The Six Mental Models For AI Investing
How AI is different than other business models
Happy Sunday and welcome to Investing in AI! Over the holidays, I write a guide to AI investing in response to a lot of the questions I received from LPs in HalfCourt Ventures. We’ve been investing in AI since 2015, and done over 100 AI deals so we’ve seen a lot of what works and what doesn’t.
If you want to download the full guide, it’s here. If you want the TL;DR version, it’s below.
AI is a collection of technologies, it isn’t just one thing, and that can make it confusing to discuss AI.
AI is causing major market dislocations. OpenAI could be to some new company what Myspace was to Facebook.
AI is moving away from SaaS metrics. Many investors have begun writing about this but, we are seeing more per-unit-of-work pricing, or similar models, not as much per-seat-per-month. When revenue isn’t a recurring flat fee, it does fit mental models of investing if you are comparing it to SaaS.
AI changes corporate strategy. Data is now a more important asset than ever before. Labeling data may be a workflow you need. Value chains are being re-shaped by AI.
AI companies can scale faster, with fewer employees
AI may not have the zero marginal cost we are used to in tech. With the rise of test-time compute and other inference related costs, marginal costs could be more variable and use case driven.
It took a while for investors to get comfortable with SaaS metrics as the world moved that way from packaged software. I think it will take time for investors to get comfortable with AI business model metrics as well. Adapting your process to incorporate these new mental models will give you an advantage in identifying early potential winners in the AI space that other people overlook. If you find this interesting, go read the whole 12 page guide.
Thanks for reading.