Changes To Investing In AI: Why I'm Moving Away From Early Stage Investing and Analysis
I no longer believe early stage has the same potential it once did.
Happy Sunday and welcome to Investing in AI! I’m Rob May, CEO at Brandguard. I also run the AI Innovator’s Community in Boston and New York. If you have a chance, please check it out. Also, until recently, I was a very active AI angel investor. Today I want to talk about why I’m moving away from that, and what to expect in terms of changes in this newsletter.
I sold my first company in December of 2014 and committed $1M to angel investing over the next few years (and eventually more). I made my first angel investment in June of 2015. Since then, if you factor in direct angel investments, my time as a VC, AngelList deals, etc… across all platforms I’ve made 133 investments in 9 years. Ninety percent of them were AI related, and almost all of them were sub $1M in revenue. Most had not revenue. And I’ve been the very first check of any kind into about 20 deals.
The returns have been pretty good. I’ve had two unicorns so far and a few more that are close. I didn’t realize how long it takes to start returning capital from early stage investments so, I was a bit too aggressive in the beginning but, now I have a few exits every year and it’s pretty balanced.
What I loved, and still love, about early stage, is the fight that entrepreneurs have to go through. It is really exciting to watch, to see the future being built, and to mentor and encourage first time founders, or commiserate with repeat founders.
But in the 18 years I’ve been involved in startups, early stage venture has changed, and I think there are many negative forces pushing on early stage startups that make it very frustrating as a place to play. On top of that, I’ve grown to lack respect for a lot of VCs. I don’t want to paint any industry with a broad brush because there are a few thousand people working in VC worldwide and many of them are great. I’ve had the chance to work with some great ones. But by and large they don’t really believe what they say they believe. Here are the things I don’t like about early stage VC.
They invest on momentum, not conviction. This leads to more fads getting funded than solid companies. Real businesses with competitive advantages take time to build. That’s not what I see mostly get funded. In fact, if you can’t pop quickly with some gimmicky trick, it’s harder to get funded.
Salaries and hiring practices haven’t adapted to the market. When I started in startups, an engineer might have to take a 20% pay cut to work for a startup, for the chance to get rich from equity. Since then, big tech company compensation has far outpaced startup compensation growth, and VCs haven’t adapted. Instead of asking someone making $120K to go to $100K, you are now asking someone in their 30s making $400K - $600K to go to $180K. And the truth be told, their big company equity is probably going to be worth more. VCs will lecture you about being a cog in a machine or whatever at a big company but, the truth is it’s harder and harder to hire great people because you have to ask them to take a 60%+ pay cut. I’ve seen too many of my investments get stuck with mediocre talent because it’s all they can afford.
Tech companies are better at defending against startups and have bigger new project budgets. Many public software companies were once venture backed startups. That wasn’t the case 20 years ago. These companies are familiar with the innovator’s dilemma and are constantly snapping up good teams, small pre-revenue startups, and are funding all kinds of projects to make sure they don’t get hit by a surprise innovation. They are better than public companies were in the past. On top of that, a new project at a big tech company might get a multi-million dollar annual budget for a few years to prove something out. But as a startup, you have to show more progress and momentum on your $1.5M pre-seed in a year or less. It’s unreasonable.
Goodhart’s Law Plus Too Much Seed Money Has Ruined Things. There has been an explosion in pre-seed and seed funds in part because you can raise these from HNWs and Family Offices and don’t need formal institutional money, which makes it easier. But there isn’t enough follow on capital, and the small funds can’t take you that far. Goodhart’s Law says that “when a measure becomes a target, it ceases to be a good measure.” Early stage startups never ask me about business building. And they sure as hell don’t have the time or money to invest in building a durable competitive advantage. They ask “what do I need to get a Series A” and they build towards that. If you wonder why so many companies get to Series B, then see intense competition, slower growth, or even collapse, it’s because they were trying to satisfy Series A investors (a la Goodhart’s Law) rather than build a real business.
Many of the most interesting tech problems are at big companies now. Years ago, most interesting tech was at startups. But for AI I’m not sure that’s true. The AI startups doing the most interesting work have all raised massive amounts of money - they aren’t small at all. Outside of those, much of the most interesting AI work is at Google, Facebook, Amazon, and that tech co-hort. Maybe as a startup you have a shot at getting someone to take a pay cut to come work on something incredibly intellectually stimulating. But now, big companies have most of those problems. It’s actually more fun there in many ways.
Too many companies make progress but have tired syndicates. This ties back to the rise of so many small funds. I’ve been invested in a lot of companies that pivot a time or two, finally figure things out, but end up with a tired syndicate. New investors don’t want to do a recap, so the founders have all this knowledge, have finally figured out some key insights from 2-3 years of working in a market, and it just dies.
Early stage investors are slow to adapt to new business models and realities. When I was raising for my first SaaS company in 2009, a significant proportion of investors were anti-SaaS and still wanted companies to measure bookings and maintenance contracts and such. I was told SaaS was a fad that wouldn’t work because it took too much capital. Now as AI companies are changing those metrics (I see most a lot monetizing more like infrastructure companies - usage based) VCs are still trying to wedge them into a SaaS metrics framework for evaluation.
Again, I’m talking in general trends I’ve seen and am painting with a broad brush. There are a lot of great investors who don’t have these issues I mentioned above, and there will always be market opportunities at early stage. But I believe the market dynamics are not in favor of venture over all, and will get worse in the next few years.
As a result, I’ve paused most of my angel investing. I’ll still do deals that I really love, but only because I personally love something about the tech or the team. I’m not going to keep looking at 1000+ deals a year, because I think the big opportunities, particularly in AI, are elsewhere.
Plus on a personal level, I’m tired of working on early stage problems. It’s not fun trying to convince people of an idea they aren’t familiar with, and it’s stressful to constantly have the existential risk of dying. It’s much more fun to think about the kind of real business strategy you get into when you have competitors, lots of revenue, multiple products, multiple channels, and sell around the world. I have my own startup to get to that level, and that’s enough for now. (No I won’t do another startup from scratch either - I’ll buy something next time and start there.)
This newsletter is pretty personal, as all my blogs and newsletters have been over the years. I don’t write to monetize, or even to build an audience. I write for me actually, not you. I don’t use ChatGPT or anything else because my goal isn’t just putting out content. My goal is to use writing as a process to think through and refine my ideas. That doesn’t work if I take shortcuts.
Over the coming months, expect more of a shift to PE and public markets ideas about AI. I’ll still write a lot of the philosophical stuff, but it will probably be less early stage focused. And there is a good chance I’ll start building in more decision making and leadership talk. Those are two topics I’ve been deeply interested in for years, have written a lot about in other formats, but that have come back to the forefront of my mind with some things that have happened recently. The business world and political world both lack the kind of leaders who believe in ideas bigger than themselves, and I believe this is a bad thing for the world.
The newer stuff won’t be right for some of you, and that’s ok. Feel free to unsubscribe and find something you enjoy more. In the meantime, thanks for reading.
great post - thanks
Love it. Thanks for sharing these great reflections, Rob.