How to Build INSANE Stock Research Dashboards with Claude

0:00The prompt was pretty simple. I want to build a dashboard to keep tabs on SpaceX. It's an interesting reminder that earnings can beat, yet other pieces of information can spook the markets. We're spending 6.4 times more into CapEx than we've got in depreciation. Early versions of AI were very much about working backwards to the answer it thought you wanted to hear. And it is actually flagging this as something that looks like it's almost unachievable. This was put together in an hour. We're pulling data direct from SEC Edgar. It's so easy to pull together and it's ostensibly free. Welcome to all our listeners.

0:38Welcome to this week's episode of What's the Big Deal? Let's dive straight in, Graham, and tell everyone what is the big deal that we're looking at this week. So today we're going to follow up a little bit on SpaceX. Obviously, we talked about SpaceX a couple of months ago in the run-up to the IPO. one of the most, dare I say, overpriced IPOs in history. I mean, you're going to have your own view on that. But we finally have SpaceX's first set of quarterly results as a public company. And we wanted to both take a look through those and also continue on the AI theme and talk about some of the applications of putting together a quick dashboard for a company like this, if you want to stay on top of it.

1:20Absolutely. As you said, we did look at the SpaceX IPO. We said it was going to be a wild ride. If you were someone who bought shares in the SpaceX IPO, definitely it's been an interesting few weeks for you. The shares did rise and now they've tumbled a little bit. So let's find out what the latest sets of results are telling us about SpaceX's business. And you mentioned that we're going to be building a dashboard. Indeed, indeed. And, you know, it's something and we were talking about it before we turned on the record button here and started filming the episode. But I think this is an area where you and I in our previous professional jobs didn't really have a use for dashboards like this.

2:01I mean, dashboard is a loose term and we've used these in a bunch of different ways in the sense that I looked at a bunch of portfolio monitoring tools to look at a kind of credit portfolio. You can in some ways say a giant quarterly portfolio review book with a couple hundred sheets in it is a dashboard in some sense. I think the thing that's interesting about some of the tools we have today is it's a lot easier to build these tools. I also think one of the really fun things about where we are is we're kind of in this era of democratization of a lot of these tools as well. So everything that we've used to put together this SpaceX dashboard today is literally just from publicly available information and for free.

2:42I mean, we lit a few AI tokens on fire to build it, but compared to some of the other stuff that gets built, it's a pretty light lift. So I actually think it's kind of an interesting use here. Absolutely. So let's start off with what you actually did, Graeme, because you set the prompt running before we started recording. You've already got the dashboard built. Talk us through that, and then maybe we can talk a little bit about the data sources. I think that's a really interesting point. Yeah, so the prompt that I get, I used Cloud Code to build this, and the prompt that I gave Cloud was admittedly pretty simple.

3:18I mean, it's been a busy week teaching, so I haven't had hours and hours and hours to put a ton of research into the exact prompt I used here. And I would definitely go back and refine. But the main prompt was I want to build a dashboard to keep tabs on SpaceX. And I'm particularly interested in a couple things. We have, in the last couple months, been talking about aggressive revenue recognition policies, especially as it relates to companies like OpenAI and Anthropic that are also going to come to market sometime soon. So I said, I'm interested in a dashboard that is going to let me know what kind of revenue recognition policies SpaceX is using and if this stuff is changing.

4:01and I'm also interested in the cash burn and the CapEx spend because you know obviously if anyone's followed the news SpaceX from from memory the the results they released on a on an earnings basis beat analyst expectations but the price fell and the price fell because just CapEx spend is like is out of control so that's the stuff that I'm I'm really interested in at a high level anyway Now, Sandy warning, I have not had time to go through and audit this. If I wanted to actually put this together for real, I would do a lot of manual checking to make sure that actually what's being presented here is correct.

4:42And there are some numbers as we go through where I had to even reprompt and say, okay, what did you do there? Because I see some multiples that just don't necessarily make sense and aren't explainable. And then I got the explanation back and said, okay, that makes a bit more sense, but it's not the data that I'd want. So there is a lot. Anytime you're going through an exercise like this, there is a lot of fine-tuning you have to do manually through reprompting to get exactly what you want. Okay. Interesting points on SpaceX about the numbers looking a bit crazy, because let's be honest, they have a tendency to look a bit crazy anyway.

5:19What are you talking about, Debs? You know, 100 times multiple of the revenue. But yeah, let's have a look at what Claude has done for you. Let's first of all just have a little overview of what's in the dashboard. What have we got there? Yeah. Okay. So right up front, we've got this section here, how fast is the cash going? So again, this is exactly what I had prompted for and kind of what I'm looking for. So we've got free cash flow related to revenue. So Q2 revenue, 7.8 billion, but negative 25 billion of free cash flow. Also interested in where the cash is coming from, what's the cash balance?

6:02We've got$100 billion of cash and marketable securities on the balance sheet. Interestingly, and quite importantly, if you can see the little footnote, little note below the figure, we've got that as before the $60 billion cursor close. One of the things as I've started to use these tools that I think is really interesting and helpful is the fact that we are now to the point where we're getting a lot of contextual awareness around stuff. So I'm not just seeing, here's the cash balance, here's the CapEx spends, but also in the context of, we know they've announced a cursor acquisition, here's how much it's going to cost.

6:41So we're kind of factoring that into our projections here. Actually, that's pretty cool. I think the cash burn is a really interesting point. We can see they're using up a lot of free cash flow at the moment, or they're using a lot of cash flow, I should say, for their CapEx. As you say, they've also spent a huge amount on the cursor of acquisition. So not sure how much the remaining cash, how long that will last for them. So that's an interesting point. And as you say, a key driver into the share price reaction. It's an interesting reminder that earnings can beat, yet other pieces of information can spook the markets.

7:15So I think that's really useful. So what else have we got besides the headline numbers? We've got some charts below. What are they telling us? Okay, we see operations paid for 12 % of CapEx. Okay, so I would need to check this, but one of the interesting things that we're seeing here that I didn't even have to prompt to get was a chart saying, trying to figure out how much of our CapEx is being covered by operations. So we see operations pay for 12 % of CapEx. So obviously, most of this cash is not coming from cash that's generated by the actual companies coming from external sources. And as we go down below, we've actually got a bridge to kind of piece together what the various sources of cash has been.

8:02Okay, so we've got CapEx by division. Not shockingly, all this CapEx is being spent on AI. I mean, if you remember from the SpaceX prospectus, you kind of had, it was like, I want to say the page was the largest TAM in human history or something like that. And it was like space, space, space, space, AI, just business services, just huge. And well, they're following through on that promise. And most of this cash is going to AI spend. August 24th to 26th, we're running our investment banking AI bootcamp. Three days in New York. This isn't theory. It's a direct adaptation of the AI training that we deliver inside the world's top investment banks, taught by the same instructors.

8:46modeling, diligence, data analysis, deal materials. You'll leave with a prompt library, our AI for IB certificates, and 27 CP credits. Details at wallstreetprep.com. Use code WTBD. Okay, so who's paying for it? So here, yeah, here we go. Here's the bridge. So we've got IPO. Now, Deb, is this right? IPO of$86 billion? What is it? Okay. That's right. It was 85 plus the green shoe. So that's the right amount of fund raise. Okay. Okay. All right. Feeling all right about that. All right. So new net debt, 12.4 billion. Other financing, 2.2. I have no idea what that is. And then we've got the CapEx spend for last quarter.

9:34Securities down six and a half billion. I don't know what that is necessarily i don't know if we're saying is i don't know you have an idea investments i've no idea no sounds to me like they've spent it on some investments but yeah yeah yeah okay all right question yeah question question to clarify so you know again as the case with all these with all these tools there's still there's still a next next level down you need to you need to go to to figure stuff out if i were building something properly by the way what i'd want is a like a clickable interface where I could click on each one of those bars and get a bit more explanation.

10:11And by the way, I have no doubt that I could actually build that, which is kind of cool. Okay, is the revenue real? Now, this I had flipped through before, and this is where I was saying earlier, I'm seeing some multiples that need a bit of, not a bit of, need just any kind of clarification and context. Let's see revenue run rate gap 3.2 times claim. Okay. So claim of a hundred billion versus implied 31.3 billion. I guess what we're saying is implied is just last quarter multiplied by four. We had about eight, eight and something billion, billion dollars of revenue. So again, I'd want a little bit more just clear explanation for that.

10:54One of the really cool things here, I think is I know we need some context around some of these, some of these multiples, but looking at this revenue run rate gap, we've got this claim of 100 billion versus implied 31.3. And 31.3 is, I think, an annualized figure based on the quarterly results. But the 100 billion is an Elon Musk claim from an earnings call. So we're actually getting really insightful data here. This is not being pulled just from the 10K. This is full market contextual awareness. And it's pretty, pretty interesting. Let's see what else we've got, uh, day sales outstanding. I mean, I'm not necessarily super interested in that from the, uh, you know, from the perspective of kind of SpaceX here.

11:38Um, I mean, I don't really have a view on it to be, to be totally honest. Uh, but CapEx and DNA looking at CapEx as a, a multiple of DNA. I mean, obviously, you know, one of the things we guide people to, when we talk about financial modeling is at some point when you reach a mature steady state, you expect this ratio to be about one for depreciation and amortization to catch up with CapEx. Here we're saying we're spending 6.4 times more in CapEx than we've got in depreciation, just kind of highlighting the pace of the growth CapEx spent here. Yeah, this is definitely a growth company. You can really see that.

12:14We talked about CapEx being huge and you can see that in the multiple that it's way above what we'd see for any normal public company 6.4 times is huge exactly exactly and you know here's i kind of like i kind of like the bridge we've got just down below we've got some we've got some commentary saying the claim needs two more quarters like nothing sorry we say the claim needs two two more quarters like nothing that has happened yet sorry it's a bit bit verbose but basically saying in order to bridge to this 100 billion we need to see we need to see growth that we just haven't seen before. We've got this kind of pre-IPO, pre-IPO gray box here, and then we've got the implied step up we need to get to this$100 billion in revenue.

12:58So it'll be interesting to figure out where Q3 comes in and if we're anywhere remotely on track to get to this$100 billion. And it's interesting to see that skepticism coming in because actually early versions of AI were very much about working backwards to the answer it thought you wanted to here and it is actually flagging this as something that looks like it's almost unachievable uh based on where they are to date versus what they're guiding to so that's quite interesting i like that yeah you know as a just as a quick aside one of the things i've found i found interesting now i don't remember maybe we're talking about i don't remember which firm this is but i believe either you and i have had discussion or had a discussion in the classroom in the last in the last couple months about one of the big investment firms as a non-voting investment committee member effectively has an AI reviewer just to kind of cut through some of the weeds, to look at things through a purely analytical lens rather than through the kind of human emotional lens you look through deals with sometimes.

14:07And I do think, you know, not that we're necessarily to the point where we're going to trust any of these things with a full financial decision yet. But it is interesting, I agree, to see that skepticism coming through and kind of jumping to a conclusion or a question rather to say, all right, is this real? Is this obtainable? Because the numbers right now make it look like it's not. So what else have we got in there? Let's keep moving down. Okay, so what else here? We've got, let's see, growth. Growth is connectivity plus a sudden AI quarter. I need a little bit more context on this header here.

14:45So, okay, we're seeing if we look at the three main segments, got a segmental revenue split, got space connectivity and AI connectivity has increased and AI, to be fair, AI revenue has gone up quite a bit. But the thing we need to put in context, obviously, or keep in context is AI revenue versus CapEx spend. Yeah. Okay, the add back wedge keeps widening. okay so that was the non-GAAP adjustments is it yeah so when i went through the q2 results as obviously a huge bridge to get from GAAP net income to adjusted EBITDA they are adding back say SBC stock-based comp to get to adjusted EBITDA i mean certainly actually i was having a discussion about this in the classroom that i was teaching this week okay what's what's your view on adding stock-based comp back, at least as it relates to adjusted EBITDA or say in QV adjustments that you're looking at when you buy a company.

15:42And I mean, ultimately, this is a real expensive business. SpaceX is paying this stock-based comp out to employees because it needs to pay those employees to retain them. And the fact that we're just adding that, in essence, we're just adding compensation expense back to EBITDA to come up with some adjusted number is pretty disingenuous. But a lot of people do this anyway. What's your view on that, Debs? Yeah, I mean, from a cash flow perspective, I get why companies do it. You know, if you're a credit analyst, you definitely want to be adding back to stock based comp because it's not a cash flow.

16:12But from a performance appraisal perspective, I would definitely, you know, keep it as a deduction. As you say, that expense, it captures the fact they're giving away a bit of their business each month, each year to their employees. And if I was a shareholder in SpaceX, I'm not. But if I was, I would care about the fact that a bit of the company is being given away every time they grant stock options to their employees. and that's what's captured in the expense. So I agree. And it's always, from a QOE perspective, quality of earnings, it's always a concern when that gap is widening. So again, it's something that the dashboard is highlighting, the wedge is widening, they're making more and more adjustments or at least the stock-based comp is growing and the other adjustments are growing.

16:52And that's always a bit of a concern, particularly when you add it to the concerns around pressures on the run rate, revenues, not maybe quite getting there, the other levers that management can pull are some of these adjustments. So yeah, great that it's flagged that. By the way, I don't know about you, there's nothing in here yet that shocks me based on what we know about SpaceX. We've got some aggressive, aggressive revenue forecasts. We've got some aggressive adjustments to get to adjusted EBITDA and all this kind of stuff. Like none of it shocks me given what we know about SpaceX and Elon.

17:28Absolutely. Okay, so we've got summary, income statement. So we've got the bridge from income loss from operations to adjusted EBITDA. So let's see, Q2 2026, we lose 143 million at the net income line. And then our adjusted EBITDA is three and a half billion. Sure. Great. That sounds awesome. And then we've got CapEx of 18.4 billion so all these numbers pale in comparison to the capex number like that yeah ultimately that is my that is my main read and takeaway here and kind of falls on the theme of everything we've been talking about the last few months here absolutely um and if we keep scrolling down at the end of the dashboard is there more to come oh there's a balance sheet okay balance sheet uh you know the one the one thing i asked for okay we've got some analyst notes here uh the one thing i asked for just for fun and i'd go back and uh and redo this a little bit i said i would like to see a chart for capex spend of elon launching cash on a rocket into the sun so we've got we've got something something kind of like this here looking at our growth in capex spend so we see we got a little a little elon stick figure riding the riding the rocket ship and then let's see the pile the pile he's flying into 100 billion dollars of cash and securities okay this needs a little bit of this needs a little bit of rework he's not really flying into it he's flying that's that's the cash pile that he's like are you comparing him to icarus he's basically so confident i love that analogy um okay great you know though i like that i like the footnotes here at like altitude is linear in dollars the shapes are honest even at the physics is not this is so good so clon does have a sense of humor 100 100 oh my goodness that's the best thing clon has produced it gets an a star just for that right so um look if anyone was using this for the real world it needs it needs some it needs a fine-tuned past for sure right but i do think i do think one of the really interesting things here is if you if you zoom out for a second and kind of think about how these tools are democratizing everything.

19:56Like when we were working in this job, I mean, I say this job as in a job where you need some kind of market data input. And for me, it's been quite a while since that's the case. I mean, it really hasn't been since Lehman Brothers days from obviously pre-2008. But even in those days, I mean, what tools are available? You had FaxNet, you had CapIQ, you had Bloomberg, you've got all the really expensive big data providers. And admittedly, they're probably still better because you've got professional teams scrubbing through numbers, but they're also insanely expensive and they're out of reach of just the individual consumer.

20:35So this was put together in an hour, right? And we're pulling data direct from SEC Edgar. The SEC's got an API that we're interfacing through direct here. We've got some contextual awareness around market commentary, analyst commentary. So given this is, it's not perfect, but it's so easy to pull together and it's ostensibly free. I do think it's a really interesting, really interesting use of the technology here. Absolutely. I mean, some of these APIs you mentioned and like the SEC one has been around for a while, but you needed to be a coder to be able to, build you know tools that could build pull the api feed in um whereas the beauty of claude now is and so on all ai tools really is that you can get you know the ai tool to build the code around that you can make a dynamic dashboard which is what we've got here it can automatically update when new results are available and also that the ai feed it pulls in a lot of fundamentals data that's stuff that's actually reported in the financials but as you said analysts tend to have to scrub those numbers, make adjustments, contextualize the information, whereas AI can now do that for us.

21:49So it's really democratizing access to data, but also data that's usable. It's not just, you know, crude raw data. It's data that's been cleaned, it's been improved for analytical purposes. So I think that's amazing. Interesting that the dashboard itself doesn't have any actual share price information. That's one thing I would want to see. And, you know, from my background in markets, you know, PMs that have, you know, got dashboards of their own portfolios, they would want to mix fundamental data with market data information, you know, data on, you know, big, large volume trades that have happened, share price movements, other market events would all be included in the dashboard.

22:29Whereas this focuses very heavily on the kind of the fundamental side, doesn't it? Oh, 100%. And to be fair, that was the prompt I But if I were actually looking for something to make a public investment decision based off of, then 100 % you need that. And especially for a company like SpaceX, things I'd be interested in seeing there are a proper annotated share price graph. Because we've got some pretty chunky events here, right? We've got the IPO. Then we've got lockup periods ending. We've got the admissions, the index, the forced buying, the kind of sell-off. So there's just a lot of stuff happening all at the same time.

23:08And we had some pretty big movements up and down and kind of putting some context around those and when and why and what happened, I actually think is quite an interesting part to the story. And I mean, from your markets perspective, what do you think analysts are going to need to see or investors are going to need to see before they finally, I'm not going to say before they finally, because it's not a foregone conclusion this is going to happen. But for ultimately them to get a bit more comfort and confidence on this capex spend and for them to change their opinion a little bit. Well, I think ultimately it's when the actual returns start to come through and the capex spend stabilizing.

23:47At the moment, you just, you know, a lot of these companies are just throwing cash at what looks like a kind of a very expensive project. The sun? Yeah, but I know there's no actual returns being generated. As we saw, they're generating profits, but not from the AI business. Most of it's coming from, it's being kind of subsidized, if you like, by other segments. So really what we like to see as analysts is actually to start to see some of the benefits of that or the returns on that investment coming through. And there's just no evidence of that. And kind of the big risk, while it's still quite nascent, is that you throw a lot of cash and you're not the one that enjoys the benefits.

24:29that you don't know where really in the value chain or even in terms of the competitive horizon, who's going to be the winner. So it could be that you're throwing cash and it's just being thrown away. It's like just a sunk cost. So we need evidence that this is generating returns and we just don't have that yet. And I think that's why the market's getting nervous at the fact that it's not just the capex spend to date, it's the fact that capex spend is continuing to ramp up. Now, if anyone's going to do it, Elon will. He's clearly very ambitious, has big targets, but he's stated some of those targets.

25:06And at the moment, we're seeing that targets maybe are too ambitious. They're not going to be hit, at least in the short term, unlikely at least. So it's interesting. But yeah, at the moment, we just don't have the evidence. And that's what the market really wants. They want data that they can hang their hat on, as it were. Well, we'll see. I'm curious to see what happens over the next 12 months, next nine months, when we get a real proper view on first-year results, see how right or wrong Elon's predictions are. Do we grow into this$100 billion revenue figure that he keeps touting around? We shall see.

25:42yeah we shall see we use some tools like this to to put a bit more easy context around around some of the numbers and keep keep on top of what's going on i'm not going to make an investment decision based on this but it's an interesting thing just to just to take a look through and to play around with a little bit yeah and you say you're not going to make investments decisions based on this but i do know people who are starting to build not necessarily dashboards but and kind of they're using AI to build instructions for their own portfolios for certain metrics and factors that they would want AI to trade on their behalf of and then kind of trialing using AI as their kind of trading idea generator and then actual for executing transactions so I think that's a really interesting space for retail investors because as you say we've now got democratization of data, which allows you to do that as an individual in a way that you couldn't previously.

26:43So I think that's a really interesting space is using AI for personal investment decisions and potentially not just the decisions for yourself, but AI making those decisions on your behalf. And so definitely that's one to watch. Interesting. Interesting. Yeah. I haven't trusted it with actual cold, hard cash yet, but we'll see. No, not with my own money, but we'll see. Yeah, exactly. Okay. So I hope you enjoyed this week's episode where we asked Claude to build a dashboard to show us SpaceX's performance, particularly focusing on its recent results. I think you'll learn a little bit about how it's performing and also hopefully learn a little bit about what's useful in a dashboard.

27:22Thanks so much for listening. And that's all from me. Thanks, Debs. And we'll see everyone next week. Take care.