513 Subordinate Voting Shares (48.5% of that class) plus the single Super Voting Share, which by itself carries 50.1% of all votes.

CSU therefore controls 74.3% of total voting rights while owning less than half of the economic interest in the subordinate shares.

REVENUE AND ORGANIC GROWTH

Here is what full-year revenue has looked like, per the audited financial statements:

2021: €742.5 million

2022: €916.7 million

2023: €1,124.97 million

2024: €1,294.9 million

2025: €1,552.3 million

That’s roughly 2.1x over four years, most of it acquisition-led but underpinned by that steady maintenance organic growth number discussed below.

Revenue growth of 20% in 2025, with organic growth of 4% for the full year. That 4% organic figure has been remarkably consistent quarter to quarter. Below are the organic numbers going back to Q1 2024 from the SEDAR+ filings:

Q1 2024: 5%

Q2 2024: 4%

Q3 2024: 4%

Q4 2024: 4%

Q1 2025: 4%

Q2 2025: 5%

Q3 2025: 3%

Q4 2025: 4%

Q1 2026: 5%

Maintenance and other recurring organic growth has been 6% or 7% almost every single quarter going back two years, the most boring and most valuable number in the whole filing. Licenses and professional services organic growth bounce around, occasionally negative, because they are lumpier and more acquisition-dependent. But maintenance being stable shows its a recurring piece that customers can’t easily walk away from, so AI has not started taking bites out of TOI just yet. As long as it keeps compounding at mid-single digits, we know prices follow earnings and eventually TOI will be back to where it belongs.

FCFA2S

Topicus uses the same free cash flow available to shareholders (“FCFA2S”) metric that Constellation and Lumine use. Myself and others that write about CSU/TOI/LMN have flagged before, but it is worth pointing out again, that this is actually a conservative, non-IFRS figure defined as cash flow from operations less interest paid, lease payments, and capital expenditures, before any capital deployed on acquisitions.

Q2 2026: € 180,000,000 (granted the Asseco dividend + Q1 payments are giving this a boost, so let’s see how the Q3 and Q4 numbers end up)

FY 2025: € 218,700,000

FY 2024: € 177,400,000

FY 2023: € 123,400,000

FY 2022: € 54,500,000

FY 2021: € 86,800,000

Q1 is always the biggest quarter since TOI’s business units mostly invoice customers for annual maintenance in Q1, cash generation is heavily front-loaded. FCFA2S in the March quarter is always much higher than the other three quarters combined. It’s a common pattern the company flags every year in its “Quarterly Results” section.

The cash goes back into acquisitions, consistent with the stated objective: “our objective is to invest all of our FCFA2S in acquisitions which meet our hurdle rate,” almost word for word the same language Constellation uses about its own capital allocation.

PEMS AND THE ASSECO STAKE

The most interesting capital allocation event post the “AI will eat software” scare has been the introduction of the PEMs strategy. Topicus was the first of the 3 CSU family members to execute a PEMs position, by building a large minority stake in Asseco, the Polish-listed enterprise software group.

Topicus started with a 9.99% position in Asseco. On February 4, 2025, it entered a binding agreement to acquire 12,318,863 additional treasury shares, representing another 14.84% of Asseco’s share capital. Regulatory approval finally came through on September 25, 2025, pushing TOI’s total stake to roughly 24.8%, high enough to require equity method accounting rather than treating it as a simple fair-value investment.

Since Topicus elected to record the investment at cost under the equity method (initial investment of €168.0 million plus €1.7 million of transaction fees, for a total of €169.7 million), it had to reverse the fair value gains it had previously booked on the smaller stake. That reversal produced a non-cash loss of €221.7 million in Q3 2025, which is exactly why Topicus reported a net loss of €120.9 million that quarter (basic loss per share of €0.94) despite the underlying business having a perfectly normal quarter operationally. Revenue that quarter was still up 24% year over year with 3% organic growth. The net loss was a paper accounting reclassification tied to a long-term strategic stake, not a sign anything went wrong in the business.

That’s a useful reminder for anyone glancing only at the headline net income line without reading the notes. It happens with these holding-company style structures more often than people expect, and it’s exactly why you have to read the actual filings.

This isn’t even the first time it’s happened. The 2021 annual financial statements show Topicus reported a net loss of €2,222.2 million on revenue of €742.5 million, a headline number that would send most investors running. The cause was a single non-cash line item: a €2,302.2 million “redeemable preferred securities expense” tied to the fair value accounting on the preferred and special shares issued as part of the original spin-out structure, before those securities were converted. The actual underlying business earned €104.4 million before tax that year, once adjusting for this line item.

THE BALANCE SHEET

As of June 30, 2026, Topicus had €323.3 million of cash against €467.7 million of debt, for net debt of roughly €144.4 million. That’s a sharp improvement from €365.9 million of net debt at December 31, 2025, driven mostly by the same Q1 seasonal maintenance billing cycle discussed above: the Topicus Revolving Credit Facility balance fell from €345.3 million to €116.5 million over the six months as cash poured in during Q1 and then held through Q2.

The debt maturity was extended during the period. On May 13, 2025, Topicus Coop amended its revolving credit facility to a single €700.0 million multicurrency line maturing October 28, 2029, replacing the older, smaller structure. On top of that, Topicus Coop layered on a new €200.0 million senior unsecured Schuldschein Loan on June 30, 2025, split across maturities in 2028, 2030, and 2032.

The cost of that debt is rising, though. Interest expense on debt for H1 2026 was €14.3 million, up 48% from €9.7 million in H1 2025, reflecting both a higher average debt balance (from the Asseco purchase and other 2025 acquisitions) and the current rate environment. Finance costs overall increased 48% for the six-month period. Worth watching if the company keeps layering on debt to fund larger, less traditional positions like Asseco.

Total assets were roughly flat over the half, moving from €2,513.2 million at December 31, 2025 to €2,515.2 million at June 30, 2026, with a €36.2 million decline in the Investments in associates balance (the Asseco dividend reducing the carrying value) roughly offset by increases in unbilled revenue and other assets. Total shareholders’ equity, including non-controlling interests, grew from €781.8 million at the start of the year to €884.4 million at June 30, 2026.

COUNTER ARGUMENTS / THINGS TO WATCH

A few things worth flagging, as stated in TOI’s filings:

The Company disclosed an ongoing tax audit by the Dutch Tax Authorities challenging the deductibility of its employee bonus program for financial years 2016 onward. As of the most recent disclosure the potential additional tax exposure, including interest, was estimated at up to roughly €8 million. Not enormous relative to the size of the business, but worth watching since it touches the compensation structure that underpins the decentralized model.

Organic growth outside of maintenance (licenses, professional services, hardware) is genuinely lumpy and occasionally negative on a quarterly basis. None of that is unusual for a serial acquirer of small software businesses, but it means the total organic growth number can wobble more than the maintenance line alone would suggest.

The dual-class, super voting share structure means minority shareholders are, by design, along for the ride. That’s the same structure that has worked out extraordinarily well for Constellation shareholders over two decades, but it’s worth being clear-eyed about what it is.

Topicus is Constellation’s European chapter written by many of the same authors, using the same strategy of buying small mission critical software companies, leaving them decentralized, measuring everyone against a hurdle rate, and letting maintenance revenue compound in the background while headline numbers get moved around by acquisitions, currency, and now the new PEMS strategy / big stake in Asseco (too bad Readytech didn’t become the second TOI test case for PEMS ☹)

Multiples: In the 2022 annual financial statements, Topicus disclosed it used valuation multiples “ranging from 1.5x to 6.0x of recurring revenue” to test its cash generating units. In the 2023 annual statements, that range had widened to “1.5x to 8.0x.” and 1.5x to 8.8x in 2024. In 2025 they stopped publishing the multiple. Multiples paid for VMS assets had been creeping up, it will be interesting to see if they start publishing the multiples again should AI factors start impacting private VMS companies like Roper expects.

The maintenance organic growth number is the one management has suggested to keep watching. It’s been sitting in the 6-7% range for two straight years, and that’s the number that tells you whether the underlying customer base is sticking around and paying more over time.

More to follow…

-CCC

Subscribe now

Yes revenue has not been growing at the same pace, but it is still growing in the mid to high teens. For a company of Constellation’s size that is pretty incredible.

With the amount of capital CSU has been deploying lately, the revenue numbers might flip back to 20%+ growth rates.

FCFA2S is CSU’s non-IFRS measure. They define it as:

“Free cash flow available to shareholders “FCFA2S” refers to net cash flows from operating activities less interest paid on lease obligations, interest paid on debt, debt transaction costs, payments of lease obligations, the IRGA / TSS membership liability revaluation charge, and property and equipment purchased, and includes interest and dividends received, and the proceeds from sale of interest rate caps. The portion of this amount applicable to non-controlling interests is then deducted. We believe that FCFA2S is useful supplemental information as it provides an indication of the uncommitted cash flow that is available to shareholders if we do not make any acquisitions, or investments, and do not repay any debts. While we could use the FCFA2S to pay dividends or repurchase shares, our objective is to invest all of our FCFA2S in acquisitions which meet our hurdle rate.FCFA2S is not a recognized measure under IFRS and may not be comparable to similar financial measures disclosed by other issuers. Accordingly, readers are cautioned that FCFA2S should not be construed as an alternative to net cash flows from operating activities. See “Results of Operations —Free cash flow available to shareholders” for a reconciliation of FCFA2S to net cash flows from operating activities.”

Consistent with management’s ethics, it is actually a more conservative figure compared to just the usual free cash flow (for example the Globe has 878M as the Free Cash Flow figure).

As an aside that is why you must always read the actual source documents provided by the company and filed on SEDAR+ to learn the full picture.

Many writers and commentators have made the same observations, but it is worth emphasizing them again, especially the 2 that really pop out.

MANAGEMENT

CSU has never once diluted shareholders. Since 2007, it has never once issued stock based compensation.

1. As at March 5, 2008, there were 21,191,530 total shares outstanding (at the time comprised of 16,903,530 common shares and 4,288,000 class A non-voting shares) (2008 Q4 MD&A)

2. While some were converted earlier in 2009, on April 3, 2012, 100% of the Class A non-voting shares were converted to common shares, on a one for-one basis. As at May 2, 2012, there were 21,191,530 common shares outstanding. (2012 Q2 Financials)

3. As at March 31, 2026, there were 21,191,530 common shares outstanding (2026 Q1 MD&A)

Instead, it forces management to buy shares in the open market.

From the 2019 Annual Information Form (“AIF”):

“The objective of our compensation plan is to reward employees for working towards our corporate goal of increasing shareholder value. We believe that shareholder value is created by managing two financial components over the long term: profitability and growth. As such, our bonus plan compensates employees at many levels of our organization based upon the profitability and revenue growth of their operating group or business unit, as applicable. The long-term focus is accomplished by mandating that at least 25% of the incentive compensation for the majority of our senior employees who earn in excess of $75,000 per annum and have bonuses in excess of $10,000 per annum be reinvested in shares of the Company that are subject to restrictions on resale for a period of three to ten years. At a minimum, these restrictions require employees to hold 100% of their shares for the first two years following acquisition, and then one third of such shares may be sold in each of years three, four and five. Senior executives are required to invest 75% of their bonus in shares of the Company that are subject to the same restrictions on resale for a period of three to ten years. Once every five years, employees may elect to receive 100% of their bonus in cash. Our bonus plan encourages employees to participate through share ownership in the value that they have created.”

Do you realize and understand how rare that these values and lack of dilution are in today’s market?

What does that say about management’s commitment to shareholders and alignment with them?

2007 Mark Leonard Quote:

“I recently flew to the UK for business using an economy ticket. For those of you who have seen me (I’m 6’5”, and tip the non-metric scale at 280 lbs.) you know that this is a bit of a hardship. I can personally afford to fly business class, and I could probably justify having Constellation buy me a business class ticket, but I nearly always fly economy. I do this because there are several hundred Constellation employees flying every week, and we expect them to fly economy when they are spending Constellation’s money. The implication that I hope you are drawing, is that the standard we use when we spend our shareholders’ money is even more stringent than that which we use when we are spending our own.”

Since 2014 Mr. Leonard voluntarily waived his entitlement to receive a salary and a bonus. (See Management Information Circulars (“MIC”) from 2015 – 2026).

Mark Miller was appointed President of the Corporation on September 25, 2025. During 2025, Mark Miller received his previously agreed to compensation for his role as the Executive Chairman of the Volaris Operating Group. Mr. Miller has voluntarily waived his entitlement to receive a salary and a bonus commencing on January 1, 2026. (See 2026 MIC)

At the Annual General Meeting (“AGM”) Mark Miller gave every indication that will continue into the future.

“I’m just continuing what Mark Leonard did, I want to help and do what’s best for shareholders.” – Mark Miller 2026 AGM

This just reinforces the above points. Management cares deeply about the company.

Management wants the company and its shareholders to succeed.

How many other companies with this kind of management team exist in the market?

REVENUE

CSU used to have a Q&A with shareholders. Below is a question and answer from August 17, 2020 Q&A.

https://www.csisoftware.com/wp-content/uploads/2026/04/qa-august-2020-final.pdf

Note this question arose during a period where CSU was experiencing negative organic growth as a result of covid related issues. See below from the 2020 Q2 MD&A.

The shareholder was concerned about the low organic growth rate:

“So when I see an aggregate organic growth number that seems low to me, I have to wonder, is it where it is because most business units are doing poorly lately (and if so, is it temporary or a more long‐term inflection in the wrong direction?), or is it because you bought more distressed and shrinking assets lately (which can still create a lot of value if you paid a low enough price for them) that mask the good results of the growing business units? So that’s the problem.”

Jamal Baksh, CFO responded by stating:

It should be noted that maintenance organic growth has been 4% in each of 2016, 2017, 2018 and 2019. If the mix of shrinking businesses were increasing it would be unlikely that the maintenance organic growth percentage would remain so stable. By focusing on maintenance organic growth, you should be able to flag any concerns around what’s going on beneath the surface.”

In 2020 it was a different concern, but the point of focusing on maintenance organic growth is still a good one, as maintenance revenue has grown from being 58% in 2007 to just under 75% in 2025.

The concern today is AI is going to destroy CSU’s revenues, customers will replace it (the weakest argument), or demand discounts on fees, reduce seat count, or otherwise use AI, or use AI to negotiate ways to reduce their annual maintenance fees. That is not playing out in the numbers.

Look at the last 36 quarters or 9 years below and then the year over year comparisons.

Maintenance organic growth has been remarkably consistent, generally between the 4-6% on a quarterly basis, and from 1-8% on an annual basis from 2025 – 2016 (6%, 6%, 7%, 1%, 8%, 1%, 2%, 5%, 4%).

Before 2016 they didn’t use to have the organic numbers in a chart but 2015 was -3%, 2014 was 4%, 2013 was 4%, 2012 was 1%, 2011 was 10%, 2010 was -4%, 2009 was -3%, 2008 was 5% and 2007 was 1%.

I think it is safe to conclude there isn’t something negative going on beneath the surface, if anything as they start implementing and improving their customer experience and software with CSU developed AI, and start to harvest and harness the vast amount of customer data they have to become even more ingrained in their customer operations, the maintenance organic revenue growth should continue to be stable and may even start growing.

More to follow…

Given I share screenshots of my trades and portfolio (see below). I wish to remain anonymous.

However, I assure you, I am a real human. A real Canadian, with a passion for investing.

Like many human beings, I am a walking, talking, living, contradiction. On the one hand, I believe strongly in the arguments for indexing and preach it widely to family and friends, as I believe for the vast majority of people that it is the safest and best way to accumulate wealth and save for retirement.

However, I have been a long time admirer of Buffett, Lynch, and more recently Trevor Scott at Tidefall Capital, Dr. George Athanassakos, the value investing professor at Western/Ivey, Prem Watsa at FairFax and Norman Rothery and his columns in the Globe, among other value investors.

For my own personal portfolio, I do not index.

I believe in a highly concentrated portfolio of high quality businesses, businesses I understand, and businesses I have a high conviction will continue to produce copious amounts of free cash flow, and will continue to compound for decades to come.

In my case, I find I agree heavily with the quote Dr. Athanassakos often refers to in relation to value investors and their beliefs: “markets are not efficient in the short run, but the tend toward efficiency in the long run.”

I believe those short term inefficiencies can be exploited.

I have never seen such a wonderful opportunity to exploit these inefficiencies as I see today; which brings me to Constellation Software Inc. (CSU), and its spun out subsidiaries Topicus.com Inc. (TOI) and Lumine Group Inc. (LMN)

Much has been written on substack and reddit about CSU/TOI/LMN and why they will be successful, but it seems the market remains stubbornly irrational and refuses to be convinced that CSU will be an AI winner.

I think this is a mistake. As you can see below, I have almost $400k invested between CSU, TOI and LMN, (100 shares CSU, 1,111 shares TOI, and 1,000 shares LMN) and any spare capital I have, I plan on allocating towards these 3 amazing companies, because I am convinced the market does not understand the power of these companies, that the vast majority of investors have behaved like sheep in their hurry to dump anything software related, and that over the long term, I will become incredibly wealthy from exploiting these inefficiencies.

In the coming days, I will do my own write-ups as to why, as I find it helps reinforce the resolve and test the thesis. Like any good argumentative essay, the arguments in favour and against must be considered before a rational, reasonable choice can be made.

More to follow.

- CCC

P.S. I am not creatively inclined and in an effort to scrub any identifying information from the screenshots, I have merged them into a pdf and then taken a screenshot of all 4, so the quality is crude.

Subscribe now

Leave a comment