Paint my hair. What the hell? Listen, sometimes I double check with AI. Okay, it's just a double check. We do have outside investors. We've raised around five million so far. We're raising another two and a half out of this SAFE. It's a lot of you, a lot of people want in on this round. So, we're growing quickly and I'm very happy to have a full email box of investors.
Listen, I can't look at every single stock. I wish I could, but PayPal is one I do want to look at. So, this is going to be a little painful. I'm going to budget two hours. Is that okay with you guys? Is that too much?
Well, we're raising money. If you're interested, you can invest. Just email me. 50k minimum. We're raising at around 20 times revenue, so it's pretty attractive. I think there's private startups going for 100 times revenue.
Yeah, it's a complicated business, right, folks? I don't know if Godel will go public. Yeah, you can invest. Yeah, I'll let you in for 25. Let you in for 25.
I just don't like talking to VCs. So, I think VCs are the worst. I've raised about 500 my whole life. I would say 495 from non-VCs, public company investors, hedge funds, mutual funds, etc. Insurance companies, corporations, high net worth. There's plenty of money outside of VCs. There's my email.
I think that, you know, the problem with our ideal, we know who our ideal buyer is. We've already talked to two or three potential acquirers. I love this space because it's so acquisitive. Now, it doesn't guarantee success at all, you know, but there's like, if you look at ICE, Nasdaq, S&P Global, FactSet, Moody's, maybe like the five huge financial technology companies. I don't include Bloomberg for obvious reasons. A lot of these companies do like 5, 10, 20 deals a year. They just buy whoever out, like Visible Alpha. Okay, it got bought out. That is one feature of Godel. Now, they did it really well, but Visible Alpha got bought by S&P Global for 500 million.
No, we don't care if you're US or foreign investors. That's a good idea, Don.
So I think, you know, there's a handful of like, it's almost like an oligopoly. There's just a couple of big chunky fintech investors. I think Bloomberg is here to stay. I mean, I think that...
I'm working finalizing a component that is going to really change the way I think charts work. And hopefully it'll be out in the next few weeks. And it's something Bloomberg would just never do. They're just too far behind. It's a lazy company. People work there till 5:00 PM. They're not working there till midnight like my guys. They don't have the heart. You know, it's an incumbent.
And I think some of the things that we're going to do are going to be more expensive than Bloomberg. So the system I have requires four H100s to work. So that's 5K a month. That's more. Our costs are more than Bloomberg's price. So Bloomberg costs Bloomberg around 200 bucks, maybe 300 to run their servers and their data, and they sell it to you for 2500 a month. This system that I'm making is $5,000 a month in my costs. So it's going to be really like, what are we going to charge for it? 50k, 40k, 30k. It's really meant for institutions.
So, we'll see. Newegg picking back up here. I might short it again. Yeah, there's colocation. There's all kinds of things that, you know...
>> He'd be in jail. All right, let's focus on UNH, though. Yeah, GPU pricing will go down. We'll figure out how to get it a little bit more efficient. But here's the thing, like the system I'm making, it's working on minute data, and I want to get it to second data or tick data, which I may need literally hundreds of GPUs for. And D.E. Shaw, okay, they have a super. You guys know about this? This is the alpha scrolly gives you. Okay, let's just call him Anton. It's called Anton. It's really good.
Massively parallel supercomputer. So quants have machines like these, but I think some don't. And you know, we have this very weird skill where we know how to work with GPUs because some of our first projects were AI things. So we know how to deploy GPUs with Ray and Kubernetes and all that stuff. And you know, we can... yeah, it's exactly anonymous. It's something like that.
And you know, you really, if you want to look at the entire stock market real time with tools like this, I mean, 240 GPUs, maybe 500 for like really real time, like millisecond style data or sub-second data. So 500 GPUs is not a cheap cluster, and I'm not sure how many users could share it. So 500 GPUs, if you're a big hedge fund, I mean, I think you need it, but 500 GPUs running 24 hours a day, 30 days a month. I got that right? Yeah. At $2, it's almost a million a month in costs.
So, it's definitely not a small thing. And I think a lot of hedge funds don't want to set this all up, to be honest. It's a lot of work, a lot of maintenance. I wouldn't use RTXs. I mean, you need the VRAM. I'd almost want the H200. More VRAM the better.
All right, let's work on UNH. I got nine minutes left on my upload. All right, UnitedHealthcare is a healthcare group. It's more about VRAM in this case and processing speed. You can email me martindl.software if we're really looking for New York people. New York people with financial experience, with Wall Street connections. Let's get to work.
Healthcare and well-being company. Okay. With a mission to help people live healthier lives and make the healthcare system work better for everyone. We have two businesses, Optum and UnitedHealthcare. Okay. Optum.
Yeah, you got to shard across VRAM. I've used some of the, I've used the two terabyte RAM machines. I feel like, you know, GPUs got the parallelism that you need here. So, but one of my colleagues thinks you could do it without GPUs, but I don't know. H200 pretty good, but need that high bandwidth memory.
