How Hedge Funds Use AI to Turn Data Overload into Investment Edge

When computers came into being, like the alpha did not erode away. It just changed the nature of the industry. It was like an expectation that now everybody would have a computer. Nobody's gonna do the numbers by hand, right? So the expectations of the market, you know, changes with the tools that are available. Right now in 2026, people are like, you know, feeling the gravity of AI. I think last week Ken Griffin on a panel was talking about how he was blown away. that AI can now do things that a PhD in finance and maths would have taken weeks and months to actually do.

Hey, and welcome to the Momentum Podcast, the show where fintech meets AI. I'm your host, Stan Altshuler, investor, founder, and advisor to some of the world's top hedge funds, asset managers, and data companies. Each week we'll dive into practical AI strategies to help you cut through the noise, beat data overload, build smarter workflows, and turn your ideas into results. We'll explore everything from AI tools for your daily workflows to building data-driven growth engines for your organizations. Here, you'll discover what's next in finance and how to stay ahead of the curve with AI. Let's dive in. Well, hello and welcome to the Momentum Podcast. I'm your host, Stan Altshuler, and today we have a special guest, Ashu, the CEO of Matterfact. Welcome to the show, Ashu.

Hey, Stan, thank you for having me.

I'm, I'm very excited to, uh, have you, and I'm sure our guests are really curious how you're seeing the world and how it's changing working with all the hedge funds out there. So you've worked at Google before launching Matterfact. Can you tell us What led you to start the company? What made you want to solve the problems in finance that you're solving right now?

It's a good question. So you wouldn't know this, you wouldn't know this about me, but I used to work at Millennium as a quant back in the day before I even started my job at Google. I've worked at Lehman Brothers right after, you know, its demise. So I saw the effects of the 2008 crisis firsthand. And honestly, like, you know, like working over there, working in financial services was the best time of my life. Like, you know, I had so much fun just seeing that like the value that you can provide. Like finance is the one place where your work has the highest relationship to the ROI that you can create. So that's what made me really excited about this space.

And there's certainly not a lack of problems that you can tackle. There's some really big problems that, you know, AI is very well suited for. And that's what I really wanted to talk about. And I will say that a lot of folks that I speak to in fintech, they take the easier path and they kind of pitch towards time-saving, you know, efficiencies, some automations, and AI can really shine with that. It also gets People's eyes light up because maybe they'll save a couple minutes here and there. You are quite on the contrary. You don't go for the time saving. You go for like a higher level thing. You talk about insights and they they lead to alpha. Can you tell me why? Isn't that harder to sell? Why you decided to focus there?

Yeah.

So you know, like I love this movie Office Space. I always come back to it. That's that movie has like yeah. That movie has like so many nuggets. So in this movie, there is this guy who who keeps filing TRP reports and he's always behind on them. If my customer base was someone like that, I would have for sure built a company about productivity, but our customers are not like that. Every hour, every minute that they spend their work on, they are in the quest to finding insights. And our job to create value over here can only come from if we align ourselves to the goals of our customers, which is to find insights, to find new ways of finding alpha, to look into places that nobody else had looked at before. You know, I always go back to it and I like, you know, think about like, you know, if I'm an analyst, sure, during the earnings season I'd be super busy and I would have like, you know, I have to like, you know, update my model within like a few hours. It's a very stressful time. But beyond that time, right, I'm always like, you know, just looking to like, you know, find new insights. I will make more money, I will be a better analyst, I'll be a better portfolio manager, not because I can fill the model faster, but because I can find good insights, good variant views. And that's where we are aligning ourselves to.

And that's very interesting. And some of the, some of the hardest working folks analyzing as much information as possible and getting them into insight faster, I think is— it's not an easy undertaking, but it's definitely a noble cause. And I think it's one that's highly valuable. One of the ways you do that is with something you call artifacts. Now I've seen them personally. I think they're like the most impressive freaking thing ever. I think this is really truly where AI can shine. Can you talk to me about why you chose to have Matter of Fact be so damn good at building artifacts where you chose to be, to have that one of the delivery mechanisms of these insights for folks?

Yeah, so the genesis of the idea was actually very simple. Like if you look at, there is like so much information out there in the world, right? There are like so many news sites, you know, you have like so many government filings and like, you know, so on and so forth. So there is data that is like, you know, spread across the world and the job of like someone trying to find insights is to actually triangulate all of these things and build insights. So like we're talking to this person and apparently they were paying like a Fortune 10 company $100,000 a year just to build out a simple mini SaaS application tracking all the data centers in the world, like just that simple application. And like So that gave us the motivation to like, you know, think of like, you know, how can we give that power to everyone in the world that they can build it, you know, that they can build this thing on their own without having to like go through a lot of people and like, you know, having so on and so forth. So I can take a moment and like, you know, show you what this dashboard looks like. So—

Can you, can you please do that so you can share your screen? And while you do that, I will say that some of these things are so powerful that the limitation is a failure of imagination. That's what it was for me when I heard the first thing about artifacts. I'm like, okay, well, maybe I can mix a little lab, but surely it won't be able to scrape every single job posting of a hedge fund and then store the data, you know, in a relational database and retrieve it and create a dashboard. It does it. You have to open your mind. Like, really, I know it sounds like cliché and corny, but you have to like rethink the limitations, rethink what's possible. Like, give it a try. Like, it's freaking possible. It will do all So yeah, as, as we lead into that, why don't you just show us? I think you've built this with a prompt.

Yeah, um, I've built plenty of these.

I'll show you one that is like, you know, very top of mind right now.

There is this like, you know, AI hype going on these days, or like, you know, all these data centers and the CapEx investment into it. And we were like, you know, building this out for APE Farm, and essentially they wanted to like, you know, track all the data centers in the world that are being announced or are being built out and what's going on in them. Like, you know, like who is building them, what is the power supply in these, and like, you know, so and so forth. So, you know, as far as we know, there are 280 projects. You can track all of these across different countries. You can track things like which states these are coming up, when do these data centers come online, and like so and so forth. You can even track, you know, everything about the power that is being supplied, uh, you know, to these data centers, where are the— who are the operators, and like, you know, how this thing is is going to work out. Like, all of this, right? This is your own personal SaaS mini dashboard for you that remains, that remains up to date and is working for you all the time, right? Like, people say that AI is going to eat up SaaS. I don't think that's true. I think it's just going to change the way SaaS gets built and the amount of software gets built. I think, like, people would start expecting a lot more insights, you know, and the insights come from like having more and more code for you.

Yeah.

This is absolutely awesome. You know, just to let our viewers know, this is something like I'm very, very excited about in this product because I've seen it firsthand. Like you can build this with a prompt. And so this leads me to my next question, Ashu, that's really important and top of mind. But you've got maybe an investor sitting here looking at this and they're looking at this dashboard and they're like, there's no way I'm gonna be able to tell A computer how to build this. Like, I don't even know what to say. Like, I don't know the capabilities. I don't know how to ask this, how to give it the proper context, how to give it the guidance to build something as awesome as this. And so my question is twofold. Let's handle the first one first before we jump into like how you succeed. How do you fail with AI prompting? And maybe take Matter of Fact out of the picture for a second, but how do people fail with AI? How do you just write the worst prompts and, or And how do you just make sure that you only use the tip of the iceberg and you never get the full value of it?

And you wouldn't believe this. We had this one guy ask our system, you know, here is the exact query he said. I'm going to Australia. I need some quick money. Recommend me good stocks to buy. Right?

That's wrong context right there.

That's an example of what—

that's funny. I was so afraid you were going to pull up one of my queries. When you're first starting out, but that's good. No, not that bad.

Right? So I was glad that they actually said that they're going to Australia. So they had something very specific in mind. But yeah. Yeah. So this example illustrates that, hey, AI is not like a magical pill that can solve all your life's problems.

It is a tool after all, right?

And if you use the tool in the right way, you will have the right output. Think of like, Jeff Bezos just said yesterday on a podcast where he said, if you go to someone that's building their house and all they have is a shovel, and now if you go to them and say that, hey, I'm going to give you a bulldozer and now you can actually build your basement faster.

Right.

AI is that sort of a bulldozer. You have to still point it to something in the right way. Otherwise, it's just going to destroy value for you.

No, that's great. That's exactly right. And what do you say to those folks that are a little bit intimidated? Well, I don't know how to point it the right way necessarily. Like, sure, I have a couple of conversations with ChatGPT. Maybe I use it for shopping. Maybe I use it to plan my itinerary to the Maldives or whatever. How do I succeed? Like, how do I do that? And then maybe just segue right into how you help folks succeed with that.

Yeah, that's a great question. So we see this all the time. When people come to us with like sort of vague queries and they're expecting very specific detail in the answer. So here is what matters, right, the state of the AI that is in right now, it is not like an expert in doing financial research or analysis. You have to spoon-feed to it the framework or the set of things that are important. So what are you looking for? what does a good research analysis on this topic look like? That is sort of what AI needs right now in terms of help or input from the user. As a matter of fact, we have tried to solve for that because what we see is what is good for a given research question varies significantly across sectors. Every company, every sector has its own set of the key important things.

It's specific.

What matters for a biotech stock is going to be very different from what matters for a consumer staples company, for example.

Or a semi, right?

Yeah, or a semi company. So in our system, we have prebuilt 1,000+ skills in the system so that if you ask a question, so far we see almost 95% coverage in terms of queries. The questions the users ask, 95% of the times we have a skill that has been prebuilt to like help the AI guide the system. Now, if the user wants to be a lot more specific, they can be, but this thing solves for the problem of the user, you know, just not having the prompt engineering skills to write a good prompt. The other thing that we have done is we have tried to make this—

That's what you're calling playbooks, is that right? Playbooks where you select the job and that's where you say you have 95% coverage, basically all the workflows that hedge fund analyst or hedge fund manager might have during the day. You just select one of the workflows, you go in there and it gives you the prompt to provide the model the right context and you tweak it.

Yeah, you got it right. So we even upleveled these skills to say that like, you know, hey, let's just make it a very simple like click box. You enter the name of the ticker, you specify like, you know, you know, what sort of analysis you want, you press enter and boom. It's done, right? You want like a daily brief? Okay, great, done. You want to like, you know, track all the rumors about this company, all the social media sentiment about this company? Boom, done. If you want to like do, do read-throughs on this given company, boom, done.

Or if you want to build a scraper and find out how many cars are on the lot in Carvana.

Yes, yes.

You should be able to do that with a, you know, a nice prompt that tells the model exactly what to do, right? So that's what I love about this approach the most is that it takes the pressure off of the hedge fund analysts who already have a lot on their mind. You don't have to be really a prompt engineer. Sure, it helps. It helps to know. But if you just know what you need to do, your workflow, that's the starting point. Click and then you can tweak it. And I think it actually teaches you how to prompt better because they're like, oh right, of course you got to tell the thing the industry. It's not going to read my mind. Oh yeah, of course you got to tell it how far do you want to look back, 3 years or 2 weeks. It's not going to be able to to make an assumption, you know, just by itself. So that's fantastic. And, and I'll say, like, you've said it's 95%, and I'm super curious, what's the missing piece?

Like, what's the other 5%?

Is there, is there really something that, like, AI still can't do? And do you see that, you know, going to 100% at some point?

So, um, you know, we are always amazed at the different kinds of ways people look at things. And, like, you know, that is their particular edge. So we're always amazed that we find new kinds of customers, new ways of looking at companies, new sort of dimensions of looking at companies and like how you analyze things. So sure, we're at 95% right now and we don't expect ever to hit that 100%. Investing is a human sport.

Yeah.

Humans have the creativity, they will always find newer ways of like Of like finding edge. And that's where, that's where this whole game is. That's why it is so interesting.

I love it.

So the next question, it kind of dovetails on what you just said. And I do like asking it from a lot of the alternative data providers too, because the idea is that you've got alpha. So you've used an analogy or a limited alpha in a sense, right? So you've used the analogy of bulldozers. If everybody has bulldozers, everybody gets a better, faster basement, right? And they build their house quicker. But you've— it's an alpha is kind of like a zero-sum game, don't you think? Like if you've got a certain amount of time to exploit an opportunity or it's just capacity constrained, let's say it's in a small cap and there's a certain amount of dollars that can chase the idea, but then everybody has that insight. Do you see kind of Matter of Fact and tools similar diluting away the alpha and then eventually all these insights are going to become table stakes? Or does it create more opportunity? How do you think about that?

So, Sam, if you just look back 50 years, right, 50 years ago we still had banks, right, and these banks used to have like a floor of people whose job was just to run ledgers, right. They would, you know, these guys were the human calculators. When computers came into being, like the alpha did not erode away, it just changed the nature of the industry. It was like an expectation that now everybody would have a computer. Nobody's going to do the numbers by hand, right? So the expectations of the market changes with the tools that are available. Right now in 2026, people are feeling the gravity of AI. I think last week Ken Griffin on a panel was talking about how he was blown away that AI can now do things that a PhD in finance and maths would have taken weeks and months to actually do. This would have been a similar experience of the people before him when computers first came about. So it is just that in 5 years our workflows would look very different, investing as a game would still very much require everyone to be competitive, everyone to be creative, everyone to think out of the box, think deeply, debate about topics. In 5 years, well, the world might look extremely different.

Absolutely. I like how you said that investing is a human sport. The asset owners still want to talk to a human being, even though their tools are improving on both sides of the market, right? The managers have better tools and the LPs are going to have better tools and better document processing and all that stuff. There's still a lot to go. Now, one of your products that I'm hearing people just raving about Is this podcast thing and this ability to search through millions of podcasts? I guess you have now 200 million. Why should hedge fund managers be looking at podcasts? Is that one of the tools that you think hedge funds should seriously be paying attention to?

Stan, that's a good question, and I'm gonna reveal some alpha here. Yeah, yeah, let's do it. Let's do it.

Only on the Momentum podcast, folks.

Yeah, only for the subscribers of Momentum. Just the subscribers, right? So, you know, if you are a listener of podcasts, you would see that like the people that come on these podcasts are actually quite high quality people, right? Podcast as an ecosystem compared to the web is still full of people that, that have certain amount of credibility, that have certain sort of insights to share. Just tracking these podcasts can give you a good sense of like, you know, what these people are saying. It is just that so far like podcasts were all audio and like nobody could listen to like 14-hour-long podcasts ever. You would find like some really amazing people on podcasts, like people from the local governments, people who are in the regulatory bodies, people like CEOs, board members, industry executives, Like every kind of people. Dwarkesh had Jensen Huang on it and he asked a very hard question. So podcast has a lot of things going on, a lot of analysis going on and it is a great place for people to track what these discussions are, where the industry is divided and what are people debating. It's a good way to have a pulse of the market. What we see is that the sort of top-tier hedge funds that are managing like lots of billions of dollars are paying a lot of attention to these podcasts. Gavin DeTolf just said, I think yesterday—

Yeah, I heard that.

Right.

In one of the podcasts, yeah.

Right. So like—

He basically said, for the folks that didn't see it, we should link to that podcast here below. But he basically said that one of the most interesting ways of finding alpha is to scan through thousands of podcasts. I don't know the number that he said, but he basically said like there's a lot of alpha in the podcast, we just need a way to gather it all.

Yeah, yeah. So now you're a smart investor, you have access to this tool that can actually like, you know, you know, scan through hundreds, like millions of these podcast episodes and get you the insights that you're looking for without you having to listen to like millions of hours of these podcasts, right? So there is alpha right there, you know. So right there, like, this is, this is the new, like, a dataset that you should be using.

I think, like, I, I've heard some people refer to this as, like, the new expert network, right? Because any question you can potentially think of asking an expert, someone has already asked in the podcast, right? Like, what can you possibly think of that someone hasn't already asked? If you want that personal one-on-one, go ahead and spend, invest weeks and invest a lot of money. And find the expert and he's going to be limited on what he can tell you. Or on demand, type in your question and you've got multiple answers from multiple sources. I mean, it's like the expert networks 2.0.

Yeah. And Stan, you should always remember expert calls can only be with the ex-operators. Podcasts are actually always with the current operators.

They're there now.

That's a good point. Yeah.

So what is your next move? You've been able to build this incredible product. That investment managers who are very tough to impress, I might say, are raving about and they're loving about. I personally get texts like, holy cow, this thing is the real deal. What is your next move? Like, what are you thinking of next in this market? You've told me yourself, like, you've got to keep on moving and the pace of your product innovation is very, very high. It's up there.

So what's the—

can you give us a little window into the roadmap, maybe a little bit without giving away too many secrets? What's like the vision for the next few years for you?

Great question, Stan. So we take ourselves the responsibility to make sure that we fully understand what is possible today with the current set of AI and where the sort of AI capabilities are going and if we can make those accessible to our customers. I'm sure no one in our client base has the time to go through Anthropic's Developer Day every single time or to sort of track all these labs, what they're doing. And what the open source community is doing and so on and so forth. So we ourselves make sure that we are putting ourselves in the right places so that we can use those capabilities and then package them in a manner that our clients can use. One place where I think there is a lot of value still to be derived, and this is just me, I think investors today are not utilizing all the data that is available to them. There is plenty of information out there that seems way too hard for someone to actually bring together just because when they were doing it by hand or they had to hire a team of 100 people or they had to do so and so forth. But now all of those things are actually possible. Right now, if you are a fund manager, I can tell you honestly, you'll have almost 70% of the resources that a person at Citadel would have, right? It is now up to you.

That's exciting.

To just take it away, right? AI can actually scrape websites, structure them, alert you when something happens, can actually build out all these models for you, can tune them, right? Like all of these things are actually possible now, right? Which would've taken like an army of engineers, analysts, infrastructure folks at Citadel to actually manage, right? So if you are a fund manager right now—

It can look impossible.

for a small fund manager. It's, it's like, it's just not in their mind. And that's, I think, part of the, like, evangelical work, like evangelizing that you're doing is like, guys, open your mind. Like, even if you're a small manager, if you're like, like under a billion, that's okay. Like, you can do it yourself. Use the tools that are there, and you're going to be as powerful as the folks sitting right next to Ken Griffin.

Exactly. And I would love to add one more thing to this. Like, when you are making a choice of these tools, go with a vendor that has your interest in mind, that is not locking you up for a long period of time just because that is how the industry has been. This is the time where the innovation, where the disruption is happening. So I would go in with an open mind and see like, you know, what can a system do to actually accelerate my process? And like, just like, you know, rethink that, right? Because I can tell you honestly, in 5 years' time, like Ken Griffin will have every analyst at Citadel using it.

Yeah, for sure. No, I have no doubt. Absolutely no doubt. Well, yeah, look, folks are realizing they can be super analysts, like they can become superhuman, cover more names, get deeper insights very quickly about different industries, have more knowledge about other industries that they just don't have time normally to learn. Now it's like that it's available to them. So I actually am a big believer. One of the reasons why I'm so excited to have you on and share that, because you're seeing that change happen firsthand.

How do you—

you've built so much there, matter of fact, And it's moving so fast. How do you keep your team motivated to come in and do the hard day's work? What's that? What's like— it's more of like a little bit about more of like a business question and a personal question. How do you find the motivation? And then how do you give that to your team for them to continue operating at this level, at this caliber?

So, Stan, you know, what motivates me is when I have these calls with our customers and they say that like what they just saw is actually unbelievable and they take it to their MD right away. So that is the bar that we're now hitting. We hear this multiple times in different conversations where folks are like, I like your product the most. We see when people are actually able to use our system in ways to actually build out their strategy. We have this user who has built out their entire portfolio with all kinds of backtesting and so on and so forth all on our system.

Crazy.

And they have now converted that into a live dashboard where they're tracking everything. So it's like they went from the idea to actually testing to now having it live monitoring running. That sort of a thing makes me really happy to see our customers getting value out of the system. And that is exactly where a founder, where an entrepreneur wants to be.

Well, you're giving them the ability to have their ideas actually take flight, to have their ideas become reality with their own hands. And I'm sure, like, they built that system with the help of Matter of Fact, but they're probably very proud of it, right? Because it's their idea. They see themselves in it. They've built it. It works. It's helping them, and they love it. So if you're— just to wrap it up, and we're running up on time, but this has been an amazing conversation— what's the one thing you would leave with? The one idea that you want to get across to fund managers? And, you know, maybe they've dabbled a little bit, you know, dipped their toe in with AI, or maybe they tried and it kind of didn't work, they fell flat, maybe they didn't know how to prompt. What's the, what's the last idea, the parting thought that you would leave them with?

I would, I would really implore them to go and watch that video of Ken Griffin, right? The person who is at the top of the industry, who has seen everything, who, who is like, you know, who has a company of the smartest quants in the world, is getting impressed by, by a system, right? Yes, AI might not be like, you know, you might have experience with some product and, you know, it might have tainted your vision of what these systems can do. I would go in with an open mind and just refer to that thing that he said. If he has been impressed by it and you have been not, there is probably some other tool out there, or probably you asked a question that I'm going to Australia and I need to find a good stock to buy.

I need some money quick.

Yeah, I need some money quick. So I would just say go with an open mind, Within 5 years, like, this is going to be the new norm for software engineers. This has become the new norm, right? And the, you know, the pace at which AI is improving, right? 2030 might be still too far away. And we might still, you know, we might start seeing improvements much sooner.

This is awesome. Thank you so much for your time today, Ashu. Always a great, great pleasure chatting with you. And for the investors and for the folks that are interested in trialing your tool, where can they find you? How should they reach out? We are going to have a LinkedIn, your LinkedIn in the description of this video. Is that the best way or how should they find you?

You can go to our website matterfact.com and just sign up over there. Or you can email me at myfirstname@matterfact.com. I always personally do onboarding for all our clients. Would be more than happy to do that.

Ashu, thank you so much for being a guest on the show. Everybody else, thank you for tuning in and sticking through the rest of this show. We love you. See you on the next one.

Thank you, Stan, for having me. Thanks.

Thanks for tuning into Momentum. If you found today's episode valuable, don't forget to follow the show and share it with your network. This helps us a lot. To explore how AI can help your firm, visit academia.im. I'm Stan Altshuler, and I'll see you next week with more practical insights to help turn data into growth and decisions into results with the help of AI.