4.18.25
industrial robotics and data, Transdigm as incredible business, true focus, soft/hard ROI, sperm racing, and Corleo the Robotic Horse
It’s so important to surround yourself with things and people that give you energy. I went on a walk with someone (👀) yesterday who I am just the biggest fan of and it had been way too long. Was so fun and a reminder to keep doing more of that. He said my weekly musings reminded him of a kinder-egg where you never knew what you were going to get inside and I just loved that description albeit a very kind way of saying WTF are you thinking about and why is it so random? A chaotic mind…. here’s the weekly musings. Two slightly longer thoughts this week…
Industrial robotics + data for reinforcement learning for robotics:
I've written in the past couple weeks a lot about data and robotics. There's something really fascinating to me about the flow of data to thing of value and it's clear to me that robotics are happening. Lot of focus on the humanoid as the form factor, but actually the robotic revolution is well underway in industrial settings.
I spoke to an SVP of a very large company who walked me through all of the places they are currently deploying robotics and it really hit me. We're here. It's still early, but we're here. Are there still tons of problems? Yes. Are the models incredibly complex and still have a long way to go for even the basic movements? Yes. Is the hand dexterity not where it needs to be? Also yes. BUT there are alternatives - like suction, like grippers, like programming one specific task, that are filling the void until the tech inevitably evolves and advances. It's also safe to say prices are as high as they ever will be and the cost curve will go down (tariffs are an x factor here of course).
As a part of this robotics wave, folks are desperate for pre-training data. Eventually we'll also need finely tuned data, but thats a step later in the funnel. The question then becomes what fidelity of data is "good enough" and how do you access that? Lots of folks going after this in different ways- whether its using videos, whether its doing simulations, or using existing systems. I feel like this point has been a recurring theme for the past couple weeks for me, but I can’t kick it.
If you are building in industrial robotics, data for robotics, or just have a hot take here, hit me up. I'm fascinated, actively learning, and would love to jam🍇.
TransDigm as an incredible business: It's an absolute monster, $75B market cap and up 300% over the last 5 years. They sell aircraft parts that can be found across commercial and military and 90% of their products are considered proprietary. The parts they focus on are mission critical but very high margin. The parts are also rarely switched out so can stay in for decades. Even more interesting, they have robust aftermarket business so essentially can sell their products multiple times. Their products have a lifetime value of 50 years!!!
We talk all the time about durable businesses, but my goodness- how many software companies do you know with an LTV of 50 years? The answer is very few. They have a real stranglehold on the market. And they are able to do this with gross margins of 61%. The aftermarket GM is even more insane at 80%. R&D spend is a measly 3% of revenue, shockingly low for aerospace industry. What that means, is they just buy proven certified components (lots of M&A and then let them operate independently) vs. inventing the future and basically operate as a PE fund in disguise.
I think about the characteristics we all love as software investors (sticky revenue, high margin, defensibility in GTM and unique flywheel, focus, etc.) and ironically, these guys have it in spades. TransDigm figured out how to turn commodity aerospace parts into an infinite-margin software investors dream. Maybe this is a bull case for deep tech albeit the lack of innovation feels very different to the core thesis of the deep tech VC space, so maybe actually not. Better reminder that there are lots of ways to make monies and tech companies aren’t the only way. 😊
I am slowly wading into hard tech solutions…startups that have a hand in the physical world. More to come in future musings. HMU if you want to jam 🍇.
True focus: Was listening to Neil Shah (so hot right now!) on Invest Like the Best and loved this point about few people actually having maniacal focus. In the spirit of transparency, it's something I simply do not have and I guess in some ways, the job of VC is antithetical to that just by nature of the task at hand. BUT, I do think it's one of the most important attributes for a founder. The ability to look at a business and deeply understand what are the very most important things that drive the business forward and just focus on those. There are so many distractions day-to-day, but ultimately its about driving urgency to the most important things because if you don't, there is someone waiting around every corner who wants to be on the path you are on. Very dog eat dog world of me, but it’s true.
Sperm Racing: WTF. actually.
Corleo the Robotic Horse: I’m a weird horse girl and this prototype video from Kawasaki is sick.
Lastly, a song i’ve been playing on repeat: Baby by Donnie & Joe Emerson
Stay weird. Stay curious.
-CBR (I’m technically an R now, but hard to kick the B!)