Constellation Software Inc. (CSU.TO)

Sharp drawdown on the AI-killing-software train, now starting to perk up. Second-biggest tech company in Canada, behind SHOP. Revenues are sticky, though growing slowly organically. Exceptional in deploying capital for acquisitions -- slowed down in last years, but reaccelerated in latest quarter. 

New initiative to take strategic long-term stakes in publicly traded software companies. Yield is 0.19%.

He just bought. It buys software businesses, and the price for those is a lot lower than a year ago. Doesn't agree that customers will rip and replace products with AI, especially with no tech support. One of the most interesting risk/reward opportunities today.

Historical valuation was in the 30s; today it's in the teens.

Trying to stabilize after the bigger downtrend. Hardware is working, but software's in a downtrend. Tradeable bounce, but he'd rather wait and catch a confirmed uptrend. He's worried of another leg lower after any bounce.

Another software chart in decay. Incredible capital allocator for so long, but AI is taking its toll. Even though some headline beats, the most important metric here is organic revenue growth (and it's continually shrinking).

Mea culpa, not a timely pick. The AI bear was let out of the cage almost immediately after this pick. Still holds because his firm just doesn't believe AI will totally disrupt these vertical-market software firms.

Demise of software is a narrative right now, but no evidence in the fundamentals. Not all of its business units will be displaced by AI, and some will actually benefit. Tends to buy companies at low multiples and earn back its money in short order.

He continues to add. Everything that could go wrong did last year. Nothing wrong with fundamentals.

Looking at the chart, he exited his remaining position at the plateau at $3500. Technically, it has to get back to that and that's $900 away (a lot of water under that bridge). 

Tough over next 6 months. Need a couple of quarters under new leadership. Still expensive at 74x PE. Business model needs to do bigger deals, which means more risk; though a lot of software stocks are on sale (but for good reason).

The dip now is a phenomenal opportunity. Metrics haven't been this low in a decade. LLM's like Anthropic are chasing big companies like CSU, which will help LLM's customers to integrate with their product. The moat around CSU is misunderstood.

Vertical software segments such as transit system scheduling or golf course management. Obvious fears of AI, and there is some legitimacy there. Its advantage is that it’s had to resolve all these “edge” issues over the years -- all the little exceptions to the rules that crop up. A new company would have to start from scratch on that front. 

New strategy of acquiring smaller stakes in larger companies -- acts as a defence against AI. Trades at 15x forward PE, cheapest in 13 years. Yield is 0.23%.

He sold it when the relative strength came off, and has since crumbled a lot. It's trying to stabilize now. He's not convinced it will rise. Given the strong sell-off, it may need to go sideways for a while.

Rollup stories tend to run out of steam at some point. Double whammy -- high valuation + concern that software is going away. Probably will have a bounce, but he can't say when, and it probably won't get back to previous levels.

Ignore the noise. Are so well positioned. Software companies are cheaper now. Barriers for entry for AI are high, he feels. CSU trades at attractive valuations. Buy for the long term.