Brady Anderson was a center fielder for the Baltimore Orioles from 1989 to 2003. Anderson was never a prolific home run hitter, averaging just 12 home runs per season over his first 10 years in the majors. However, in 1996, he hit 50 home runs, which was the most by an American League player since Roger Maris hit 61 in 1961. To be fair, Anderson has never been directly linked to steroid use, but . . .

During the steroid era (late 1990’s - early 2000’s) home runs went through through roof. While overall the number of home runs went up, the most dramatic statistic was the number of players that hit more than 40 home runs

We are now coming out of the steroid ERA for investing. ZIRP was our steroids, and unicorns were our home runs. After all, we are in an industry that is predicated on home runs. As most people know, venture funds don’t work unless they have at least one home run, i.e. at least one company in a fund that returns most if not all or a multiple of the fund. Over the last 12 years, especially in 2020 and 2021 our home runs went way up and they have been coming down for the last couple of years.

What we were doing wrong for the last 3-5 years is now obvious. Indexing. You find something that has some momentum in any category that "could be big". As long as you are in the top 2 or 3 company, you make decent money. In baseball terms, that's making solid contact with the ball. Swing less discriminately and put the ball in play. In pre and post steroid era making solid contact usually led to a routine fly ball. In the steroid era it resulted in a home run.Just as home runs went way down when steroid testing went up, so have unicorns as interest rates have increased. We are beginning to see many of ZIRP-fueled unicorns turn into routine fly balls; companies like Hopin, Convoy, Bird Scooters, Olive AI etc.

As investors, we've been making solid contact with companies, but ZIRP led to these looking like home runs. I figured this out early on at Salesforce Ventures. We were deploying $1B a year, and our IRR was insane as were our realized cash returns. I avoided investing in companies once they passed $1B in valuation, because then I felt the gravity of absolute valuations was too great. At the time there were not that many companies valued > $5B. Where I broke this rule was with the investment in Snowflake, where we invested $400M. That is the fattest pitch I have ever seen. We thought we were going to invest at $8B, but we ended up at $12B or 13B. I would have invested at $20B. The Snowflake sales team were order takers not sellers. Salesforce was like that in the early days. The product sold itself. If you ever see that in a company, put as much money in as you can, because you only see it 1 or 2 times in your career. But I digress.

Back to the baseball analogies. The benefit of hitting a baseball is it only takes a few seconds to leave the park. You hit 50 home runs for a season, sign your contract and make millions. Unfortunately, VC investments take a decade to leave the park and then some. With increased interest rates, it's as though the winds picked up to 50 mph while the ball was at apex about to leave the park turning many of our companies into routine pop flies, maybe some doubles.

So what do you do if you can no longer take steroids, but you are paid to hit home runs. You can do 1 of 3 things:

So what does this look like for investing?

Keep doing what we have been doing seems like a flawed strategy even in the face of AI. For that to work tech markets overall or at least an entire category of tech has to see inflated valuations, which isn't likely. This happened in the 90's in the run up of the internet and then again in the early 2020's due to inflation. As much as we hope, it's hard yet to see that AI has that same impact. There are many analyses showing SaaS valuations vs. interest rates over time, and how the two are inversely correlated (this one by Stephens.com)

Swinging for home runs every time with lots of strike outs looks like making a lot of bets, see what's working and quadruple down on your winners. This seems to be the current strategy of larger funds who have become more aggressive at the seed stage and A stage in AI. It's doing what we were doing, just with smaller checks. Find two people from Uber and Stripe who went to Stanford. Cut them a check. Do that a bunch of times and be ready to pile in on the one 1 or 2 winners. Capital is also an advantage here. Crowd out seed investors with large checks and you can also price out a lot of A investors with large checks. I think this might work for large funds where they can make up for their losers and deploy $200M+ in a few winners, return 2x on $7B and call it a day. It does present an agency issue where it probably doesn't work well for many (younger) individual investors at large, but I think it works well for the firm. There's a lot of capital out there, and increasingly VC is becoming investment banking with an emphasis on AUM. I talked about this in my last post - Sea Change in VC.

Then there's waiting for the fat pitch. Well, that's harder. Historically lots of fat pitches come either when you're early at the beginning of a new tech platform such as, semiconductors, personal computers, internet, cloud and mobile. Unfortunately, it’s not yet clear that AI will be that. Again, I'm SUPER bullish on AI, but for the foreseeable future it will be an accelerant to lots of business and will  disproportionately help incumbents, existing growth companies and PE firms. I am writing about this in a post titled "AI is Electricity Not Steam". I coined that phrase (in my Notion notes), and then I heard Satya Nadella say it. So basically Satya lifted it from me.

Where are the fat pitches then? Well, even Munger says it's a lot harder now to be an investor. There's a lot more capital and smart people chasing investments. The internet has also reduced the level of information asymmetry from a select few in the know to everyone. I like to think of it how Ted Williams mapped out the strike zone into 77 zones, each the size of the baseball.

He only swung at the ones he could hit. As he put it, "My first rule of hitting was to get a good ball to hit. I learned down to percentage points where those good balls were. The box shows my particular preferences, from what I considered my “happy zone” - where I could hit .400 or better - to the low outside corner - where the most I could hope to bat was .230. Only when the situation demands it should a hitter go for the low-percentage pitch." Ted Williams had quicker feedback than we do, so it's harder. Also, as investors our strike zones change. It's too hard to do the rinse and repeat SaaS 1.0 playbook of replacing on-prem and analog solutions with cloud solutions.

The benefit of shifting from on-prem to cloud was that there were lots of fat pitches. The wind was at your back, and there were lots of new categories to be created. The value proposition was so much better vs. on-prem, which was very expensive, created vendor lock-in, updates required full system updates, you basically had your own data center running software . . . . . .When SaaS 1.0 came in the value proposition was so good. It was even better when you were replacing pen and paper. Identifying the best pitches was much easier, and there were far fewer players swinging at them. In the mid 2000's SaaS was not at all sexy. Everyone was moving to consumer on the heels of Facebook. SaaS was not viewed as really attractive category for several more years. So you were brave for swinging at SaaS at all in that category. You had to be non-consensus.

There are trends or drivers of disruption right now that will create big opportunities; investments in the modern data stack, AI and the overspend on a bunch of SaaS tools. The combination of the modern data stack makes data more mobile throughout the enterprise. The advancement of AI in combination with lots of data that is better organized and more mobile throughout the enterprise, will provide the fuel needed for AI to do it's magic. I believe this will lead to a consolidation of applications in horizontal and vertical apps both in the enterprise and for SMBs.

But I don't think just consolidating existing workflows and bundling apps is sufficient to dislodge incumbents like we saw in SaaS 1.0.  Even if your stuff is a little better, offers some automations and is cheaper the burden to switch is too high. The incumbents will also offer AI and automation, maybe it won't be as pretty and the automation isn't as good. But the pain of moving off an incumbent is high, and you have to offer something way better and novel. Salesforce has been proving this forever.  The UI has never been great. But try as they might, no start-up CRM has come close to challenging them  in the last 20 years with maybe the exception of Hubspot. They've all died or stalled. AI alone won't change that. Salesforce is also doing AI and doing it well.

That means you need to develop your own Ted Williams strike zone for fat pitches. Maybe you look at biotech, materials, defense tech etc. I am heavily focused modern apps that are heavily reliant on data and AI, but that doesn't offer the same step function in improvement as moving from on-prem to cloud. SaaS, mobile and Cloud offered an entirely new delivery method of delivering software (in a new form factor and interface with mobile), whereas AI is providing a more efficient unit of work. I borrowed this from Peter Thiel, which I thought was a brilliant way of thinking about it.

I focus predominantly on enterprise SaaS, data, AI and fintech in the context of SaaS applications. With that context here are some of the themes and characteristics I am pursuing.