AI & Avocado Toast

Our early thoughts on investing in the AI technology shift as a pre-seed fund.

In Q2 of 2024, funding to companies in AI more than doubled quarter over quarter to $24 billion — representing 30% of all dollars invested per Crunchbase [Crunchbase Report].  One of the questions we get all the time is how we’re thinking about and investing in AI as pre-seed investors.  Our thinking is evolving as the market grows and adopts AI, but we are spending a lot of time reading, experimenting and talking to founders and experts trying to understand where our small checks to very early-stage founders can have the biggest impact in the long-term. 

Most recently, we convened a group of technology leaders and investors to chat about generative AI over breakfast - an “AI and Avocado Toast” morning. The attendees ranged from founders actively building AI companies to later-stage investors and family offices curious about AI trends. FirstMile and the founders of the Regis Company (FirstMile Fund II), Mike Vaughn and Joel Janov (two of the sharpest folks we know in terms of leveraging a variety of generative AI models to help solve problems for their customers), kicked off the breakfast conversation. 

Some of the key takeaways from the breakfast discussion were:

• Generative AI is not just hype.   A lot of people have seen fancy demos or advertisements only to find they cannot figure out how to effectively incorporate it into their work or lives.  We and most of the folks in the room agreed that Gen AI is going to change our lives dramatically; however, the pace of that change is yet to be determined.
• We believe that generative AI is a transformative technology that will power the next technology shift.
  • Much like the PC in 80s, Internet in the 90s and mobile and cloud in the 00’s, these technology shifts enabled a wave of other innovative companies to flourish.  
  • We’re fortunate to have Bill Miller on our team, who co-founded Storage Networks in the mid-90s.  As internet connectivity and bandwidth grew, he saw the opportunity to attach storage to the network.  He became an early pioneer of cloud storage and cloud infrastructure, helping to usher in a wave of cloud-based application companies.
•  It takes a deep understanding of the capabilities, strengths and weaknesses of each AI model to ultimately harness them in a useful fashion but they are very powerful tools in the right hands.

As a firm, we map today's AI investment landscape into three broad buckets: Platform, Tooling, and Applications.

Platform
In the world of generative AI, Platform companies are building the foundational models.  Much like building a cloud infrastructure company, this is a very capital-intensive endeavor.  Companies like OpenAI, Google, and Anthropic are spending billions of dollars to develop and train these models.  As a firm focused on capital-efficient companies, this is not a category where we see much opportunity.  But we do expect the competition at this level to help us immensely, see later bullet.

Tooling or “Picks & Shovels”
To get something useful out of a foundational model, you often need connections to data sources or interfaces to engage with the models in a structured way.  There are also “picks and shovels” types of investment opportunities here, the most obvious of which is the chips to run the models, which is why NVIDIA has surged to become the most valuable company in the world.  We believe there could be some interesting gaps in this area that a fast-moving startup is best suited to fill…but we’re also treading carefully.  This space is moving very quickly and we’ve already seen platform companies add capabilities that turn a really interesting startup idea into a feature that’s baked into the model.

Applications
This is the area we are spending the most time on.  We’re looking less for “AI companies” and more for companies that use AI to solve customer problems and operate more efficiently themselves.  Just as almost all of our portfolio companies leverage cloud infrastructure, we have been seeing them now use AI as a component that supercharges their solution. The investment and competition at the platform or LLM level is creating a fantastic set of capabilities that startups can access at a very reasonable cost. The team at Regis, for example, is combining a deep knowledge of corporate learning and development with new AI capabilities to deliver a better solution that is faster and cheaper for their customers.

These rapid advances in AI capabilities make it an exciting time to be an investor. If you’re leveraging AI to solve a painfully large customer problem, we’d love to chat with you.