Variant Perception

Where We Disagree With the Market

The market is treating the €385M Asseco Poland stake as a Constellation-style playbook extension — the "PEMS" (Permanent Engaged Minority Shareholder) framing that has propagated across substack, sell-side notes, and a unanimous 3-Buy / 0-Hold / 0-Sell analyst consensus with a +56% mean price target. The evidence in this report reads it differently: Topicus diverted €385M — 176% of FY25 FCFA2S of €219M — into a 23.14% minority position in a listed Polish IT-services company that is not VMS, not private, not bolt-on, and not 100%-owned, in a year when ROIC fell from 15.8% to 10.4% and FY26 Q1 VMS deployment collapsed to €22.5M. That is what forced deployment looks like when the addressable European bolt-on pipeline has run out of capacity at the 5–6× EBITDA / ~20% IRR hurdle. The variant view is not that the stock is "too cheap" or "too expensive" — it is that consensus is paying for a CSU-equivalent multiple on the wrong denominator: €402M of headline FCF rather than the €120M of FCF-after-acquisitions that actually accrues to owners after the engine has refuelled. Both variants resolve inside six months: FY26 H1 cumulative VMS deal spend (visible by the Q2 print in early August) and the AGM scoping of Asseco intent tomorrow.

All figures in this file are in euros (Topicus' functional currency). The USD version of every number sits in the sibling file variant-claude-USD.md.

1. Variant Perception Scorecard

The score reflects three things. Consensus is unusually visible for a TSXV small-mid cap: three covering analysts, all Buy, mean PT C$143 (+56%), with substack and Seeking Alpha bull cases reinforcing the same frame. Evidence is strong on the headline-cash and Asseco-strategic-signal disagreements (both rooted in audited disclosures), but weaker on the multiple-anchoring view because it depends on a forward judgment about ROIC trajectory. Resolution is fast: the AGM is tomorrow, the Q2 print lands in ~80 days, and Q3 closes a clean YoY GAAP comp by early November.

2. Consensus Map

The map matters because three of the six issues — Asseco intent, ROIC mechanics, and the CSU-multiple anchor — are linked. If Asseco is forced deployment rather than opportunistic extension, then ROIC drift is not mechanical; it is the early signal of a structurally lower-IRR future. Once both fail, the CSU-anchor on the multiple loses its foundation. The variant view exploits that linkage.

3. The Disagreement Ledger

Disagreement 1 — Asseco is forced deployment, not playbook extension

Consensus would say the Asseco stake is a tactical extension of the Constellation playbook into engaged-minority investing, validated by CSU's own SABRE move, and that the €385M earns roughly the same IRR as a VMS bolt-on because TOI got an entry-price discount versus the WSE quote. The evidence disagrees on cause-and-effect: TOI redirected 176% of FY25 FCFA2S into a public-market minority of a Polish IT-services company in the same year that VMS deployment ran below the long-run trend and ROIC dropped to a five-year low; six months later, Q1 26 deployed only €22.5M. That sequence is the firm telling you, through capital allocation rather than words, that the addressable European bolt-on pipeline at 5–6× EBITDA / ≥20% IRR can no longer absorb a full year of FCFA2S. If we are right, the market has to concede that the next decade compounds at incremental IRRs closer to 12–15% rather than 20%+, and the CSU-equivalent multiple loses its analytical floor. The cleanest disconfirming signal is FY26 H1 cumulative VMS deal spend ≥€100M at inferred 5–7× EBITDA, visible at the Q2 print in early August.

Disagreement 2 — The headline cash machine overstates owner economics by ~3×

Consensus would say €402M of FCF and €219M of FCFA2S are the correct denominators; acquisitions are discretionary growth investment, not operating cost, and the proof is that TOI can pause M&A and return cash (as it did with the €208M FY24 special dividend). The evidence disagrees: for a permanent serial acquirer where the moat is the acquisition engine itself, M&A is not discretionary — pause it for two consecutive years and you become Enghouse, with a 4× EV/EBITDA multiple. The five-year cumulative FCF-after-acquisitions is just €480M (36% of headline FCF), and FY25 deployment was funded by €442M of net new debt plus the inaugural €200M Schuldschein, not by cash conversion. If we are right, the market has to concede that the right valuation denominator is closer to €120–150M of FCF-after-acquisitions, which makes the current multiple ~3× more expensive than the headline 12× P/FCF implies. The cleanest disconfirming signal is FY26 net debt trending below €200M while VMS deployment runs inside trailing FCFA2S — i.e., the cash conversion picks up the funding load without leveraging the balance sheet further.

Disagreement 3 — The CSU-equivalent multiple is not yet earned

Consensus would say TOI's playbook and its margins are CSU's playbook and margins, so the trading multiples should converge — and indeed they have (TOI 11.5× vs CSU 12.6× EV/EBITDA). The evidence disagrees on what generates the CSU multiple: 25 years of disclosed capital discipline, IG-adjacent debt access, the Mark Leonard letters that anchor every analyst model, and a controlling shareholder that does not deviate from the playbook. TOI has four years of independent track record, just deviated from the playbook with Asseco, has a CEO who is also Chairman, sits on his own comp committee, and took a second CEO role at Your.World in May 2024 without subsequent disclosure. The moat heatmap shows TOI tied with or below CSU on five of six dimensions and tied on one. If we are right, the market has to concede that the right multiple sits between CSU (12.6×) and Enghouse (4.3×) — roughly 8–9× EV/EBITDA, implying a fair value closer to C$55–65 than the consensus C$143 target. The cleanest disconfirming signal is AGM commentary tomorrow that scopes Asseco as terminal and the CEO role at Your.World as time-bound, combined with a Q2 print that re-anchors VMS deployment at hurdle.

4. Evidence That Changes the Odds

The most important evidence row is #2 — the €480M five-year FCF-after-acquisitions versus €1,347M reported FCF. It is the cleanest re-framing of valuation arithmetic anywhere in the report. Every other row is a signal; this row is the denominator change that would force a re-rate even without any operational deterioration.

5. How This Gets Resolved

The matrix concentrates resolution into the next ~90 days. The AGM tomorrow is the only opportunity all year for a real signal on Asseco intent and CEO dual-role; the Q2 release in early August carries the cleanest cumulative deal-cadence and maintenance-organic data point. Items 3–6 update through Q3 and Q4 — slower-moving but mechanical.

6. What Would Make Us Wrong

The most honest test of the variant view is to name what would break it before the market does. Three pieces of evidence would change our mind.

The first and largest is a clean FY26 H1 acquisition vintage. If Topicus announces ≥€100M of VMS bolt-ons between now and the Q2 print at inferred multiples of 5–7× EBITDA — and the AGM Q&A frames Asseco as a permanent, terminal minority position with no further deployment plans — then our central claim that Asseco is forced deployment becomes untenable. The Q1 26 lull would be re-read as a deal-timing quirk in a deliberately lumpy cadence, not as the pipeline running out of capacity. We are not 50/50 on this; we are roughly 35/65 against. But a deployment surprise of the right size and shape would force a re-rate of our variant view inside one quarter, and that is exactly the cleanest disconfirming evidence we can ask for.

The second is sustained ROIC recovery in FY26. Our framing depends on the FY25 ROIC drop from 15.8% to 10.4% being a structural early signal, not a mechanical artifact. If the FY26 audit prints ROIC back at 13–15% while Asseco carrying value compresses through equity-method income absorption, the bonus formula does its job, and finance costs roll over with the Schuldschein refinancing not being needed — then the multiple anchor to CSU survives and our gravity-point claim (8–9× EV/EBITDA, fair value ~C$55–65) is wrong by half. The variant view rests on the second derivative of returns, and a positive second derivative dissolves it.

The third is AGM commentary that explicitly scopes both governance items. If management uses the AGM to (i) name Asseco as a "permanent engaged minority with no incremental deployment intent," (ii) characterize the CEO's Your.World role as time-bound or transitional, and (iii) introduce a credible succession plan or named COO at the Topicus level — then the governance discount that our variant view leans on collapses inside 24 hours. The base rate for management to use a low-coverage proxy meeting to resolve a strategic-ambiguity question is low (this is a press-release-only company), but the format change to hybrid in-person Q&A for the first time in company history suggests it is less low than the base rate would imply.

We are also wrong if maintenance-line organic growth holds at 6%+ for two more quarters AND the engine still re-accelerates. That is the floor scenario where the customer moat is the load-bearing wall and the acquirer-side disagreement is irrelevant: at 6% maintenance organic on a €1.1B base, plus price escalators and bolt-ons, FCFA2S compounds in the high teens regardless of Asseco. We do not assign this floor scenario much weight because the FY25 evidence runs against it, but it is the cleanest path to "operating engine carries TOI through capital-allocation noise."

The first thing to watch is the AGM tomorrow — specifically, whether management uses plain English to scope Asseco as terminal and the CEO's dual role as transitional, or whether the meeting produces another year of strategic-ambiguity on the largest capital decision in company history.