Podcast Summary: Chit Chat Stocks – Will AI Kill Constellation Software? With Drew Cohen From Speedwell Research (Ticker: CSU)
Date: February 11, 2026
Host(s): Ryan Henderson, Brett Schafer
Guest: Drew Cohen, Founder of Speedwell Research
Overview
This episode dives into Constellation Software (CSU), a Canadian software conglomerate known for its legendary capital allocation and decentralized structure. The company is facing its biggest drawdown ever—down 50% from highs—prompting concerns that advancements in AI could disrupt its business model. Host Ryan Henderson, with guest Drew Cohen from Speedwell Research, explores whether these AI-driven fears are justified, examines Constellation’s competitive advantages, discusses leadership transition risks, and unpacks valuation, capital allocation, and culture.
Key Discussion Points & Insights
1. Constellation’s Sharp Drawdown and the Role of AI
- Current Situation:
- Constellation Software is in an unprecedented drawdown, with the stock down 50% from recent highs ([00:51]).
- The sell-off is entirely narrative-driven: “AI, that's it in a nutshell and everyone is concerned that they have no idea what this means.” – Drew ([01:54]).
- Financials currently show no deterioration; concerns are hypothetical and future-oriented.
2. Constellation’s Business Model: A Refresher
- Focuses on acquiring niche vertical market software companies serving small, fragmented industries (e.g., cemetery operators, chicken coop management, dental offices) ([03:21]).
- The strategy: Buy companies with limited organic reinvestment potential, extract cash flows, and redeploy into new acquisitions.
- Run as highly decentralized groups—autonomy and founder retention are key.
- “It is ultimately more about pulling the cash flow out of the business and putting it elsewhere, rather than reinvesting in a massive TAM opportunity…” – Drew ([04:45]).
3. Bear Case: Why AI Is Seen as a Threat
- Primary Market Fear:
- AI enables extremely rapid, cost-effective software creation.
- The concern: Cheap, custom AI-generated software could replace Constellation’s legacy products, even to the point of full AI agents automating end-customer tasks ([08:18]).
- Drew’s Counterpoints:
- Business moats extend beyond product quality: distribution, customer inertia, service/support, and criticality of software to operations add significant switching costs ([09:33], [10:56]).
- “Whenever a business builds a product, having a better product than a competitor is not sufficient enough to build a business…” – Drew ([09:33]).
- Much of CSU’s software is so old and on-premise that even before AI, it was already technically replaceable, but customers rarely switch due to mission-critical risks and low appetite for disruption ([13:00]).
- Customization and integration needs, retraining costs, and complexity deter replacement by AI solutions.
4. AI Agent Risk: A Real Concern?
- AI agents that could automate end-user tasks or interface directly with data/software represent a hard-to-quantify risk ([17:22]).
- “It's a little harder for me to really understand how that could all play out...there's just a lot of open possibilities as to what is possible. And so all of that is scary.” – Drew ([17:34]).
- However, trust and readiness for businesses to hand over critical processes fully to AI is very low, at least for the next several years.
5. Other Bear Cases & Key Risks
- Tail risks are inherent in technology. “There’s a reason why Warren Buffett doesn’t invest in technology companies…” – Drew ([17:22]).
- Potential for accelerated churn or industry transitions if AI rapidly becomes viable for CSU’s customer base ([19:30]).
- “These are not growth businesses...some already have questionable terminal value (e.g., software for linear TV stations).” – Drew ([19:30]).
6. Evolution in Acquisition Strategy
- Transitioning to larger acquisitions as small vertical-market software deals become less plentiful ([22:46]).
- Still enjoys some competitive insulation on mid-market dealflow, finding special situations and carve-outs (e.g., Optimal Blue, Allscripts).
- New moves include minority investments (e.g., Seco, a Polish IT company), and possible explorations into verticals like healthcare IT, or even non-software areas ([22:46], [26:23]).
7. Management Transition: Mark Leonard’s Departure
- Mark Leonard retired abruptly for health reasons, replaced by long-serving deputy Mark Millard.
- No significant cultural change expected due to highly decentralized org structure and deep internal promotion ([27:06]).
- Potential risk: “Where I can see him having the most impact is them going into new verticals...” – Drew ([27:42]).
8. AI Will Likely Expand—Not Shrink—The Acquisition Universe
- Lower barriers (AI allows faster, cheaper software creation) might spark more start-ups, which could boost CSU’s pipeline ([29:31]).
- Using AI, incumbents also expand into new functions, creating cross-vertical competition, but Drew sees ecosystem “consolidation” as likely only at the margins.
9. Valuation & Performance
- CSU now trades at around 20–21x free cash flow, with a long history of 15%+ FCF growth at 20%+ ROIC ([32:31]).
- Adjustments: Drew adds back the IRGA liability related to Topicus as a one-time event ([32:40]).
- “If you were able to basically own this whole business today ... the reverse DCF would show you low double-digit, mid-teens, high-teens [annual] return depending on how much you want to assume they're able to reinvest that capital.” ([33:40]).
- The likelihood of buybacks is higher in the future but remains a rarity given CSU’s approach and high hurdle rates ([36:00]).
- Stock underperformance over 5 years with flat earnings would be a major red flag, likely implying that the AI bear case or elevated churn had come true ([39:41]).
10. What Most Investors Miss
- “Just because you have a product doesn’t mean you have a business.” ([40:48])
- Drew uses the Friendster/Facebook anecdote to illustrate how re-platforming and operations, not features, frequently determine winners.
11. Constellation’s Culture: What Sets It Apart
- A true meritocracy with forced employee stock ownership (from bonuses) creates strong alignment ([43:23]).
- Competitive, decentralized structure nurtures careful capital allocation and return focus.
- Persistent experimentation (e.g., organic vs. acquisition growth tests) and intellectual humility characterize leadership and communications ([43:23], [45:54]).
- “Having skin in the game isn’t just participating in the upside, it's participating in the downside too. And that changes the decision framework ...” – Drew ([47:35]).
Notable Quotes & Memorable Moments
- On AI replacing CSU’s stack:
- “Even before the AI, a single engineer could recreate these tools. But it wouldn’t succeed because you won't be able to get [the customer] to switch … It’s not a good risk-reward.” – Drew ([11:05])
- On Decentralized Ops Culture:
- “They try to keep as much of the original employees there as possible. Not dismantling the business, which is part of the selling point…” ([07:03])
- On Investor Overreaction:
- “It's somewhat a case of the baby being thrown out with the bathwater before kind of figuring out and thinking through what the actual ramifications would be.” ([01:54])
- On Real Bear Case:
- “You have to have humility whenever you're dealing with the future in something as open ended as AI ... Warren Buffett doesn't invest in technology companies for a reason...” ([17:22])
- On Culture:
- “A strong meritocracy with a lot of buy-in because employees are forced to buy stock ... creates a sense of ownership, skin in the game, and that changes the decision framework.” ([43:23], [47:35])
- On Forced Ownership Model:
- “I forget the exact stat...but Constellation Software created more ... over a thousand millionaires or something like that because of this forced stock buying and the business doing really well.” ([47:53])
- On Business Moats:
- “Business isn't just about the product. It’s the sales, the service, the switching costs, the distribution ... all these other factors.” ([09:33], [41:00])
Key Timestamps
- 00:51: Intro and CSU’s historic drawdown
- 01:54: Why is it selling off? The AI bear narrative
- 03:21: Explanation of business model and deal sourcing
- 07:03: Discussion of post-acquisition integration and decentralization
- 08:18: Bear case: “AI can disrupt everything”
- 09:33: Moats beyond product quality
- 13:00: Mission-critical nature and customer switching aversion
- 17:22: Real AI risks, tail risks, and humility
- 22:46: Evolving acquisition size/strategy
- 27:06: Mark Leonard’s retirement, succession
- 29:31: Will AI expand acquisition target pool?
- 32:31: Valuation and performance metrics
- 36:00: Potential for stock buybacks
- 39:41: What would cause CSU to flatline?
- 40:48: “Just because you have a product doesn’t mean you have a business.”
- 43:23: Culture, employee ownership, and why it’s durable
- 47:35: Skin in the game and impact on decision-making
Conclusion
While the AI narrative has caused a dramatic shift in sentiment and stock price for Constellation Software, fundamentals remain robust. The company’s deep-seated moats—distribution, mission-criticality, and extreme customer inertia—protect it from rapid disruption. Management transitions and the exploration of larger deals or new verticals pose meaningful risks, but the culture of careful capital allocation and true ownership alignment persists beneath the surface drama. Drew and the hosts suggest that, for long-term investors, CSU’s risk/reward is once again compelling, barring an unforeseeable leap in AI that rapidly reshapes the software landscape—an outcome they see as possible but not probable in the near future.
Drew Cohen’s Plugs: ([48:58])
- Research: speedwellresearch.com
- Free resources: speedwellmemos.com
- YouTube: “Drew Cohen Money”
- Free newsletter: drewcohenmoney.com (“5 Minute Money”)
Note: The summary above is faithful to the original tone and content, omitting advertisements and non-content sections.
