Denise Elshar warns of AI infrastructure risks using railroad parallels and CoreWeave debt signals
Summary
Denise Elshar argues that the current AI infrastructure boom mirrors the 19th-century railroad era, where transformative technology coexisted with widespread corporate bankruptcy. She cites historical data showing that nearly one in four railroad companies failed after the 1873 panic, suggesting that industry success does not guarantee individual profitability. Elshar highlights CoreWeave as a warning sign, noting its bond yields jumped approximately 15% in October due to investor concerns over high debt and capital expenditure needs. She contrasts the rapid obsolescence of AI chips with the durability of railroads, warning that price competition could hurt leveraged providers. Finally, she advises investors to avoid issuer concentration risk by recognizing the overlap between broad index funds and tech-specific ETFs.
Key takeaways
- 1 In October, yields on some of the bonds jumped around 15%, even a little bit higher.
Denise Elshar · 11:37
Episode sections
- 0:04 – 2:17
Introduction and AI Market Correction
Denise Elshar introduces the podcast's philosophy on systematic wealth building and notes that AI stocks have recently entered a technical correction, prompting a discussion on the sector's risks and historical parallels.
- 2:17 – 5:15
Historical Parallel: Railroad Boom vs. AI
The speaker compares the current AI infrastructure build-out to the 19th-century railroad boom, noting that while the technology is transformative, many railroad companies went bankrupt despite the industry's success.
- 5:15 – 7:06
Wealth Creation in Ancillary Industries
Elshar highlights Andrew Carnegie (steel) and John D. Rockefeller (oil) as examples of investors who profited from the railroad era by supplying essential inputs rather than owning the railroads themselves.
- 7:06 – 10:19
AI Infrastructure and Differentiation
The discussion shifts to AI data centers and semiconductors, emphasizing that market enthusiasm is shifting toward discrimination, requiring investors to analyze competitive advantages, balance sheets, and profitability.
- 10:19 – 12:16
CoreWeave Case Study and Bond Yields
CoreWeave is presented as a 'canary in the coal mine' for AI infrastructure; its bond yields jumped ~15% in October, signaling investor concern over the company's high debt load and capital expenditure needs.
- 12:16 – 15:22
Technological Obsolescence and Price Competition
Unlike railroads, AI chips face rapid obsolescence. The speaker warns that if supply outpaces demand, prices could fall, hurting highly leveraged infrastructure providers even if AI adoption grows.
- 15:22 – 19:25
Portfolio Concentration and Issuer Risk
Elshar warns investors about issuer concentration risk, where holding stocks and bonds of the same company creates overlapping exposure to the same financial health risks.
Assets discussed
| Asset | Kind | Stance | Context |
|---|---|---|---|
| NVIDIANVDA | equity | neutral | And instead of locomotives, we're building NVIDIA chips. And instead of coal and steam, we need electricity and lots of it. |
| MicrosoftMSFT | equity | neutral | Let's say you own the S P 500 Index Fund, and then maybe a tech ETF, and maybe a few individual stocks like a Microsoft or an Amazon or an Apple. |
| AmazonAMZN | equity | neutral | Let's say you own the S P 500 Index Fund, and then maybe a tech ETF, and maybe a few individual stocks like a Microsoft or an Amazon or an Apple. |
| AppleAAPL | equity | neutral | Let's say you own the S P 500 Index Fund, and then maybe a tech ETF, and maybe a few individual stocks like a Microsoft or an Amazon or an Apple. |
“Stance” is how the speakers talked about the asset, not a recommendation.
Claims worth verifying
AI stocks have entered a technical correction.
Denise Elshar ·1:57
How to verify: Check S&P 500 Information Technology Sector Index (XLK) or Nasdaq Composite (IXIC) price action for a decline of 10% or more from recent highs during the period 2026-08-13 to 2026-11-13. For recent events, verify 2026-08-13 through 2026-11-13. Use another period only when the speaker explicitly names it.
Almost one in four railroad companies went bankrupt following the financial panic of 1873.
Denise Elshar ·4:33
How to verify: Review historical financial records from the Panic of 1873, such as those archived by the Federal Reserve or historical economic texts, to count railroad bankruptcies versus total railroad companies. For recent events, verify 2026-08-13 through 2026-11-13. Use another period only when the speaker explicitly names it.
CoreWeave reported more than $100 billion in contracted future revenue backlogs earlier this year.
Denise Elshar ·10:39
How to verify: Search CoreWeave’s official investor relations page or press releases for announcements regarding contracted revenue or backlog figures from 2026. For recent events, verify 2026-08-13 through 2026-11-13. Use another period only when the speaker explicitly names it.
Yields on CoreWeave bonds jumped around 15% in October.
Denise Elshar ·11:29
How to verify: Check bond market data providers (e.g., Bloomberg, ICE Benchmark Administration) for CoreWeave’s bond yield movements in October 2026. For recent events, verify 2026-08-13 through 2026-11-13. Use another period only when the speaker explicitly names it.
Andrew Carnegie worked for the Pennsylvania Railroad early in his career.
Denise Elshar ·6:15
How to verify: Consult biographies of Andrew Carnegie or historical records of the Pennsylvania Railroad’s employee lists from the mid-19th century. For recent events, verify 2026-08-13 through 2026-11-13. Use another period only when the speaker explicitly names it.
CoreWeave provides specialized cloud computing power for companies developing and running artificial intelligence.
Denise Elshar ·10:39
How to verify: Review CoreWeave’s website or business description on their investor relations page to confirm their service offerings. For recent events, verify 2026-08-13 through 2026-11-13. Use another period only when the speaker explicitly names it.
AI chips become technologically outdated relatively quickly, similar to consumer electronics.
Denise Elshar ·12:16
How to verify: Review industry reports on semiconductor lifecycle management or technology refresh cycles from firms like Gartner or IDC. For recent events, verify 2026-08-13 through 2026-11-13. Use another period only when the speaker explicitly names it.
Issuer concentration risk occurs when an investor holds both stock and bonds of the same company.
Denise Elshar ·18:01
How to verify: Review financial education materials from regulatory bodies like the SEC or FINRA regarding portfolio diversification and concentration risk. For recent events, verify 2026-08-13 through 2026-11-13. Use another period only when the speaker explicitly names it.
Primary sourceDo your own research
- Verify CoreWeave's October bond yield movements using Bloomberg or Refinitiv fixed-income data.
- Check CoreWeave's public filings or press releases for the $100 billion revenue backlog figure.
- Review historical bankruptcy rates for US railroad companies post-1873 using the Interstate Commerce Commission historical reports or academic economic studies.
- Analyze the overlap between major tech ETFs (e.g., QQQ, XLK) and broad market indices (e.g., SPY) using fund prospectuses or portfolio holdings data.
- Assess the depreciation schedules and R&D spending of major AI chip manufacturers to evaluate the speed of technological obsolescence.
- Compare the capital expenditure requirements of AI data centers against the revenue visibility of providers like CoreWeave using SEC filings.
- Research the historical profitability of ancillary industries (steel, oil) versus primary infrastructure owners during the Gilded Age.
- Check current electricity and cooling infrastructure capacity constraints in major data center hubs like Northern Virginia or Silicon Valley.
Full transcriptexpand
- 0:04
- Unidentified speaker (UNKNOWN): There's
- 0:05
- Denise Elshar: a lot of noise when it comes to money and investing. Headlines, market predictions, stock tips, and social media experts telling you what to do. But real wealth is almost never built that way. Real wealth is built quietly, systematically, and over time.
- 0:21
- Unidentified speaker (UNKNOWN): And
- 0:21
- Denise Elshar: the truth is, most people were never taught how investing actually works. Most people think it investing is about picking stocks. It's not. It's about building a framework that allows wealth to compound over decades. And that's what this podcast is about. I'm Denise Elshar, and this is Private Wealth the Edit. I
- 0:41
- Denise Elshar: spent my career managing money and investments for institutions and high net worth families. And over time, one when things thing become very clear to me. The way sophisticated investors think about money. is very different from what most people are taught, or not taught, but do. This podcast is about closing that gap. Each episode, I'll focus on what actually
- 1:01
- Denise Elshar: matters: cutting through complexity to bring clarity to how wealth is built. I'm here to give you an institutional edge, one private conversation at a time. Welcome back to the edit where we edit out the noise and make sense of investing one. private conversation at a time.
- 1:17
- Unidentified speaker (UNKNOWN): I
- 1:17
- Denise Elshar: hope you enjoyed our last episode. It was the Halloween edition where I talked about something that spooks even the seasoned investors, and that's a stock market crash or correction. Those scary moments when stock prices fall and everyone panics, and the temptation to do something, anything to stop the pain can be overwhelming. But in most cases, the best thing to
- 1:37
- Denise Elshar: do is don't touch that dial and let your portfolio and diversification work their magic. And before we get back to our building blocks and building portfolios, I thought it would be good if we spent a little bit of time on AI. And to be honest, particularly because markets have wobbled a little lately on AI investments, and as of this taping entered a technical
- 1:57
- Denise Elshar: correction, and it feels like the elephant in the room. So we can't talk about market corrections without addressing this elephant. and what some consider a really big risk. So I thought before we get back to our regularly scheduled program, let's address Mr. Elephant and let's talk about AI and what's happening in financial markets. To do that, I want to take us back
- 2:17
- Denise Elshar: actually 150 years ago to the railroads and see what lessons that we we've learned from that past period. The last time we saw such a transformative technology. Because right now we're living through what may be one of the biggest technological revolutions of our lifetime, really, and that's artificial intelligence. And it is everywhere. It's in the news,
- 2:37
- Denise Elshar: it's in our phone, it's at our workplace, and it is certainly in our investment portfolios. And for a while it seems like a company, all they had to do was mention mentioned the word artificial intelligence, and investors couldn't buy the stock fast enough. But something interesting. is happening. Wall Street's starting to get a little bit more selective. And to understand why, we're going to
- 2:56
- Denise Elshar: travel back 150 years to a time when America was obsessed with another revolutionary technology, the railroad. So all aboard, because today's edit is about the AI gold rush, the railroad barons, and why the people who make the most money from a revolutionary technology aren't always the people who make
- 3:14
- Unidentified speaker (UNKNOWN): the most.
- 3:15
- Denise Elshar: that built
- 3:16
- Unidentified speaker (UNKNOWN): it. So
- 3:16
- Denise Elshar: let's go back to the 1800s. America was in the middle of an extraordinary transformation. Railroads were connecting the country in ways nobody ever imagined. Before railroads, moving goods across the country was really expensive and it was slow and actually sometimes just impossible. But suddenly with railroads, farmers were able to sell their products hundreds of miles
- 3:36
- Denise Elshar: away. Factories could reach new customers. and cities expanded. It was really, it was a revolution. And investors knew it, and money poured into railroad companies. Everybody wanted a piece of the action. Railroad tracks were being laid everywhere, and new companies were being formed and new routes were being announced. It was an enormous financial expansion and projects were going everywhere.
- 3:56
- Denise Elshar: Does this sound familiar? We always say history doesn't repeat itself, but it does rhyme. But today, instead of building railroad tracks, we're building data centers. And instead of locomotives, we're building NVIDIA chips. And instead of coal and steam, we need electricity and lots of it.
- 4:13
- Unidentified speaker (UNKNOWN): And
- 4:14
- Denise Elshar: just like the railroad boom, everybody wants to own a piece of the AI action. But here's the part of the railroad story I think is fascinating. The railroads absolutely transformed America. That is not debatable. But a staggering number of railroad companies went bankrupt. Almost one in four went bankrupt following the financial
- 4:33
- Denise Elshar: panic of eighteen seventy-three. So how was that possible when a technology was so successfully? Here's what happened.
- 4:45
- Denise Elshar: Works. They laid tracks, they built stations, they bought trains, and they raced to connect as many cities as possible.
- 4:53
- Unidentified speaker (UNKNOWN): But
- 4:53
- Denise Elshar: as more companies entered the business, they started building competing routes, and too much capacity meant the prices crashed. And remember, these companies had borrowed large amounts of money. So those interest payments didn't disappear just because revenues disappeared. So eventually the economics stopped working.
- 5:11
- Unidentified speaker (UNKNOWN): The
- 5:11
- Denise Elshar: railroads were incredibly successful at moving America forward.
- 5:15
- Denise Elshar: but a lot of their shareholders and bondholders lost money along the way. And that brings me to the story of two of the wealthiest Americans in history, Andrew Carnegie and John D. Rockefeller. Those were the people who really made money during this era. And here's where the story gets interesting. When we think of America's great industrial
- 5:35
- Denise Elshar: fortunes, those two names always come up. but neither built their fortunes owning railroad companies. They understood something else. They didn't necessarily need to own the railroad to make money from the railroad. You need to understand what the railroad needed and what opportunity it created. Andrew Carnegie
- 5:55
- Denise Elshar: recognized that railroads needed steel and lots of steel. And Carnegie built an enormous steel business supplying that demand. Now, Carnegie did work for the Pennsylvania Railroad early in his career, and he also invested along some railroad related ventures like sleeping cars and bridges. But his great fortune ultimately came
- 6:15
- Denise Elshar: from the steel. Then there's Rockefeller. Rockefeller built standard oil, and as the railroads expanded, Was they helped create an enormous market for transporting oil and refined petroleum products across
- 6:26
- Unidentified speaker (UNKNOWN): the world.
- 6:27
- Denise Elshar: country. So Rockefeller used the railroad networks to move his products, negotiated favorable rates, and built one of the most powerful businesses in America. Both men definitely benefited enormously from the railroad expansion, but neither needed to bet his fortune on a particular railroad company to succeed.
- 6:47
- Denise Elshar: Now, there were plenty of wealthy railroad owners too, so I'm not going to suggest. that nobody made money owning railroads. The point is that some of the biggest fortunes were created in the industries around the railroads, around the technology. And I think that's an important lesson here because sometimes the best way to profit from a
- 7:06
- Denise Elshar: revolution isn't to own the revolution itself, it's to own the businesses that benefit from it. So that brings us back to AI. What does the AI railroad boom look like? Let's look at what's happening today. Technology companies are spending extraordinary amounts of money building out the infrastructure for artificial
- 7:26
- Denise Elshar: intelligence. We're talking about data centers, semiconductors, electricity, cooling system. It's an enormous investment cycle. And the comparison to the railroads is pretty compelling. The idea that the data center is the railroad track and that the chips might be the locomotives and the electricity is the coal burned fuel.
- 7:45
- Denise Elshar: But here's the question I want you to think about. Does the fact that the world needs all this infrastructure mean every company building it will be a good investment? Of course not. This is something I think investors are beginning to realize, and we're starting to see differentiation in the markets. For the first part of the AI boom, the market was rewarding
- 8:05
- Denise Elshar: almost anything associated with artificial intelligence. It seemed like attaching AI to your growth story was enough to attract investors. But this is starting to change, and I think that's a very important development. It's actually a really good development, to be honest, because the market's moving from enthusiasm to discrimination,
- 8:25
- Denise Elshar: from asking, is this company involved in AI to asking how exactly is this company going to make money from AI? And more importantly, how much money will it have to spend to make that money? And those are very different questions and all very important.
- 8:42
- Denise Elshar: So not all AI companies are created equal. And that brings us to differentiation. Differentiation is a fancy well. Wall Street word for something pretty simple. Not every company is the same, even if they're operating in the same industry. Some businesses have incredible competitive advantages. Maybe they
- 9:02
- Denise Elshar: have loyal customers or strong are strong balance sheets. Others are competing by offering essentially the same product as everybody else. And some are probably borrowing huge amounts of money just to stay in the game. When an industry is growing quickly, investors sometimes overlook those differences, but eventually it catches up to them
- 9:22
- Denise Elshar: and everything starts to matter and it matters a lot. For an investor, I think there are four important things to consider. First, does the company have something that makes it special? It's Is there a barrier to entry? Is there a competitive advantage? Is there something special about it? Second, what does its balance sheet look like? Is it
- 9:42
- Denise Elshar: financing its growth from profits? Does it have money? Is it borrowing money? What do the future bills look like to pay back that money? Third, is the company actually making money? And I don't just mean growing revenue, because growing revenue growth and profitability are not the same thing.
- 9:58
- Unidentified speaker (UNKNOWN): And
- 9:59
- Denise Elshar: then finally, how much are investors paying for all this? Because even a wonderful company can be a terrible investment if you pay too much for it in stock. Now, let me give you a very timely example of that. It's a company called Core Weave. And Core Weave has become the canary in the coal mine of the data centers. And maybe you've never heard
- 10:19
- Denise Elshar: of Core Weave, but if you've been following AI stocks, you've probably seen its name recently. So Core Weave provides specialized clouds. cloud computing power for companies developing and running artificial intelligence. In simple terms, instead of every company buying its own very expensive AI computers, Core Weave builds and operates
- 10:39
- Denise Elshar: that and rents that computing power to customers. It's a smart idea. It sounds like a pretty good business, right? And the growth really has been extraordinary. The company reported more than $100 a hundred billion in contracted future revenue. backlogs earlier this year. So that's a lot of business. There's always a butt though, but all that building computing power is
- 10:58
- Denise Elshar: extremely expensive. You need the latest chips. The data center needs to be housed. You need electricity, you need cooling, and all the other infrastructure that goes along with it.
- 11:09
- Unidentified speaker (UNKNOWN): And
- 11:09
- Denise Elshar: Cora Weave has borrowed enormous amounts of money to fund its expansion, which brings us to what happened in the bond market recently. Now, you know, I spent a large part of my career in fixed income. So when I see something happening in the bond market, I pay attention. And Cora Weave's bonds have been sending a pretty interesting signal.
- 11:29
- Denise Elshar: In October, yields on some of the bonds jumped around 15%, even a little bit higher. Let me explain why that happened.
- 11:37
- Unidentified speaker (UNKNOWN): When
- 11:37
- Denise Elshar: you buy a company's bond, you're basically lending money to that company. And the yield is the return investors are demanding for the risk of lending money to that company. So when investors started demanding a higher yield of 15% or more, that's a sign they're becoming very concerned about the risk.
- 11:57
- Denise Elshar: Now, that doesn't mean Core Weave is going bankrupt, and it doesn't mean the business isn't growing, but the market's asking a much tougher question. Will the company generate enough cash from all these investments to justify the enormous amount of money it's borrowed? And what happens if interest rates stay high? And what
- 12:16
- Denise Elshar: happens if the company needs even more money to expand? So, unlike railroads, which could last for decades, AI chips are becoming technologically outdated. relatively quickly. Sound familiar? You buy a computer, you buy an iPhone, a year or two later, there's a new version. So you may have to keep spending money,
- 12:37
- Denise Elshar: enormous amounts of money, just to remain competitive. And that makes the financial equation even more complicated for a company like Coreweave. Investors aren't necessarily rejecting AI, they're questioning the financial structure of some. of the businesses trying to capitalize on it. And that's not a bad thing. That's actually a really healthy
- 12:56
- Denise Elshar: development, but there will be winners and there will be losers. And that is exactly the kind of distinction we saw during the railroad boom. The demand for transportation was real, but that doesn't mean every railroad could earn enough money to pay its debt. Let's talk about shovels. Now there's another chapter. to the story, and it's one I
- 13:16
- Denise Elshar: find particularly interesting. We've all heard the old expression about the California gold rush. The people got rich weren't necessarily the ones digging for gold. They were the ones selling picks and shovels. And there's a lot of truth to that idea. But here's my question What happens when everybody starts Start selling shovels? Because that's when things get
- 13:36
- Denise Elshar: interesting. Competition increases, supply expenses. expands, and eventually prices have to start coming down.
- 13:43
- Unidentified speaker (UNKNOWN): In
- 13:43
- Denise Elshar: the AI world, we could see this happening with computing power and cloud services and even the AI models themselves. Right now, companies are racing to build more capacity because demand is enormous. But what if all that construction eventually produces more computing power than customers need at today's prices? I'm not saying this is. a
- 14:03
- Denise Elshar: today problem, but if we're thinking a little bit in the future, this could evolve. What happens when comp competing providers start lowering prices to win business? Or when newer, faster, more efficient chips make yesterday's expensive technology look less valuable? Sounds like a cell phone, no? If you're the
- 14:23
- Denise Elshar: customer, that's fantastic. Your AI costs come down and you can do more with less. And maybe AI becomes more affordable and businesses could even consider using it for more things. But if you're the company that borrowed billions to build that infrastructure, falling prices could become a serious problem. And that's what brings us to a really
- 14:42
- Denise Elshar: interesting possibility. AI could have an enormous economic success while some AI infrastructure investors expect experience disappointing return. I think the lesson here is the difference between a revolutionary technology and a good investment, because they're not necessarily the same thing. And that's something institutional
- 15:02
- Denise Elshar: investors spend a great deal of time thinking about. It's not enough to identify an exciting industry. You need to understand the business operating within that industry, how they make money, how much capital they need. and what makes them different from their competitors, and whether the price you're paying for that investment makes sense. Now, if you own
- 15:22
- Denise Elshar: a broad index fund, you probably already have exposure to many of the AI companies, and that's not necessarily a bad thing, but it's worth understanding how much of your portfolio is concentrated in the same technology theme. Because owning lots of different stocks doesn't mean you're taking lots. of different risks. And for those of you who own AI stocks
- 15:42
- Denise Elshar: outright, here's a question I'd like you to ask yourself. Am I buying this company because everyone says AI is the future? Or because I understand why this particular company can turn the future into sustainable profits? That is the big question. And remember, we don't need to predict exactly which company will win.
- 16:03
- Denise Elshar: the AI race. That's the beautiful thing about building a diversified portfolio. We can participate in innovation without betting our financial futures on just one company, one technology, or one story. So, in closing out, let's bring this back to the railroads. 150 years ago, investors were convinced that railroads would change the world,
- 16:23
- Denise Elshar: and they were absolutely right. But not every AI company is going to
- 16:26
- Unidentified speaker (UNKNOWN): be a good thing.
- 16:27
- Denise Elshar: winner, just like not every railroad company made money. And here's your edit for today: the first phase of a revolution rewards the story. The next phase rewards the economics. And if there's one thing I want you to remember, it's this the technology can win when the investor loses, because in the end, it's not about getting on the
- 16:47
- Denise Elshar: train, it's about making sure you bought a ticket on the right one. and that you haven't paid too much for that ride. Before we wrap up, let's turn to this week's AI edit, where we take what we've talked about today and bring it to life in your own portfolio with an AI prompt. Each week, I give you a simple AI prompt to help you apply the lessons from the edit today to your
- 17:07
- Denise Elshar: own financial life.
- 17:08
- Unidentified speaker (UNKNOWN): And
- 17:09
- Denise Elshar: given everything we've discussed today, I think this week's prompt is particularly useful. You may have a lot more experience. exposure to AI than you realize. Let's say you own the S P 500 Index Fund, and then maybe a tech ETF, and maybe a few individual stocks like a Microsoft or an Amazon or an Apple.
- 17:27
- Unidentified speaker (UNKNOWN): And
- 17:27
- Denise Elshar: you might think you're well diversified, but many of those same companies could appear across all three of those investments. And it's not just with stocks. As we've talked about, tech companies are spending an enormous amount. of money building AI infrastructure,
- 17:41
- Unidentified speaker (UNKNOWN): and
- 17:42
- Denise Elshar: many are turning to the bond market to help finance that. So you could own a company's stock in your equity portfolio and its bonds and your fixed income portfolio without even realizing it. This is something that we call issuer concentration risk when too much of your portfolio is exposed to the same company, even across different types of
- 18:01
- Denise Elshar: investments. And while stocks and bonds do behave differently, they are both exposed to the financial health of that company. So if something goes wrong, both could lose value. So this week's AI edit is designed to help you look through your stocks and bonds and identify overlapping holdings and understanding where your exposure
- 18:21
- Denise Elshar: to AI is and to individual companies where you may be more concentrated than you think.
- 18:27
- Denise Elshar: So the prompt is on my website at privatewealthedit.com and it's under the episode tab. Just pick it up, copy, and paste it into your favorite AI tool. But just remember to verify the results and discuss any financial decisions with your financial advisor. Do not ever trade off of anything that comes out of an
- 18:47
- Denise Elshar: AI bot. That's it for this week's edit. Thanks for joining.
- 18:51
- Unidentified speaker (UNKNOWN): Thank
- 18:51
- Denise Elshar: you for listening today. If you found this useful, follow the show or share it with someone who you think might find it interesting. Send it to your kids. If you have any questions that you'd like to ask or topics you'd like me to cover, please email me at privatewealthedit at gmail.com. And remember, I'm here to give you an institutional edge, one private
- 19:11
- Denise Elshar: conversation edge. This podcast is for informational purposes. only and does not constitute investment advice. All investing involves risk and you should consult your own advisor before making financial decisions.