# 137 backs AI companies with positions above the models

Markdown copy of https://realalpha.app/s/137-backs-ai-companies-with-positions-above-the-models/ on Real Alpha, the best minutes from long interviews with the people building AI. Cite and link the page, not this copy.

- Speaker: [Christian Garrett](https://realalpha.app/people/christian-garrett/) · Christian Garrett is a Partner at 137 Ventures and co-founder of Onodrim Industries. · Backed SpaceX for 16 years at 137 Ventures and co-founded Onodrim, a European defense and industrial platform.
- Host: [Molly O'Shea](https://realalpha.app/people/molly-oshea/) · Founder, Sourcery · Ex-Upfront Ventures investor; has interviewed Alex Karp and Palmer Luckey.
- Show: Sourcery
- Episode: [SpaceX, Palantir, Anduril & Onodrim | The Return of Conglomerates](https://realalpha.app/v/4J7tjyVCrqA/), Nov 13, 2026
- Moment: 16:07 to 19:27 of the episode (3 min), playing on https://realalpha.app/s/137-backs-ai-companies-with-positions-above-the-models/; the episode from there: https://realalpha.app/v/4J7tjyVCrqA/?t=967

## Quote

> “We have not done a ton there but those are companies that we felt were pretty well positioned within sort of having the **third-party structural positioning**”
> — [Christian Garrett](https://realalpha.app/people/christian-garrett/) at [16:52](https://realalpha.app/v/4J7tjyVCrqA/?t=1012) · Sourcery

## Takeaways

1. Cognition and Hex sit at the harness and application layers.
2. Third-party positioning can reduce reliance on any one lab.
3. He weighs rapid growth alongside business durability.

## About this segment

Christian Garrett explains why Cognition and Hex appeal as companies at the harness and application layers, with third-party positioning and fast growth. He sees those traits as helpful when underwriting durability in a shifting model market.

From “SpaceX, Palantir, Anduril & Onodrim | The Return of Conglomerates”, Nov 13, 2026. [Watch the full episode](https://realalpha.app/v/4J7tjyVCrqA/)

## Transcript

Transcribed by machine (Whisper) from the episode’s audio, so a word may be off here and there.

[16:14](https://realalpha.app/v/4J7tjyVCrqA/?t=974) we get into the portfolio, because you mentioned a couple of these areas of things that you're working on or looking out for in market dynamics and the structure there. So of those, I don't know, five or six pieces that you laid out like what are the most vulnerable parts and then at the end of the day is pricing just based on momentum and stories or like how much of that is really tied to the numbers it's hard to say i mean i think you know there are areas that we've focused on with higher conviction and we've tried to you know look for companies that we felt have more durability and defensibility which is always our focus on underwriting um so you know we've obviously invested quite a bit in cognition we've just co-led hex's round which is a company we've invested in for a long time. And you know both of those sit sort of at the harness layer and the

[16:58](https://realalpha.app/v/4J7tjyVCrqA/?t=1018) app layer. We have not done a ton there but those are companies that we felt were pretty well positioned within sort of having the third-party structural positioning which is very beneficial. You know not every company wants to be reliant on a lab for certain types of products. Definitely on security, definitely on coding we've seen right and being able to rub the different models and we think also in analytics which is where you know very in a lot of data workflows which is where Barry and Hex sit. I think those companies, once again, we believe have good structural positioning and it's easier for us to underwrite some of the durability of the business.

[17:28](https://realalpha.app/v/4J7tjyVCrqA/?t=1048) And then second is we're just back in growth. All those companies are growing extremely fast, both of them. I think you've obviously seen, whether it's with XAI and SpaceX, the ability to see the value accrue to the hyperscaler and the compute layer, regardless of outcomes. We believe that open source is clearly going to be a large percentage of models and deployment, just like it is within software you know where do you pick your bets how do you sizes bets not been easy and there's been certain categories we have not done right or invested as much and that's less of a view of how well the companies are doing it's just been harder for us to sort of think about and grasp the durability of the growth and what market dynamics will look like in the future so in essence i mean a couple areas we've made bets in a couple areas that we're still

[18:10](https://realalpha.app/v/4J7tjyVCrqA/?t=1090) spending time getting more conviction in i don't think i have any more better informed view than someone else. And I think, you know, this is something that, you know, we're all spending time underwriting. I will say, though, that clearly you've had an incredible set of companies that have been built at layers that we haven't invested in. But, you know, if you look, I mean, OpenAid Anthropic are two of the most impressive businesses that have been built in a generation and incredible products that we're all using today. And, you know, just because there's a moment in time of volatility or questioning, you know, sort of what the out years of growth will look like, what the economics will look like, that's going to happen with the sort of fast moving and transforming market and a lot of disruption, you're going to have

[18:53](https://realalpha.app/v/4J7tjyVCrqA/?t=1133) all these things kind of shake out. But on the other side of the volatility, you can sort of peer into because just look at right now, I mean, I'm sure we all have family members that are using ChatGPT, right, and using forms of agents and Muse and Dot and GrokBots. And I mean, I'm still impressed by seeing, you know, cousins and friends and, you know, grandparents using these products. That's a clear sign that, you know, these obviously have a lot of runway on impact in the world and are still early in their diffusion.

[19:19](https://realalpha.app/v/4J7tjyVCrqA/?t=1159) And, you know, like all things, markets overextrapolate the good and the bad.
