Some thoughts on what the hell is going on in venture

Seven of them, in fact.

IYKYK.

1. It took a decade for the median VC (and not just the best VCs) to understand how to quantitatively analyze the characteristics of a SaaS business. Now that this knowledge is table stakes, it's on the verge of becoming obsolete. What even is "ARR" in a consumption-based AI company? Do our benchmarks make sense anymore when gross margins dip below 50% because of compute costs that now live in COGS? How confident should we be about lifetime value if "AI wrapper" companies end up not having any moats?

2. One of the advantages of investing into SaaS was that beyond a certain point it was reasonably predictable. A firm that only invested into Series A companies doing $1-2m of revenues and growing 300-400% annually could expect a decent enough proportion to make it to a scale where early investors should make solid returns even after dilution (say, $50m ARR). Compare to now, where it seems questionable whether some of today's early stage winners are even going to be in the same line of business 24 months from now. Underwriting growth durability in this environment seems really, really hard. (Conversely, founder dynamism seems a whole lot more important.)

3. A portfolio company of ours recently went out to raise its Series D. Over and over, they were told by various growth equity firms that the only deals under consideration for investment right now are AI companies and "Rule of 40" type companies that have both growth and profits. Non-AI and non-profitable? You’re out of luck.

4. Back in the day, seed investment for startups went into various capital costs that were necessary to get the initial product out the door… stuff like servers, hardware, and co-located office space in Menlo Park. That went away when cloud computing and remote work created a new reality that "it's never been cheaper to start a company and build an MVP." Predictably, these lowered the barriers to entry and permitted the creation of lots and lots more tech startups. Now, capital intensity is back, but in the form of compute. Does this mean that barriers to entry are back up, competition within markets is going to go down, and VCs will become kingmakers again? We'll see.

5. A significant percentage of the GPs who raised and deployed their first funds in 2020-2022 are now completely unreachable. They don't respond to founders or other investors, and in some cases they appear to have left the industry without informing their LPs. The "zombie firm" situation is more extensive than I think a lot of people are aware.

6. VC attention is more narrowly focused than I have ever seen it. In prior manias, there was a much more expansive set of things that investors had to keep track of to understand what was going on (for example, keeping up with crypto meant swimming through dozens of different protocols, exchanges, forks, business models, regulatory schemes, et cetera). Whereas in the AI mega-trend we see 10,000 different VCs writing the same Substack post about what the same five companies (Nvidia, OpenAI…) are up to. It's weirdly monolithic, which is odd because the actual use cases for the technology are so broad and diverse.

7. Now more than ever, when evaluating new opportunities I try to ask myself, "why am I seeing this deal?" This seems like an investing climate where it's extremely easy for the overconfident investor to lose their shirt.

Shout out to Matthew Giampetroni and Connor Sundberg for recent conversations that inspired some of these thoughts.