How I Think About AI Investing in 2024

How the jagged frontier impacts investing

Happy Monday and welcome to Investing in AI! My last post caused a bit of a stir, and generated the most email feedback of anything I’ve written in the past couple of years. (Evenly split between agree and disagree.). But the most common question I got asked is where I’m investing now, and where I think the opportunities are. Today I’ll attempt to lay out my thinking and answer that question.

First of all, I didn’t intent to imply in my last post that there are no early stage opportunities. (I define early stage as pre-seed and seed in that post, but realize many investors may include a few later rounds in “early stage.”). The post was just highlighting that the combination of changes in the venture ecosystem plus the rise of AI and it’s unique properties might be changing the shape, size, and number of early stage opportunities in ways that make investing there more difficult.

The second things is that, while there will definitely be some early stage opportunities that remain, I think the bigger opportunities, in aggregate, are not in that market segment. I believe, as I stated in the previous post, that publicly traded tech companies have become better at defending against the innovator’s dilemma, so the place to go next is markets that haven’t had to defend against it so far.

If I had to list out the key drivers in my thinking on where to invest next, I’d lay them out like this:

• The “jagged edge” of AI makes experimental approaches better than analytical approaches.

• Contrary to the conventional wisdom that startups are the most experimental, big tech companies are very experimental and can run experiments longer and dedicate more resources to AI experimentation because the gains will be larger.

• The cost to develop software is going to drop dramatically over the next 5-7 years, changing build vs buy decisions and making it hard to build a “traditional” SaaS company. Actually, let me correct that statement. It will make it so easy to build a traditional SaaS company that competition will erode margins. You don’t want to invest there.

• The traditional online methods of customer acquisition are going to weaken due to the rise of agents.

• AI in the short to mid term still relies heavily on data, data, data, and many existing companies have (or can collect) unique data sets easier than startups can.

I think a lot about technology changes through the lens of economic complements. I use the mental model that two things are linked and one drops dramatically in price, increasing demand, so demand for the other goes up but, maybe AI doesn’t affect the production/supply of that thing as much. It’s why we built Brandguard. I was thinking about the explosion of marketing content generative AI creates, and then, what will be in more demand as a result? — The review and approval of that content.

Given this lens, here are things I currently think about when investing in AI:

• Don’t focus on stage, focus on insights about AI, and apply them to whatever stage is appropriate. This maps to the jagged edge concept of AI - good and bad applications of AI aren’t always obvious so when you find one, lean into it at whatever stage it best applies.

• Find business models that are more than software, because pure software will become less defensible over time. More and more I like models that bundle software with something physical: sensors, services, a needed human touch, etc.

• Consider companies that can benefit from strong ecosystems, particularly in GTM. The SaaS wave of the last 15 years was weak on channel partnerships as an early GTM focus. I think that will change. If you have a lot of channel partners or other integration and ecosystem partners who benefit from your success, there are many people who are incentivized to make sure you aren’t replaced.

• Look to apply AI in industries that have not had to defend much against the innovator’s dilemma. This could mean bigger more vertically integrated business models but, in many cases I think that will be the right play.

To add some clarity to where I stand on investing in AI - I think the opportunity is massive. But unlike other types of technology that came before, it won’t all start with startups and expand up market. There will be opportunities everywhere, but I think pre-seed and seed AI companies that look like the pre-seed/seed companies of the past decade, will not be nearly as successful.

As always, thanks for reading.