Skip to content

AI Investing Lessons: Why Revolutionary Technologies Can Still Be Bad Investments

Tuesday, November 17, 2026

During their recent episode of the Value Options Podcast, Travis and Carlisle discussed AI Investing Lessons: Why Revolutionary Technologies Can Still Be Bad Investments. Here's an excerpt from the episode:

[Tobias]

I've been writing a white paper, which I'll have out in the not too distant future, where I just went and looked at historical tech revolutions and what happened through those historical tech revolutions. The one that I remember most clearly is 1999, the .com 1.0, when there were a lot of these profitless .coms. It was clearly a new style of business where you could have that website. For a while, it was like just having a website made you a .com.

Now, it's kind of like that's just table stakes. There's not a business in the world. Every lawn mowing business has a website where you can reach out to somebody and book your lawn getting mowed.

So that's not the technological leap. But at that time, Buffett came out and he said, it could be a world changing technology and still be a terrible investment. And what really matters is durable competitive advantages.

And I think that's really the focus of this white paper that I've done. I just took his idea and then I've worked through some historical tech revolutions. And as you say, there's railways was one that was sort of transformational.

Electricity was transformational. Hard to imagine now that electricity is such a huge technological advancement. But for a long time, even manufacturers were resistant to electricity and they hadn't organised their shop floors in a way because they were used to using steam engines where you'd have one big steam engine that drove everything on the line and then electricity made it possible to put it in a different kind of modular fashion.

And that took a few decades to kind of work its way into factories just because it takes a long time to start a new factory. But we're in a totally different world now where it's software really that drives the world. And that's where AI is native to the software.

But I still think that some of those, there are lessons you can take away from that. And one of them is that there are these sort of different layers of the technology. There's the infrastructure layer and there's a little bit of a, we're not entirely sure, is this a brand new technological revolution or is this just a continuation of the telecom computing?

And that Carlotta Perez, who's one of the people I massively have a huge amount of respect for and her ideas, and I sort of discussed them a little bit in this white paper. One of the things that she said is that this is just a continuation. It's not a brand new tech revolution.

Not that that really matters, but other than just for thinking about where we are in the infrastructure layer. So there's an infrastructure layer, which is the fibre optic cables. And as everybody well knows, the fibre optic build out led to some bankruptcies, even though ultimately that was the thing that created the possibility of running really fast internet over fibre so much faster than twisted pair.

And then there's a builder layer that goes on top of that. And you can think of that as being sort of Google and those sort of internet native businesses that wouldn't exist except for that underlying infrastructure. And then there's a retail layer that goes on top of that, that uses Google as part of its business.

And so the challenge, I think, for this revolution is to work out whether Anthropic and OpenAI are that builder layer or that retail layer. And I can see a pretty strong argument for them being maybe a little bit of a hybrid, maybe a little bit in both because you can certainly get a lot of functionality out of them. As you say, you can just ask it to build what you want.

But my question is, do they really want to be that retail layer or will it be like electricity or like the internet where every business is going to have AI built into its business and it's probably going to be supplied by one of those labs? Although there's always going to be a limit on what they can charge because the functionality of those free AIs will always keep them kind of honest. If you need the bleeding edge AI because you're doing bleeding edge research, then you need those ones.

But if you can do with just a little bit less functionality, maybe use a free version.

[Tim]

I mean, exactly. I mean, there's so many similarities in the different AI products now. And then, I mean, we're not even talking about like China.

But if you look at what Meta has done with Muse and how that's kind of more retail related version, but now they're getting more into the business aspect of it, it doesn't seem like there's a huge competitive moat between those. So I think that that's a big risk. But then you also have huge disruption for companies such as Google potentially where their search business could potentially be cut out a little bit by the AI models when they're doing stuff like that.

There's not as many potentially search queries. But then on the flip side of that, well, guess who's producing their own chips? A huge, huge volume on is going to be selling them and has built what's going to be one of the biggest chip companies within Google.

And there's so many. And chip design, they have so many different swings at the back. And that's one of the smart things that these large hyperscalers have done is that they've given themselves a lot of opportunity to where they're just not a one or two trick pony anymore.

You know, it's impressive what they've what they've done. And so if I think that's where valuation matters, Toby. Yeah, I saw one.

Yeah. Yeah. I saw one person who kind of compared, you know, somewhat the trepidation on the Anthropic IPO, how, you know, the price is kind of a little less than what they said.

Now, of course, they're racketing it back up. But with the way it was with with Google when they when they IPO in 2004, because, you know, we when when Google IPO, they had taken a controlling market share for who. But everyone had just lived through the tech bust.

And so, you know, what once again, what's the barrier to entry there? You know, is Google going to be able to sustain itself? And I mean, it was still a high multiple, you know, 10 times sales or whatever it was, but it wasn't it wasn't like I mean, there was a lot of hesitancy.

I remember I was working at Vanguard at the time and everyone was like, oh, you wouldn't want to touch that, blah, blah, blah. But it ended up being a phenomenal buying opportunity. So we'll see.

I mean, that's a very different valuation from where Anthropic's going. But we'll see.

📈 Want to see how value investors Tobias Carlisle and Tim Travis use options trades?

See the actual cash-secured put and covered call trades Tobias Carlisle and Tim Travis send subscribers — with ongoing recommendations as positions develop through the Wheel strategy.

👉 Start your 7-day free trial.

Link in my channel profile or use the link below.

👉 https://valueoptionsletter.com/a/acquirers-multiple

Disclosure: This is an affiliate link. If you subscribe, I may earn a commission at no additional cost to you.

Educational content only. Not financial advice.