Two ways big bets on data centers might affect you (Two Indicators)
Two ways big bets on data centers might affect you (Two Indicators)
11 hours ago•Planet Money•NPR
Podcast18 min 28 sec
Listen to Episode
Note: AI-generated summary based on third-party content. Not financial advice. Read more.
Quick Insights
  • Treat Oracle (ORCL) as a higher-risk AI infrastructure investment: its $169 billion debt load, near-junk credit rating, and reliance on OpenAI’s roughly $300 billion commitment make its outlook unusually sensitive to project delays and customer finances.
  • Before adding AI-exposed stocks or bond funds, check how much your portfolio depends on the data-center build-out; rising borrowing costs, construction delays, and weaker-than-expected demand could undermine returns.
  • The insights provide no specific buy recommendations, price targets, or timeframes for NVIDIA (NVDA), Alphabet (GOOGL), Amazon (AMZN), Meta (META), or Microsoft (MSFT).
Detailed Analysis

Oracle (ORCL)

  • Oracle has shifted from a profitable, relatively low-cost software business toward a capital-intensive data-center build-out, financed with substantially more borrowing.
  • Its total debt is reported at $169 billion, nearly double the level of two years earlier. S&P downgraded Oracle in July, leaving its rating one notch above junk status.
  • Oracle depends heavily on OpenAI’s commitment to pay about $300 billion over roughly five years for computing capacity. Because OpenAI’s financials are not public, its ability to meet those commitments is difficult for outsiders to assess.
  • A data-center project for OpenAI in New Mexico faced potential delays. Oracle sent creditors a force majeure notice that could allow it to delay payments, though the company publicly said the project remained on schedule.
  • Oracle is part of a circular financing relationship: OpenAI pays Oracle for cloud capacity, Oracle buys chips from NVIDIA, and NVIDIA invests in OpenAI. The podcast identifies this interconnectedness as a potential source of broader financial strain.
  • The transcript also says Larry Ellison guaranteed $40 billion to support his son David’s Paramount Skydance merger deal with Warner Bros. Discovery, adding to investor concerns.

Takeaways

  • Oracle’s data-center plans may benefit if demand for AI computing continues, but the company’s high debt and reliance on OpenAI make it a riskier way to gain exposure to that growth.
  • Consider how much a company’s investment case depends on a single customer or a chain of interdependent deals. The podcast notes that Oracle’s credit-default-swap rates were rising, signaling that credit markets saw increasing risk; it does not say Oracle was on the verge of bankruptcy.
  • A fall below investment-grade status could make borrowing and refinancing more difficult and could force some investors, such as certain pension funds, to avoid Oracle’s bonds.

NVIDIA (NVDA)

  • NVIDIA supplies chips used to build data centers and is involved in circular arrangements with Oracle and OpenAI. In the example discussed, Oracle buys NVIDIA chips, while NVIDIA invests in OpenAI, which pays Oracle for computing capacity.
  • The podcast describes NVIDIA as being “all over the place” in these interconnected AI deals, which could transmit financial trouble between companies if one part of the chain weakens.

Takeaways

  • NVIDIA is exposed not only to demand for its chips but also to the financial health of customers and partners in the AI infrastructure ecosystem.
  • The transcript offers no NVIDIA price target or direct stock recommendation. Its main caution is that interconnected financing can make risks harder to isolate.

Alphabet (GOOGL)

  • Alphabet is identified as one of the major hyperscalers investing in data centers and issuing bonds to finance infrastructure.
  • In August, Alphabet sold $25 billion in bonds and received orders worth more than four times the amount offered, indicating strong investor demand. However, investors wanted somewhat higher yields, suggesting that borrowing costs and pricing were becoming more important.
  • Alphabet also sold bonds denominated in Australian dollars for the first time.

Takeaways

  • Strong demand for Alphabet’s bond issue suggests investors were willing to lend, but the higher yield demanded shows that demand does not mean financing is cost-free.
  • The transcript discusses Alphabet’s bonds, not a specific outlook or recommendation for its stock.

Amazon (AMZN)

  • Amazon is named as a major hyperscaler building data-center capacity and borrowing to fund future infrastructure needs.
  • The podcast says Amazon had more debt outstanding than any other non-financial company in the comparison discussed. It also notes that Amazon’s bonds compete with U.S. Treasuries for investor dollars.

Takeaways

  • Amazon’s data-center spending could support future AI services, but borrowing at this scale makes financing costs and the returns on infrastructure spending relevant considerations.
  • The transcript provides no Amazon price target or stock recommendation.

Meta Platforms (META)

  • Meta is identified as a major hyperscaler planning data-center investments. The podcast says companies such as Meta may rent data-center buildings rather than own them.
  • Some project-related bonds can last longer than the leases supporting them. If leases must be renewed before bonds mature, investors face lease-renewal risk as well as the risk that the facilities could become technologically outdated or no longer needed.

Takeaways

  • For Meta, the key issue raised is whether long-term demand for data-center space will justify the cost and duration of the commitments.
  • The transcript does not offer a Meta stock recommendation or price target.

Microsoft (MSFT)

  • Microsoft is listed among the large U.S. companies whose fortunes can affect investors with broad U.S. stock exposure. It is also named as a hyperscaler and as a participant in some of the interconnected AI deals.

Takeaways

  • Microsoft’s AI and data-center investment is part of a broader industry build-out. The discussion highlights potential infrastructure and financing risks, but does not single out Microsoft’s debt or provide a stock-specific view.
  • No price target or direct recommendation is mentioned.

OpenAI (Private company)

  • OpenAI is a major customer of Oracle’s data-center capacity and has reportedly promised to pay Oracle about $300 billion over roughly five years.
  • OpenAI has delayed going public, and its financial statements are not publicly available. The podcast says this makes it difficult for investors to judge whether its projected growth and payments are sustainable.

Takeaways

  • OpenAI is not presented as a publicly traded investment in the transcript. Its importance for investors is indirect: Oracle’s financial outlook depends in part on OpenAI’s ability to meet its commitments.
  • The lack of publicly available financial information is a specific transparency and customer-concentration concern raised in the discussion.

AI data-center and hyperscaler bonds

  • The five hyperscalers sold more than $90 billion in bonds in the prior year and had sold more than $130 billion so far in 2026, according to the figures cited in the podcast.
  • These bonds are competing with U.S. Treasuries for investor funds. The podcast says investor demand is present, but issuers may need to offer higher yields to attract buyers.
  • Risks discussed include construction delays, rising borrowing costs, leases expiring before bonds mature, and data centers becoming obsolete or unnecessary if demand forecasts prove too optimistic.
  • Some financing arrangements use complex structures and vary from deal to deal, creating additional uncertainty for bondholders.

Takeaways

  • Investors considering bond funds should recognize that those funds may hold debt tied to the data-center build-out, as well as bonds from large technology companies.
  • Look beyond headline demand for AI infrastructure: the ability to repay depends on projects being completed, demand persisting, and revenues supporting the debt over long periods.
  • The podcast gives no specific bond recommendation or yield target.

U.S. Treasuries

  • The podcast says new hyperscaler bond issuance is competing with Treasuries for investor dollars and helping push Treasury yields higher.
  • It also notes that rising interest rates increase borrowing costs across the economy.

Takeaways

  • Changes in demand for corporate bonds can affect the broader bond market, including Treasury yields.
  • The transcript does not provide a Treasury price or yield forecast.

S&P 500 and broad U.S. stock exposure

  • The episode says that investors in U.S. stocks are likely to have significant exposure to mega-cap technology companies because the largest S&P 500 constituents include Microsoft, NVIDIA, and Amazon.
  • The build-out could contribute to productivity and financial gains, but the episode also highlights the possibility that AI demand or data-center economics may disappoint.

Takeaways

  • Broad index exposure can still carry substantial concentration in a small number of large technology companies.
  • Investors may want to understand how much their portfolios depend on the success of the AI and data-center investment cycle. The podcast presents both potential gains and risks, not a forecast that the boom will succeed or fail.

Paramount Skydance and Warner Bros. Discovery

  • The podcast says Larry Ellison guaranteed $40 billion to support his son David’s Paramount Skydance merger deal with Warner Bros. Discovery. It describes this as a factor that made Oracle investors nervous.
  • The discussion does not provide a separate investment outlook for Paramount Skydance or Warner Bros. Discovery.

Takeaways

  • The deal is mentioned as a potential source of concern around the Ellison family’s financial commitments, rather than as a stock recommendation.
  • No price target or specific investment recommendation for either company is given.
Ask about this postAnswers are grounded in this post's content.
Episode Description
Your money might be fueling the AI boom … just as some AI data center bets are starting to look riskier. If you own stocks, that money is likely tied up in giant tech companies like Microsoft and Nvidia. And if your 401(k) has money invested in the US bond fund, you might be lending tech companies the money to build data centers.  Today on the show, two stories from our short daily podcast The Indicator from Planet Money on how risky these data center bets are for investors and the companies making them. And, of course, for all of us.  Read:  Our book: Planet Money: A Guide to the Economic Forces That Shape Your Life  Our weekly longform Planet Money newsletter Our weekly Indicator round-up newsletter Follow:  Instagram TikTok YouTube Facebook The original Indicator episodes were produced by Corey Bridges and Julia Ritchey. They were engineered by Kwesi Lee and Cena Loffredo, fact checked by Sierra Juarez and edited by Kate Concannon. This episode of Planet Money was produced by Vito Emanuel and Emma Peaslee. Alex Goldmark is our executive producer. Music: NPR Source Audio - "Saunter" and "Super-Fi"  Support public media with NPR+ and enjoy perks for over 25 podcasts like this one. This show’s perks include bonus episodes and sponsor-free listening. Learn more at plus.npr.org. See pcm.adswizz.com for information about our collection and use of personal data for sponsorship and to manage your podcast sponsorship preferences. NPR Privacy Policy
About Planet Money
Planet Money

Planet Money

By NPR

Wanna see a trick? Give us any topic and we can tie it back to the economy. At <em>Planet Money</em>, we explore the forces that shape our lives and bring you along for the ride. Don't just understand the economy – understand the world.<br><br><em>Wanna go deeper? <em>Subscribe to </em><em>Planet Money+ and get sponsor-free episodes of Planet Money, The Indicator, and Planet Money Summer School. Plus access to bonus content. It's a new way to support the show you love. Learn more at plus.npr.org/planetmoney</em><br></em>