说话者承认过去对FIX的看跌/谨慎立场是错误的，因为低估了人工智能数据中心的需求；该股大幅跑赢其320美元的公允价值估计。

what we could see to the downside then we can certainly also see to the upside in companies like Dell or actually the next company that we'll talk about which is Comfort Systems.

And it's kind of funny because both Dell and also Comfort Systems probably even more Dell they were pretty boring pitches. I still remember us sitting and discussing Dell and thinking gosh this is one of the least inspiring companies that we have yet looked at and now those are the companies that just completely crushed it.

I was really shocked when you told me just how well Comfort Systems has done in the last year because I I haven't really been tracking it that closely after we set it aside. But then I realized of course it has to do with AI. And so for context, Comfort Systems is a serial acquirer, kind of a Birkshshire Hathaway style decentralized holding company, which takes the approach of buying small regional contracting businesses and then allows them to keep running things while they handle the back office and the capital allocation.

And so it was really a mature looking business. if you were simply looking at organic growth. However, even in early 2025, you could actually already see that there were some early signs of these AI data centers and semiconductor factories becoming a bigger and bigger share of the company's business because comfort systems is actually uniquely specialized in being able to provide HVAC services to complex centers like data centers.

So I don't think I have any regrets. Not because I don't like making money, but more because the factors shaping this business are so macro and I of course have very little confidence in my ability to predict the future and especially not with something like AI. So clearly I did misjudge the stage of the cycle that we were in as I tried to be cautious, right? I wanted to to really hedge against the possibility that we were nearing the end of the AI spending cycle, which it seems like we absolutely have not.

And if that had occurred and when it does eventually occur, Comfort Systems will take a double hit from revenue growth declining and then the multiple declining too as a response to that. So, I didn't think it was cheap at a 25 times multiple, but I can't lie, looking at the price today, it does feel a little surreal. Uh, I did my fair value estimate at around $320 per share, and, you know, I proudly said with a margin of safety, I would be happy to buy shares at maybe $290 or below. And, uh, the stock is trading at $1,700 today.

But looking at comfort system, it's just funny. I mean, to quote the investment decision from back then, between $320 and $340, you're probably just going to get an average return. and any price much higher than that risks earning a below average return. Well, 15 months later and a 5x later, it's safe to say that shareholders did quite well.

I would probably guess that Comfort Systems has gotten a little ahead of itself, too. But I don't want to be wrong twice now.