Data center buildouts face financing bottlenecks, energy constraints, and insolvency risks for smaller players
Summary
Speakers argue that financing and infrastructure, rather than hardware availability, are the primary bottlenecks for data center construction, with hyperscalers like Meta and Google dominating due to their ability to secure equipment and energy. The episode highlights the massive power demands of AI racks, which consume ~300kW per chip, necessitating advanced cooling and driving natural gas demand. While electrical and cooling equipment providers benefit, the reliance on imported components from Taiwan and Korea widens the US trade deficit. Concerns are raised about the insolvency risk of private credit-backed entities like CoreWeave and the valuation risks associated with SpaceX.
Key takeaways
- 1 Eaton has been a massive beneficiary to this. So Eaton, for example, is they're doing the transformers at the last mile. So they're putting the transformer equipment into the data center.
Unidentified speaker (speaker_1) · 7:47
- 2 CoreWeave faces a classic insolvency risk with ~$9B cash against $20-30B in debt and estimated negative free cash flow of ~$27B over the next 12 months, requiring significant new debt financing.
Unidentified speaker (speaker_1) · 14:37
Episode sections
- 0:07 – 1:30
Financing Bottlenecks and Hyperscaler Dominance
Speaker 1 argues that financing, not just hardware, is the primary bottleneck for data centers. Larger hyperscalers like Meta, Google, and Microsoft are better positioned to secure equipment and energy due to their ability to write large checks and guarantee purchases, whereas smaller builders face longer timelines and tighter financing conditions.
- 1:30 – 4:06
Infrastructure Challenges: Power, Cooling, and Heat
Building a data center requires securing land, bringing in energy supply (taking up to a year), and installing equipment. Modern racks consume ~300kW per chip (10x older tech), generating immense heat that requires water cooling and efficient air conditioning to prevent component melting.
- 4:06 – 7:36
Local Pushback and Community Benefits
Data centers face NIMBY opposition due to noise and visual impact. However, companies like Meta are mitigating this by providing direct financial benefits, such as $50,000 checks to teachers in Louisiana for the Hyperion data center, turning local opposition into support.
- 7:36 – 9:06
Beneficiaries in Electrical and Cooling Equipment
Companies providing electrical equipment (ABB, Siemens, MCore, Vertiv, Eaton) and cooling solutions (Johnson Controls, Carrier) are major beneficiaries. Generac and Caterpillar are also benefiting from backup power generation and construction equipment needs.
- 9:06 – 10:49
Natural Gas Demand and Supply Dynamics
Data centers will drive significant natural gas demand, with 15GW of buildout adding 4-5 BBTU/year. This demand, combined with LNG export plants, may close the supply-demand gap that has existed since shale production grew post-2012, as US gas supply growth has stalled due to low prices and a plateauing Permian oil field.
- 10:49 – 12:02
Imports, Trade Deficit, and International Supply Chains
Approximately 80% of data center components (chips, memory, servers) come from international suppliers, primarily Taiwan and Korea (TSMC, Samsung, SK Hynix). This import reliance widens the US trade deficit and dampens the GDP impact of the domestic data center build-out.
- 12:02 – 13:40
Space Data Centers and SpaceX Valuation Risks
Space-based data centers face technical hurdles like radiation and launch vibration. SpaceX is identified as the most at-risk 'darling' stock due to its $2.2T valuation lacking the profitable business base of other hyperscalers. A potential Tesla-SpaceX merger is discussed as a way to unlock executive compensation.
- 13:40 – 29:07
Private Credit Risks and CoreWeave Insolvency Concerns
CoreWeave is cited as a high-risk entity with ~$9B cash vs. $20-30B debt and negative free cash flow. The speaker views it as a 'donut' (zero value) due to insolvency risk. Private credit funds are facing gating issues due to defaults in software stocks, reducing available funding for data center builders.
- 29:07 – 30:59
Conclusion and Sponsor Read
The speakers conclude the discussion with a metaphor comparing SpaceX's financial flexibility to other entities. The host then transitions to the podcast outro, thanking listeners and introducing the sponsor, Moran Wealth Management.
Assets discussed
| Asset | Kind | Stance | Context |
|---|---|---|---|
| NVIDIANVDA | equity | neutral | In fact, you know, you've seen inventories go up at both uh in NVIDIA and Broadcom. So they are delivering the |
| Meta PlatformsMETA | equity | neutral | So that's Meta, Google, SpaceX, um, Microsoft. They're the folks that are really getting all of this. |
| AlphabetGOOGL | equity | neutral | So that's Meta, Google, SpaceX, um, Microsoft. They're the folks that are really getting all of this. |
| MicrosoftMSFT | equity | neutral | So that's Meta, Google, SpaceX, um, Microsoft. They're the folks that are really getting all of this. |
| AmazonAMZN | equity | neutral | So when you look across the other hyperscalers, other mag seven names or the and the big companies that are valued that high, it's you know, like I said, it's Google, Meta, Microsoft, Amazon. |
| TeslaTSLA | equity | neutral | I heard today that the the the the merger between uh uh uh Tesla and SpaceX is in is is uh sooner rather than later. |
“Stance” is how the speakers talked about the asset, not a recommendation.
Claims worth verifying
Data center racks today consume ~300kW per chip, which is 10x the ~30kW consumption of pre-AI racks.
Unidentified speaker (speaker_1) ·3:01
How to verify: Check technical specifications for AI server racks (e.g., NVIDIA HGX systems) vs. traditional server racks; compare power density data from industry reports (e.g., Uptime Institute). For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Meta is building a data center called Hyperion in Louisiana and providing $50,000 checks to teachers in the local school system.
Unidentified speaker (speaker_1) ·4:52
How to verify: Search for 'Meta Hyperion Louisiana data center' and local news reports regarding teacher payments or community benefits agreements. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
CoreWeave has ~$9B in cash, $20-30B in debt, and negative free cash flow of ~$27B over the next 12 months.
Unidentified speaker (speaker_1) ·14:37
How to verify: Review CoreWeave's S-1 filing or latest 10-Q/10-K on SEC EDGAR for cash, debt, and free cash flow metrics; compare with Wall Street consensus estimates. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Primary sourceApproximately 80% of data center components (chips, memory, servers) come from international suppliers, primarily Taiwan and Korea.
Unidentified speaker (speaker_1) ·10:49
How to verify: Review US Census Bureau trade data for imports of semiconductors and computer equipment from Taiwan and South Korea; check industry reports on supply chain origins. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
15GW of data center buildout adds 4-5 BBTU/year of natural gas demand.
Unidentified speaker (speaker_1) ·9:47
How to verify: Verify the conversion factor of 180 BBTU/GW stated by the speaker; calculate total demand based on projected GW additions from EIA or industry forecasts. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
NVIDIA and Broadcom have seen inventories go up, indicating the GPU shortage is less of a problem.
Unidentified speaker (speaker_1) ·2:21
How to verify: Check NVIDIA and Broadcom earnings reports or investor presentations for inventory levels and guidance on supply chain normalization. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Primary sourceSpaceX has a $2.2 trillion market cap.
Unidentified speaker (speaker_1) ·25:40
How to verify: Check current valuation of SpaceX from financial news sources (e.g., Bloomberg, Reuters) or recent funding round valuations. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Two IPOs of data center companies were pulled due to funding difficulties.
Unidentified speaker (speaker_1) ·15:37
How to verify: Search news archives for withdrawn IPOs in the data center sector around the time of the podcast (Nov 2026). For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Permian oil production has plateaued, stopping the growth in associated natural gas supply.
Unidentified speaker (speaker_1) ·20:03
How to verify: Review EIA or DrillingInfo data for Permian Basin oil and gas production trends over the last 1-2 years. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
China is building nuclear power plants, increasing demand for uranium.
Unidentified speaker (speaker_1) ·18:52
How to verify: Check World Nuclear Association reports or IAEA data on new nuclear reactor construction in China. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
SpaceX does not need debt and has other businesses to run.
Unidentified speaker (speaker_3) ·29:07
How to verify: Review SpaceX's capital structure and revenue breakdown in private financial disclosures or credible news reports on their funding rounds and non-space revenue streams. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
SpaceX is an SEC registered investment advisor.
Unidentified speaker (speaker_0) ·30:15
How to verify: Search the SEC's IARD system or advisorinfo.sec.gov for 'Moran Wealth Management LLC' to confirm registration status; note that the speaker attributes this status to Moran, not SpaceX. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Primary sourceA copy of Moran Wealth Management's current written disclosure statement is available upon written request or at advisorinfo.sec.gov.
Unidentified speaker (speaker_0) ·30:35
How to verify: Visit advisorinfo.sec.gov and search for Moran Wealth Management LLC to view their Form ADV Part 2A. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Primary sourceDo your own research
- Verify CoreWeave's debt maturity schedule and cash burn rate via their latest 10-Q or 10-K filings to assess insolvency risk.
- Check Eaton's (ETN) order backlog and revenue guidance for data center-related electrical equipment in their quarterly earnings reports.
- Review Meta's (META) community benefit agreements and local political contributions in Louisiana to verify the $50,000 teacher payment claim.
- Analyze US natural gas supply/demand balance sheets from the EIA (Energy Information Administration) to assess the impact of 4-5 BBTU/year data center demand.
- Examine US import data from the Census Bureau to quantify the trade deficit impact of data center components from Taiwan and Korea.
- Assess SpaceX's valuation multiples relative to its revenue and profitability metrics, considering its private status and lack of public financial disclosures.
- Check private credit fund liquidity reports for gating mechanisms and default rates in software sectors to evaluate funding availability for data center builders.
Full transcriptexpand
0:00Unidentified speaker (speaker_0): This podcast is produced by Moran Wealth Management, an SEC registered investment advisor. Nothing in this episode constitutes his personalized investment advice.
0:07Unidentified speaker (speaker_1): We've talked about it earlier, but you're starting to see uh less financing available, and I think that'll be the bottleneck that surprises everyone with the data centers, is just paying for
0:16Unidentified speaker (UNKNOWN): them.
0:18Unidentified speaker (speaker_0): You're listening to the Quarter Over Quarter podcast. Each episode, your hosts, Tom Moran, and I, Don Drury, discuss.
0:27Unidentified speaker (speaker_2): different angles of wealth for those who are focused on accumulating it, retaining it, or having fun spending it. Welcome to the show. Gentlemen, we're going to talk about data centers and the bottlenecks there and some of the problems that are rearing its ugly head. Tyler when people think about AI, they think about software, they think about LLMs, but I picture and I hear more and more really large electric bills and potentially not enough power to even supply what's coming online. Who's going to get burned? Who's going to be able to eke through and make a successful entree into deploying these data centers?
1:30Unidentified speaker (speaker_1): In my mind, it's going to be the bigger hyperscalers who you can simply write a bigger check. And we're already seeing that happen. So there's a multitude of issues when you think about getting a data center out. So first of all, you have to buy the land. That's mostly already been done in most instances. Um, then you have to bring the energy supply. So this is typically going to take a year. It depends on which state you're in. But if you're in a state where it's regulated, you can go a little bit quicker. Um
2:01Unidentified speaker (UNKNOWN): you're
2:02Unidentified speaker (speaker_1): getting these. sites uh cited maybe a year and a half. The closer they are to generation, the better because you don't have to cite power lines to them. Just to put in perspective, some of these data centers are about a gigawatt. Um it's enough for a small city to to power them. And then not only that, but like when you build the data center, get any actual equipment to go in it. So there's uh there's chillers that go into that, there's cooling devices that go in that, there's the GPUs themselves. The GPUs are the big expenses, and you know, that's where the shortage was originally was just getting the GPUs made. Um, now that's less of a problem. In fact, you know, you've seen inventories go up at both uh in NVIDIA and Broadcom. So they are delivering the chips at this point, and so the real bottleneck is to your point, is becoming energy and powering these things. And not only is it getting a power line to the data center. it's getting the equipment inside. So the transformers to break these things down. To put in perspective, uh uh before AI started, if you built a data center, you would have a rack of servers and those things were running somewhere around 30 kilowatts an hour, which is uh take a bunch of but like maybe a typical house, right? They got a bunch of your computers on, your TVs on. The racks today, each chip is running to 300 kilowatts an hour. It's just a multitude of 10 times, it's it's 10 times bigger. And that also throws
3:21off more heat. So you then need to spread these things out. You need to start water cooling them. You need to uh build data centers that are incredibly efficient from an air conditioning and cooling perspective to keep the chips cool. Because if you do not, the components around the chip will literally melt onto themselves because it's so hot at 300 kilowatts. So those are some of the multitude of challenges. And the people that can get the equipment and write the bigger checks and guarantee, hey, we will buy um this commodity from you, this amount from you, you know, and give certainty to the vendors, they're the ones that are getting everything. So that's Meta, Google, SpaceX, um, Microsoft. They're the folks that are really getting all of this. Um, and then the danger really is the people that are going to be last to this,
4:06Unidentified speaker (UNKNOWN): are
4:06Unidentified speaker (speaker_1): they gonna get their data centers online in a reasonable amount of time? And you're starting to see that. So we've seen our force, our first force majeure from uh Oracle a couple of weeks ago, where Oracle cannot get the components to start the data center in time for as they've committed to. Uh, we're starting to see a little bit of that. And now uh for new data centers that have not broken ground yet, you have the non-speaking. not in my backyard phenomenon.
4:29Unidentified speaker (speaker_2): Nim NIMBY is around for sure.
4:32Unidentified speaker (speaker_1): These things create a horrible, dull buzz if you live within a quarter mile of them. It's a constant buzz. It's not a pleasant, it's not doesn't create a pleasant environment um for their neighbors. And so that's a big component for it. Now, at the same time, is they are spending so much money on these things that they're able to lift. up the community from a perspective of tax dollars. So we're um the famous example is uh Meta, which is down in Louisiana, is building this giant data center, and it's called Hyperion. And they have started cutting $50,000 checks to all the schools, to all the uh teachers in the school system. So,
5:10Unidentified speaker (UNKNOWN): you know,
5:10Unidentified speaker (speaker_1): there's flip sides to it as well. So the needless to say, the school teachers in that part of the Louisiana. Louisiana are ecstatic about the data centers.
5:17Unidentified speaker (speaker_2): Oh interesting. So last night we shot uh our our second episode of Smoke and Stocks. And Tony pulled out a fact. Yeah, I I'm not gonna say it pulled out a fact. He he said something that took me by surprise.
5:38Unidentified speaker (UNKNOWN): He
5:38Unidentified speaker (speaker_2): said that there's a data center in,
5:43Unidentified speaker (UNKNOWN): I
5:43Unidentified speaker (speaker_2): believe it was Utah, that was so energy intensive that it raised the temperature of the city. I think he originally said the state, and I and then he goes, Oh no, no, no. I said the city. Ten degrees at night. That seems improbable if all of these chillers, coolers are there to ensure the racks don't melt down.
6:20Unidentified speaker (UNKNOWN): Well,
6:20Unidentified speaker (speaker_1): they're actually, so what would be happening is they're the with the mechanism of the chillers chiller is they're moving the heat out of the building.
6:26Unidentified speaker (speaker_2): and and into
6:27Unidentified speaker (UNKNOWN): the air. So that could be it could be
6:29Unidentified speaker (speaker_1): now. I don't know. I haven't heard that story. Um it sounds it sounds fascinating. It it could be the case. You do have um architectural impacts on heat. So, for example, uh it's actually called a heat island because you build so many roads. So, LA, Atlanta, what you'll see is you'll actually see hotter temperatures, typically in the northeast side of these cities, and more thunderstorms because. as the weather comes over the city, there's so much glass and so much pavement it causes the heat to rise, and then it it actually does impact the wet the weather and atmosphere. So it it could be possible. It's crazy.
7:04Unidentified speaker (speaker_3): Now I uh I know you're concerned about you know overbuilding and and and then the amount of debt these some of these companies are taking on and uh and what they serve as the debt. Are there are there but there are opportunities in data center? centers and that the c the effici uh the c I always like use the analogy, the the miners didn't make money in the gold rush, but the people own the saloons, the hardware stores, the picks and shovels. Um we have a we have uh uh uh an AI portfolio and and we have uh uh stu some of the stocks include things that would be beneficiaries. Uh what areas benefit from the Denis diners data centers?
7:36Unidentified speaker (speaker_1): Uh so when you look at the let's look at the electrical equipment code going in
7:38Unidentified speaker (UNKNOWN): there.
7:38Unidentified speaker (speaker_1): So you have ABBs in there, you have Siemens's in there, you have MCores in there. The big one has been companies like Vertiv companies. Eaton has been a massive beneficiary to this. So Eaton, for example, is they're doing the transformers at the last mile. So they're putting the transformer equipment into the data center. On the cooling side, there's been a lot of uh companies that have done well there. Um the big one, Johnson control. Controls, has been a big winner. You've even seen things like Carrier do well. Um, and it's even moved beyond that. So companies that are even adding uh power. So the backup power. Recently we've seen Generac, which is known for its home standby generators, uh it it jumped a bunch the other day because they signed a deal with um, I believe it was with m Meta to provide uh generation capacity uh backup generators for their new giant dime data center. So you're seeing that and you're seeing that also with even Caterpillar. So caterpillars benefit from that because they're providing both generators and the excavators to move all this uh property to move all the uh dirt around to clear these data centers. Even the construction companies, there's a a great little construction company in Texas called Sterling Construction. I think it's up about
8:47um gosh, it might be twenty 20 or you know fifty 50x X since the beginning of this started, where they're cons doing most of the data center construction is in Texas, which has been one of the hotbeds because they're close to natural, close to the natural gas and some of the other resources there. Um, I think the next thing that will actually be the beneficiary of these data centers as we continue to build them will actually be natural gas. So as you watch that evolve, you can see the different waves of investment come in and people realize it. So when it first started, you saw the obvious, okay, we gotta get the chips. And you saw NVIDIA and the chips do well. And then you saw the oh wow, we gotta power these things. Um the utility started doing well. Anyone who had a nuclear plant started doing well. Um and then you saw okay now we got to cool these things since the chips are so hot then every all those providers got there. And then oh hey we have to have the backup supply for energy. So then that happened. And what's really interesting to me is if they they're could have going really to built really build 15 gigawatts this year to the let's just say they do 15. They're planning for 25 and it's going to be difficult to bring all of that online.
9:47But if it's 15 gigawatts, each gigawatt translates into a 180 um Bs of energy of natural gas demand. So for if you do 15 gigawatts, you're looking at adding four to five um
10:01Unidentified speaker (UNKNOWN): uh
10:02Unidentified speaker (speaker_1): B's of supply uh uh a billion BTEs of supply a year. Like that is a monster increase in demand for natural gas. And when you combine that with the LNG plants that we're currently building, which should add another seven or eight, you're probably gonna close the gap there between supply and demand that's been that's it's been oversupplied for since shale came online really in 2012, 2014. So I think that that's probably the next opportunity across data centers. It's it's actually going to be powering the power plants to um to uh to run the data centers. Now
10:38Unidentified speaker (speaker_3): I know you're bullish in international summary. I heard that uh half of the components going into building these tech stacks are coming from international suppliers and sources. So is that another reason why international is doing well?
10:49Unidentified speaker (speaker_1): Well, that's all pretty much covered from coming time. Taiwan and Korea, right? So we're seeing uh and it might be even closer to 80% when you look at what's going into the data centers. Um and you can see that in the statistics. So uh one of the interesting things that's happened recently is you've seen um you've seen the trade deficit widen significantly, and it's all because of the large increase that we're putting into chips, whether it's memory components, which are coming primarily from Taiwan and Korea. So that's Micron, Samsung, and SK Heinex. They're making most of them there. Mike now, Micron does have plants here in the US, but they're on the smaller side. And then also the servers and the GPUs, which are the other expensive component here, they are coming primarily from Taiwan. Most of that is made at uh Taiwan semiconductors. semiconductor. So that's it's actually uh dampened the effect on GPU. GP on excuse me and GDP uh for the data center build out, which has been the strongest part of our economy because it is being partially offset by um net exports because we're importing
11:49most of the equipment.
11:51Unidentified speaker (speaker_3): I know we have another uh podcast we're doing on uh or uh uh data centers uh uh uh you satellite know, uh orbital data data centers. Uh uh briefly d uh Philip, what are your thoughts there?
12:02Unidentified speaker (speaker_1): That's that's really tough. Now the the interesting part about that is people want to do it because A, uh there's you can cool it instantly and there's endless power in space. So you throw out a solar array, uh theoretically you can have a solar ray big enough to power the data center and then you don't have to worry about cooling because you're in the vacuum of space. Now the flip side of this is there are a whole uh bunch of characteristics that make it very difficult to have uh um GPUs and computer chips in space. A is the lift-off can be the shaking and vibrations for the liftoff can be difficult for the chips. And then B is you do get gamma radiation that shoots off of the sun in and waves, and those have in the past knocked out satellites and knocked out other equipment, which could also affect the data center. So you can could ultimately put that up there and then have a giant solar flare, if you will. and perhaps do some serious damage to the data center.
12:56Unidentified speaker (speaker_3): I mean that's it's uh it's a junkyard up there. You know, and uh and uh you can't just call your the your local repairman and say, Hey, would you come in and fix my uh my uh uh well where is it? Oh it's it's it's uh it's orbiting the earth. So so it's at that that uh I w I wonder if that uh how how they factored that into the the uh the the life. expectancy because I know one of the things I'm concerned they they've linked into the depreciation schedule on a lot of these data centers to the point of absurdity. Well you know for uh um I I wonder if they factored all the potential risk. And uh
13:28Unidentified speaker (speaker_2): let me piggyback onto that because they lengthened uh the life cycle of that and now uh all the big private credit, private equity firms have come together to create vehicles to provide funding. And I know
13:49Unidentified speaker (speaker_3): Tyler's a big fan of that.
13:50Unidentified speaker (speaker_2): Well it it and it y you meant Yeah, yeah. I of course he is. Tyler loves to talk about the circular lending and I am curious because we're hearing more and more about this, what are your thoughts on Core Weave?
14:08Unidentified speaker (UNKNOWN): Um the
14:11Unidentified speaker (speaker_1): Core Weave is an interesting company where they have
14:17Unidentified speaker (UNKNOWN): they
14:17Unidentified speaker (speaker_1): have s grown so fast and they've done it primarily using debt. And they are they have committed to build so many more data centers that they're going gonna to need to find a lot more debt. And the problem is is the market is now sniffing out that they're not making enough money to pay for the data centers. So for example, right now they have um roughly nine billion in cash. I think they have somewhere between twenty and thirty billion of debt. And over the next 12 months, um they're going to have a negative free cash flow, which is estimated by Wall Street to be around you know, 27 billion. So like they're going to have to find all of that debt. They're short on capital. Um, which in my mind, I mean that's that's a classic insolvency, right? So you have a company where you have more debt than you have cash. You are running a negative free cash flow. You need people to give you, you know. 10, 20 billion dollars to continue funding your
15:17operations. If they don't give it to you, you're bankrupt. So that's a very real concern. And I think in the last few days, we've seen the market start to realize that same concern. And you've seen now two IPOs of data center companies that are similar be to pulled because they're having trouble getting funding. Um, you know, the the the other problem is uh private credit has been a big funder of this recently uh through the insurance companies. And now private credit is having issues raising new funds themselves because a lot of their products have been gated recently. As there was concerns over the the irony, is of course most of that was software stocks, which started having difficulties, which led to some of uh the defaults within the credit cr in the credit funds and some lower returns than expected. Which so the irony is the AI. caused the gating in the private credit, which now the private credit can't raise the money to be the incremental dollar for the data centers like CoreWave.
16:10Unidentified speaker (speaker_2): So uh your valuation of Core Weave? It's zero. I'm sorry, say that again. I thought I heard you said say zero. That's a donut. Yes, sorry. It's a donut. Fantastic. I like donuts. Yeah. I
16:23Unidentified speaker (UNKNOWN): like
16:23Unidentified speaker (speaker_1): donuts too.
16:24Unidentified speaker (speaker_2): But but not to invest in. Yeah,
16:26Unidentified speaker (speaker_1): not to invest. in. It's uh which is interesting because the stock uh last looked yesterday was in the 60s, but it was uh it's it's been falling rapidly the last two two trading days.
16:36Unidentified speaker (speaker_2): So you run our natural resource portfolio, and one of the
16:44Unidentified speaker (UNKNOWN): w
16:44Unidentified speaker (speaker_2): part of your thesis for natural gas is AI data centers. and I assume for nuclear and some of the other things. Pull out AI data centers and they don't exist, they're whatever the case may be. Is there still enough underlying fundament are there There enough underlying fundamental drivers to still hold the thesis? that natural gas, that nuclear is still a great investment?
17:24Unidentified speaker (UNKNOWN): Um
17:25Unidentified speaker (speaker_1): on nuclear pr on nuclear probably not, um because frankly they're extremely expensive to build. Um and the reason that they're pushing for them is that in with AI you need a constant supply of energy. So you can't have intermittent energy, you can't really use solar, you can't use wind. Or if you do, you have to use huge batteries with them to make that work. And the companies that are doing a lot of the AI are um you know California based companies. It's Google, it's Meta. Um
17:53Unidentified speaker (speaker_2): well they have easy regulation, so that shouldn't be a problem
17:56Unidentified speaker (UNKNOWN): with California. Problem but they're
17:57Unidentified speaker (speaker_1): You know, they're concerned about the environment, rightfully so. And so they want wanna, to you know put their good house. housekeeping label on these things. And so they would prefer to have uh sources of power that are not um harmful for the environment. So that takes
18:11Unidentified speaker (UNKNOWN): when
18:12Unidentified speaker (speaker_1): you look at the power sources that you can have to really power a data center, they have to be consistent, high load. So you're looking at coal, natural gas, and nuclear. And so across the board, nuclear is obviously the cleanest. So that's why people are pushing uh for nuclear. Um there's a the topic is you know small modular reactors, which is probably where we're we're going next. And these are small nuclear power plants that are effectively about one-tenth the size of a traditional nuclear power plant, uh, and they fit on the size of a uh a shipping container. And we know how to build these. They're they're in all of our submarines, they're in all of our aircraft Craft carriers. We can build them, we know how to, uh, but there's just been a delay of of trying to use them for civilians. civilian resource for s for civilian civilian needs in particular AI data centers. So I think from nuclear, that is a bigger risk. Although, you know, from a uranium perspective, China is still building nukes left and right. So there's gonna be a higher demand regardless for coming from China for uranium. But for the US it
19:08Unidentified speaker (UNKNOWN): might be
19:09Unidentified speaker (speaker_3): building on doing a lot of coal too, aren't they?
19:11Unidentified speaker (speaker_1): A lot they're doing both. Coal, uh coal and uh and um uh uranium for nuclear um the other side of the point to natural gas would be what it really does for natural gas is it delays the tightening of the market. So you're gonna get there in two to three years through two cop two things happening. One is the number of LNG export plants will continue to come online and that's a constant demand for natural gas. So you're looking ultimately over the next three years, it could be somewhere around 15 to 20 B's of demand, which is would suck up the entire market where we are today. And similarly, on the supply side, we are not seeing the growth in US supply of natural gas that we had in the past for two reasons. One is the low prices simply slowed down drilling for natural gas wells by themselves. And uh the B part of that is the Permian, which has been the huge growth for the oil. those wells spit out thirty forty percent of the production of those wells is natural gas, and the permine has finally stopped. So it used to be called associated gas because as the permine went from three million barrels a day to nine million barrels a day, that also came with a huge increase in gas. And now that that's stopped, um, we're not seeing this huge growth from natural gas that
20:23we've seen since quite frankly two thousand nine.
20:33Unidentified speaker (speaker_3): Aaron
20:33Unidentified speaker (UNKNOWN): Powell
20:33Unidentified speaker (speaker_1): Demand is picking up and in fact demand's picking up for coal because of what we've seen in Iran. So
20:38Unidentified speaker (speaker_3): you're supposed to call it beautiful coal now.
20:40Unidentified speaker (UNKNOWN): It's clean
20:41Unidentified speaker (speaker_1): yeah, clean, beautiful
20:41Unidentified speaker (UNKNOWN): coal.
20:43Unidentified speaker (speaker_2): Big, beautiful coal. Big, beautiful
20:45Unidentified speaker (UNKNOWN): coal.
20:45Unidentified speaker (speaker_3): Stop calling it AI anymore.
20:49Unidentified speaker (speaker_2): Oh yeah, what it I super
20:50Unidentified speaker (speaker_3): super intelligence. Why didn't they just say advanced intelligence that we wouldn't have to say, we still say AI.
20:59Unidentified speaker (UNKNOWN): Yeah.
21:02Unidentified speaker (speaker_1): You have my vote, Mr. Moran. Or
21:05Unidentified speaker (speaker_0): average intelligence.
21:09Unidentified speaker (speaker_1): Uh yeah, you know, it's uh interesting. So with coal, if you look at what happened in with natural gas, um a lot of natural the gas comes out of the Middle East. So it has to go through multiple checkpoints, including the Strait of Hormose, um, it has to go through the Strait of Malacca. Coal, on the other hand, is primarily coming from Alaska, you know, and the United States. And so we're shipping it straight across the Pacific. There's no choke points. So if you're an Asian buyer of an Asian buyer of LNG and you need energy and you need energy and you're looking around and you're worried about not getting LNG because of these new choke points, which have shown they can be closed down, you're gonna start switching the coal. Um and that's what we're seeing. So we're seeing higher coal demand out of particularly the Asian economies that can't provide their own energy, whether it's Japan, Korea, um, and the Philippines.
21:52Unidentified speaker (UNKNOWN): All
21:52Unidentified speaker (speaker_2): right, gentlemen, we're getting close to the end of the show. Let's let's have a few questions, uh a little speed round here.
22:01Unidentified speaker (UNKNOWN): Um I've
22:02Unidentified speaker (speaker_2): got to put my glasses on, and I left my good glasses at home, of course. And I have probably 20 pairs of these in my office. So when Tom comes and forgets
22:14Unidentified speaker (UNKNOWN): his.
22:18Unidentified speaker (speaker_2): Yeah. So I I'm using one of the ones I got
22:21Unidentified speaker (speaker_1): you.
22:22Unidentified speaker (speaker_2): Um what's one of the most
22:26Unidentified speaker (UNKNOWN): miserable?
22:26Unidentified speaker (speaker_2): misunderstood bottlenecks of this whole AI build-out? Because we talk a lot about the bottlenecks, but what's one that people misunderstand?
22:40Unidentified speaker (speaker_1): Um, the money it takes to build them. And I think that's what we're seeing right now. So we're finally seeing the endless um ability to finance the data centers. is finally being put to question. And that really started with to be quit, I'll be super quick, but it started with a lot of the data center builders had
22:58Unidentified speaker (UNKNOWN): uh
22:58Unidentified speaker (speaker_1): jobs of cash because they were these giant companies like Google Meta and then Microsoft. And they've kind of burned through that. And then now you've seen we've talked about it earlier, but you're starting to see uh less financing available. And I think that'll be the bottleneck that surprises everyone with the data centers, is just paying pain for them.
23:14Unidentified speaker (UNKNOWN): So
23:14Unidentified speaker (speaker_3): you don't think it's the electric grid? I would have thought the electric grid that they you know, the amount of power and energy it takes to fuel these things. I mean they were just we're uh we're already at a uh a d uh a archaic a grid system and you throw uh the all this demand on top of it uh puts a lot of pressure on and that's why we're getting some political pushback with it's uh it's affecting uh utility costs and uh now we're gonna having have people, you know, uh not not, you know uh have moratoriums against data centers. So
23:42Unidentified speaker (speaker_2): Yeah. Um computing is becoming more and more efficient, right? So the more efficient it becomes, does that translate into less energy consumption?
23:59Unidentified speaker (UNKNOWN): It
23:59Unidentified speaker (speaker_1): does, and uh it it translates into less demand for data centers. So that's the irony of the whole thing is um what we're seeing now with some of the open source models, and we do it here ourselves, is we can have the GPU here on site. We can load up the AI um a model onto it, and we can run it without having to query a data center somewhere. And it's highly efficient. And so
24:23Unidentified speaker (UNKNOWN): I
24:23Unidentified speaker (speaker_1): imagine down the line, and I know Apple's working on this and other comp uh and so is Samsung is can you put a stronger chip in your phone? So you're doing the compute on your phone, you're doing the compute on your computer, you're bringing it closer to what we call the edge. And as that happens, we might not need as many data centers as we're forecasting. That could be the um the saving grace to the the whole story, quite frankly.
24:45Unidentified speaker (speaker_2): But that also could hurt a lot of people. There will hurt. A lot of investors, a lot of companies. that are allocating heavily into the massive data center build out.
24:58Unidentified speaker (speaker_3): And the life expectancy they've depreciated over if their useful life is now much shorter. I guess you use it as a farmhouse. I mean uh
25:11Unidentified speaker (speaker_2): A grow house.
25:13Unidentified speaker (UNKNOWN): So I mean it's
25:13Unidentified speaker (speaker_3): uh I mean that that that may that may uh may not end well.
25:18Unidentified speaker (speaker_2): Uh yeah, it could be ugly. Um last question. Who's the darling that has the potential to be the dog?
25:33Unidentified speaker (UNKNOWN): In
25:33Unidentified speaker (speaker_1): in my mind it's it's probably SpaceX and Elon Musk. I think um Wow. They they have a $2.2 trillion market cap the and they don't have the
25:47Unidentified speaker (UNKNOWN): the
25:48Unidentified speaker (speaker_1): other businesses that are supporting it, right? So when you look across the other hyperscalers, other mag seven names or the and the big companies that are valued that high, it's you know, like I said, it's Google, Meta, Microsoft, Amazon. They have great other businesses that are highly profitable. SpaceX doesn't have that, right? So Twitter is does not make a bunch of money. It's a small company. Um, they're really relying on building these data centers and having an LLM through Glo Groc that's they're gonna get paid on. And I think they're the most at risk when you when you look around. Um with the and they know they have the rocket business, but you know, we've done the math, and even if you're firing two rockets a day, we can't get anywhere close to the valuation you need. Um so I think when you look at the bells of the ball, the really popular stocks that have uh people were clamoring for and that were excited about the IPO, huge IPO, right? Um in my mind, that's the one that probably has the most to uh lose if this doesn't play out well.
26:39Unidentified speaker (speaker_3): I heard today that the the the the merger between uh uh uh Tesla and SpaceX is in is is uh sooner rather than later.
26:46Unidentified speaker (UNKNOWN): Mm-hmm.
26:47Unidentified speaker (speaker_1): And it will be Tesla by buying and SpaceX because Elon has control of SpaceX from a voting perspective, so he can he could approve that. And there's also uh there's an incentive compensation plan that they put together at Tesla about the two years ago, with you remember it was in the press because there's a lot of fights about the Delaware court said they couldn't do it, they moved to Texas. Um, but the way that's structured is it's uh based on his market cap. So presumably if if SpaceX keeps its two $2 trillion dollar market cap and it merges into this Tesla and Tesla's got a trillion dollar market cap. You now have a $3 trillion market cap and you unlock all his incentive comp. Um so he would get another trillion of stock or something like that. I don't know the exact number, but it's a big number. So he it benefits Elon's uh significantly if he can have Tesla buy SpaceX.
27:28Unidentified speaker (speaker_2): You may be speaking out of both sides of your mouth. You just said SpaceX is the one most at risk. Therefore, I assume you wouldn't allocate to it. Correct. However, if Tesla may take out SpaceX, I assume there's a premium and potentially an arbitrage one could take advantage of.
27:49Unidentified speaker (speaker_1): You'd be getting paid in Tesla stock to make that transaction happen, which is
27:54Unidentified speaker (UNKNOWN): so
27:54Unidentified speaker (speaker_1): you have the effectively the valuation of both companies becomes one, and then you have a lower valuation. So you're looking at a much bigger downside because now you've taken another company. which doesn't make a bunch of money that's trading for a trillion dollars. So you've compounded your problem from a investment perspective, unless you sell at the deal.
28:11Unidentified speaker (speaker_3): If the cost of days uh being domestic data centers on Earth uh gets get keeps going up and getting prohibitively high, uh doesn't Musk have an advantage in building them or uh orbital data centers?
28:23Unidentified speaker (speaker_1): Theoretically he would.
28:24Unidentified speaker (UNKNOWN): Yeah,
28:25Unidentified speaker (speaker_1): but It just it's just this it's it's time. and probability, right? So you ha you have to forecast for that. So when we when we look at SpaceX and then we try to value it, we put some really radical assumptions on that. We assume that he does that. We assume that the rockets are going off constantly. We assume that he can make the big his bigger rocket work, which dramatically lowers lift costs. So we do all those things in our assumptions and we still can't get close to some of the numbers that
28:47Unidentified speaker (speaker_3): so if I buy a roadstra road strip
28:48Unidentified speaker (UNKNOWN): I
28:48Unidentified speaker (speaker_3): might not be um the company may be a bankrupt, I may not get it repaired
28:52Unidentified speaker (speaker_1): I'm not gonna say bankrupt but it's uh you're it it's gonna
28:55Unidentified speaker (UNKNOWN): be difficulty it's not a zero it's not a zero
28:58Unidentified speaker (speaker_3): it's a donut with without a hole
29:00Unidentified speaker (UNKNOWN): with
29:00Unidentified speaker (speaker_1): some
29:00Unidentified speaker (speaker_2): sprinkles
29:01Unidentified speaker (UNKNOWN): they don't have the debt that the other
29:03Unidentified speaker (speaker_1): they don't they don't have the debt that core weeb does so core we're just as they need debt today.
29:07Unidentified speaker (speaker_3): Right.
29:07Unidentified speaker (speaker_1): SpaceX they can stop and they don't need that debt. They can they have other businesses they can run. They're not you know
29:13Unidentified speaker (speaker_2): so core weebs Weaves a donut, SpaceX is an apple fritter.
29:18Unidentified speaker (UNKNOWN): I
29:18Unidentified speaker (speaker_1): would say so.
29:19Unidentified speaker (speaker_2): Fantastic. Gentlemen, thank you. Uh thank you for joining us for another edition of Quarter Over Quarter. If you enjoyed today's podcast, please like, subscribe. That helps us tremendously. Share with your friends. We look forward to seeing you on a future edition of Court. Quarter Over Quarter. That's all for this episode of Quarter Over Quarter. If you like the show, then consider giving us a like and subscribe to show your support and ensure you don't miss any new episodes. To learn more about what we do at Moran Wealth Management, visit our website at Moranwm.com. At Moran, our team of professionals are here to help deliver personalized wealth and asset management strategies for every step of life.
30:15Unidentified speaker (speaker_0): Moran Wealth Management LLC is an SEC registered investment advisor. Registration as an investment advisor does not imply a certain level of skill or training. The publication of Moran Wealth Management's videos should not be construed by any consumer. or prospective client as Moran Wealth Management's solicitation or attempt to affect transactions and securities or the rendering of personalized investment advice. A copy of Moran Wealth Management's current writ disclosure statement as set forth on Forum ADV discussing Moran Wealth Management's business operations, services, and fees is available from Moran Wealth Management upon written request or at advisorinfo.sec.gov. All information in the videos is provided solely for convenience. purposes only, and all users thereof should be guided accordingly.