AI-Driven Investment Firm
Why hasn’t anyone created an AI-driven investment firm yet? You would think some young people would come together and pitch this to LPs
Trending Content - Hedge Fund Forum
Career Resources
What do you think quant funds are doing?
Where have you been for the last 20+ years? It's called quant, they're just not being as buzzwordy. Specifically marketing themselves as "AI-driven" - there's like a dozen or more funds that come to mind from just the last 4-5+ years.
Not just publics but privates as well. Think an AI-driven version of Altimeter or something. Long duration AI-driven idea generation on public side and analysis of historical VC data to come to a sound decision on private side.
There have also been quant-driven VCs for a while too. One of the founders of Rentech did this with Differential Ventures. Applied AI in finance is just the buzzwordy rebrand of quant finance. You could argue firms like Bridgewater that build a giant ML model to ingest macro data and spit out investment/hedging ideas are also "AI-driven".
Bro thinks chatgpt is gonna outperform 20+ years development of MIT math PhD level algorithms.
LPs aren't morons like VCs are so they aren't going to allocate substantial capital to a couple of Stanford CS dropouts full of buzzwords when the above exists instead.
Not just publics but privates as well. Think an AI-driven version of Altimeter or something. Long duration AI-driven idea generation on public side and analysis of historical VC data to come to a sound decision on private side.
Historical data has nothing to do with VC. That's not how any of it works. VC is about access and connections. There is no "trading history" for private investments so that kind of information is irrelevant for an AI to try and extrapolate decisions from. In order for something like this to work it would literally have to be predicting the future.
Luhhshhehgwg
Quants like DE Shaw do not utilize AI as a core driver of their investment decisions. They use satsitcal models that take into account historical price data and such to effectively predict the directionality of a stock but for the most part, the investment decisions are being driven by mostly humans. I was saying why haven’t anyone tried to create a full stack investment organization majority run by AI agents with maybe 2-3 investment professionals.
AI like LLMs is just ML at scale with a neural net/memory function for recall, so in a way they are indeed using AI.
Yes but they are using historical data to come to a viewpoint regarding the direction of a security. They are not using and training the underlying LLM to provide actionable alpha-generative insights.
You... don't know what you're talking about. That's literally what shops like Citadel, RenTech, etc. have teams entirely dedicated to. Maybe it's not in an LLM format where they're using a chat interface but the output and actions that happen downstream for it are identical to what you're describing.
What do you think an LLM would do in your scenario? Pull alpha out of thin-air and predict the future like an investing Jesus? Of course not. It's basing its initial analysis on historical data and extrapolating external macro or sector-specific factors (really anything you want to feed it) to then generate a recommendation. This isn't a novel idea it's something HFs started optimizing for years ago. You do realize that the Chinese open-source foundation model DeepSeek is literally the side project of a quant fund that wanted to use their excess GPU capacity right?
What's your background or experience with LLMs or quant investing?
Haven't been on public side for a long time so I'll leave to others with more experience. However, I was in a statarb/quant pod at an MM 10+ years ago and was researching how to derive signals from social media. The signal was too weak to meaningfully drive alpha but we had the tech to evaluate sentiment pretty easily.
"LLMs allow for a holistic, real-time synthesis of market-moving narratives that traditional quant systems weren’t designed to handle before."
It sounds like you're mistaking the front-end capabilities of LLMs, outputs in natural language responses i.e. "narratives", as a source of incremental alpha generation. Sure it packages it up for a non-technical person to understand but it's not much different than what's being done today at modern quant funds.
On the private side, half the edge is access. Sure having a super sophisticated LLM could point you towards Snowflake in 2020, but it doesn't mean you'll actually get a chance to invest. If you go earlier stage you'll have even less data to make assumptions or theses on. Sure you can make thematic bets based on trends, but at that stage you're backing the people more than a business.
I used to work at a sophisticated, well resourced VC that had in-house engineering for data projects. The CIO is a good friend of mine and he said that they've yet to meaningfully attribute any single investment success to their AI efforts. It helps during diligence and pitching but the deals themselves are sourced and won via relationships.
Buddy thinks ChatGPT is the unlock for self driving cars
Regents Gate Capital
ah yes, a bunch of 20yo that just finished jerking off across uni dorms pitching a GPT-based investment firm to Qatar Investment Authority side by side with 40yo STEM degrees from Harvard who use Nasa-level programs that were already doing 20 years ago x250 more things than what AI does today for stock markets
how no one has thought about this??? extremely unique POV OP, great observation probably you should pitch those middle eastern LPs on your gemini based investment firm (can't afford GPT)
go get em!!
I work at one of these big blue chip funds. There’s no way anyone could compete. We know AI is a huge driver for growth in the future and we’re so heavily invested in it across all domains. Whatever some small group of people can do, trust me is being done at scale already. AI is a great tool but not replacing a full fledged analyst *yet
Also no LPs wanna give money to something unproven with no track record. So you have to run this whatever strategy for a couple years successfully first. It’s a chicken and egg problem.
Seriously, the big HFs are like 10 years ahead of everyone else in this industry on all of these types of things.
.
Ad aperiam inventore odit dignissimos in dolor. Et ad enim voluptas voluptatum doloremque reprehenderit. Qui nostrum temporibus placeat ut.
Tempora quia voluptatem deserunt voluptatem explicabo voluptas. Doloremque eum ut dolores corporis eum. Magnam animi ut tenetur voluptates ducimus. Nesciunt dolor deserunt quae eum.
See All Comments - 100% Free
WSO depends on everyone being able to pitch in when they know something. Unlock with your email and get bonus: 6 financial modeling lessons free ($199 value)
or Unlock with your social account...