The Blue Wave Meets AI: What the Midterms Mean for Data Centers | The Real Eisman Playbook Ep 84
The Blue Wave Meets AI: What the Midterms Mean for Data Centers | The Real Eisman Playbook Ep 84
Podcast51 min 30 sec
Listen to Episode
Note: AI-generated summary based on third-party content. Not financial advice. Read more.
Quick Insights
  • Favor cash-rich Microsoft (MSFT), Alphabet (GOOGL), and Amazon (AMZN) for AI exposure; their established infrastructure and financial strength may help them through either faster or slower AI spending.
  • For longer-term AI-chip exposure, consider NVIDIA (NVDA), Broadcom (AVGO), AMD (AMD), or Micron (MU), while recognizing that construction delays could pressure them in the near term; chip supply and demand may remain tight until late 2028 or early 2029.
  • Prefer lower-debt Nebius (NBIS) over highly leveraged CoreWeave (CRWV) among neoclouds; the discussion highlighted financing costs and weak returns as significant risks for CoreWeave.
  • Avoid treating all AI-linked stocks alike: the guests were notably bearish on Intel (INTC)’s execution and cautious on Oracle (ORCL) because of its financing and expansion risks.
  • The guests expected technology stocks to perform well over the next six to nine months, but framed this as a selective market rather than a broad buy-the-theme opportunity.
Detailed Analysis

AI Data Centers and Infrastructure

  • The guests expect AI adoption and data-center demand to keep growing, but see political opposition making U.S. data centers harder, slower, and more expensive to build. One guest suggested that roughly 10%–15% of planned data centers could be affected; this was an estimate, not a firm forecast.
  • The discussion framed a slowdown in AI capital spending as potentially constructive if it limits debt-fueled overbuilding. One guest said CapEx could peak in 2027, with free cash flow beginning to rebound in 2028. Another described a scenario in which data-center growth moderates to 20%–25% rather than an anticipated 70%, extending the investment cycle.
  • The guests cited strong chip demand in Taiwan and Korea and said supply and demand may not reach equilibrium until late 2028 or early 2029.
  • Bullish, but selective: the guests expect AI use cases to keep spreading, while emphasizing that the market may increasingly distinguish companies with durable demand and sound financing from those dependent on continuous capital raises.

Takeaways

  • Focus on businesses that can benefit from sustained AI demand without relying on aggressive borrowing or uninterrupted rapid construction.
  • Treat the potential slowdown in data-center construction as a risk to new-build suppliers, but not necessarily to existing infrastructure or companies with strong balance sheets.

Hyperscalers: Microsoft (MSFT), Alphabet (GOOGL), and Amazon (AMZN)

  • The guests argued that the large cloud providers could benefit whether AI spending accelerates or slows: more AI could drive cloud revenue, while less spending could improve cash flow.
  • Their existing data centers and ability to build in other locations were described as advantages if local opposition makes new construction more difficult.
  • The discussion also suggested that enterprises spending less on AI infrastructure could have more budget available for other technology purchases.

Takeaways

  • The guests were broadly positive on the hyperscalers’ ability to weather different AI-spending scenarios, citing their established infrastructure and financial capacity.
  • The transcript did not give price targets or specific buy recommendations.

NVIDIA (NVDA), Broadcom (AVGO), AMD (AMD), and Micron (MU)

  • The guests described continued strong demand for AI chips, while acknowledging that political and capital-spending constraints could moderate data-center growth.
  • One guest said chip demand in Taiwan and Korea was exceptionally strong and expected supply-demand balance only by late 2028 or early 2029.
  • The guests characterized NVIDIA, Broadcom, AMD, and Micron as companies that could benefit if AI infrastructure continues expanding at a more sustainable pace. They also said chip stocks were trading at roughly half to a quarter of a market multiple in their scenario, without specifying the valuation measure.
  • NVIDIA was also described as supporting the neocloud business model by helping create additional buyers for its chips beyond the largest cloud providers.

Takeaways

  • The discussion supports a positive long-term view of AI-chip demand, but also points to possible near-term pressure if data-center construction or capital spending slows.
  • The valuation comparisons were the guests’ characterizations; no specific price targets were provided.

Oracle (ORCL)

  • The guests were notably cautious. Oracle is trying to expand as a cloud provider while taking on substantial data-center spending and financing needs.
  • One guest said Oracle had made a large commitment to OpenAI, described in the discussion as a $300 billion bet, while its core business was suffering. Another said Oracle’s database business remains cash-generative, with roughly 50% margins, but that its balance sheet was more challenged than those of the major hyperscalers.
  • The company was characterized as a “high wire act”: successful execution could lead to substantial upside, but if the expansion or OpenAI relationship fails to deliver, there could be significant downside. The guests also criticized the clarity of Oracle’s investor communications.

Takeaways

  • The discussion presents Oracle as a higher-risk AI infrastructure investment than the established hyperscalers: its core business provides support, but its financing needs and execution risks are substantial.
  • The guests did not offer a specific price target or a direct buy/sell recommendation.

Intel (INTC) and AMD (AMD)

  • Intel was described as having a weak quarter and an unconvincing outlook, with investors concerned about its ability to navigate the business.
  • One guest called Intel a “broken company,” pointing to gross margins in the 30% range compared with competitors’ margins in the 50%–70% range and criticizing its chip-manufacturing execution.
  • AMD was presented as a contrast: the guests said its margins were in the 50% range and that it handled weakness in the PC market better than Intel. They praised CEO Lisa Su.

Takeaways

  • The discussion was bearish on Intel’s execution and more positive on AMD’s relative operating performance.
  • The guests emphasized that a capable leader or favorable deals may not be enough to fix Intel’s underlying manufacturing and execution problems.

AI Search and Alphabet (GOOGL)

  • Alphabet’s search revenue growth was discussed as slowing to 12%, down from 19% in the first quarter. One guest called it the quarter investors had feared and said it raised concern that AI search could weaken Google’s highly profitable search business.
  • The guests noted that Alphabet acknowledged earning less advertising revenue from AI search than from traditional search. They also discussed uncertainty about whether advertisers will pay the same rates when chatbot activity, rather than human users, drives traffic.
  • A more optimistic view was that Alphabet could recover if it successfully monetizes Gemini and its broader AI ecosystem. One guest said even a 30%–40% success rate in monetizing AI could support significant stock upside, though this was a personal view, not a formal target.

Takeaways

  • The central issue is whether Alphabet can preserve its advertising economics as users shift toward AI-based search.
  • The transcript gives reasons for caution, but also identifies successful Gemini monetization and Alphabet’s cloud business as potential supports.

Apple (AAPL)

  • Apple was described as a comparatively safe technology stock, with a strong balance sheet and cash flow and relatively light AI capital spending.
  • One guest said Apple could benefit from AI-enabled hardware and from its large installed base: roughly 1.5 billion iPhones and 2.5 billion iOS devices were cited. The guests also noted that Apple receives a share of some subscription revenue through its ecosystem.
  • A new hardware cycle, including a foldable device, was discussed as a possible opportunity. The view that 20% of the world could access AI through Apple devices was a guest estimate.

Takeaways

  • The guests saw Apple as a potential way to gain exposure to consumer AI without investing directly in costly frontier-model development.
  • The discussion’s thesis depends on Apple turning its installed base and services ecosystem into meaningful AI-related revenue; the guests did not provide a price target.

Palantir (PLTR), Snowflake (SNOW), Datadog (DDOG), CrowdStrike (CRWD), Shopify (SHOP), and JFrog (FROG)

  • These companies were named as software businesses the guests saw as benefiting from AI adoption, with accelerating revenue or improving growth prospects.
  • Palantir was highlighted for its enterprise AI offering and its role as a “control plane” or intermediary that can help companies use different AI models while protecting sensitive information.
  • Snowflake and other companies were cited as having advanced AI-related offerings compared with more traditional software providers. Cybersecurity firms such as CrowdStrike were also described as positioned to defend their markets as AI models expand into security-related functions.

Takeaways

  • The guests favored software companies they believe can turn AI adoption into revenue, rather than simply face competition from AI tools.
  • Their comments were positive, but they did not provide specific price targets or recommendations for these stocks.

Salesforce (CRM), ServiceNow (NOW), Adobe (ADBE), and Other Traditional Software

  • One guest said traditional software companies have a window to adapt their products and monetize AI, but that the transition may be difficult and take time.
  • ServiceNow’s modest guidance miss drew a sharp market reaction, which the guests attributed partly to investors’ low tolerance for disappointing results.
  • On Adobe, one guest argued that reproducing software features does not necessarily replace the enterprise workflows, governance, collaboration, and services customers receive from Adobe.
  • Salesforce, ServiceNow, Adobe, and UiPath were cited as companies that need to demonstrate how they will adapt and monetize AI.

Takeaways

  • The discussion was mixed: AI may pressure some software products, but the guests argued that established enterprise workflows and services can be more durable than the underlying code alone.
  • For these companies, the key question raised was whether they can translate AI into useful products and revenue before investors lose patience.

AI-Exposed Software: UiPath (PATH), Pega (PEGA), Appian (APPN), NICE (NICE), and Five9 (FIVN)

  • The guests identified workflow automation and call-center software as areas more directly exposed to AI competition.
  • UiPath, Pega, and Appian were cited as examples of workflow companies that could be affected; NICE and Five9 were cited as call-center companies facing potential pressure.

Takeaways

  • The discussion was cautious on these businesses because AI could automate or change parts of the services they provide.
  • The transcript did not give price targets or quantify the likely impact.

Cybersecurity: CrowdStrike (CRWD), Palo Alto Networks (PANW), and Zscaler (ZS)

  • The guests said cybersecurity companies may benefit as enterprises look to protect their data and manage AI systems.
  • They also noted that AI model providers could expand into cybersecurity, creating competition for established cybersecurity firms such as CrowdStrike, Palo Alto Networks, and Zscaler.

Takeaways

  • The sector may benefit from greater demand for AI security, but the guests also highlighted competition from AI providers entering cybersecurity.
  • Their comments were therefore positive on the opportunity but not uniformly bullish on every company.

Neoclouds: CoreWeave (CRWV), Nebius (NBIS), and IREN

  • Neoclouds were described as smaller data-center operators that borrow to build computing capacity and rent it out, without the broader businesses that support major hyperscalers.
  • The guests warned that data-center construction constraints and more expensive financing could hurt the group. One said CoreWeave’s borrowing costs could reach 15% for new borrowing, compared with an estimated blended cost around 9% at the time.
  • CoreWeave was criticized as generating low-single-digit returns while borrowing at higher rates. One guest said its equity holders could be left with little value if the economics do not improve. Insider stock sales were also raised as a concern.
  • The guests distinguished Nebius, saying it had little debt and a stronger focus on profitability. One guest thought both CoreWeave and Nebius could be leaders, while also cautioning that not all neoclouds would survive.
  • The transcript compared the business model to Bitcoin mining: operators may do well while compute prices are rising, but could struggle if prices level off.

Takeaways

  • The guests were cautious on neoclouds overall, with particular concern about leverage, financing costs, and returns on capital.
  • Nebius was presented as relatively better positioned because of its lower debt and focus on profitability; this was a guest opinion, not a specific recommendation.
  • CoreWeave’s economics were described as especially challenging for shareholders, even if customers and chip suppliers benefit from its capacity.

Data-Center Equipment and Power: Arista Networks (ANET), Cisco (CSCO), Dell (DELL), GE Vernova (GEV), and Quanta Services (PWR)

  • Arista, Cisco, and Dell were discussed as suppliers connected to existing data centers and technology infrastructure. One guest argued that demand from already-built facilities could support them even if some new data centers are delayed.
  • The guests also said the market may be distinguishing between companies tied to new construction and those benefiting from installed infrastructure.
  • GE Vernova and Quanta Services were mentioned as potential losers if fewer data centers are built, since reduced construction could weigh on related equipment and power demand.

Takeaways

  • The discussion suggests a distinction between suppliers to existing data centers and businesses more dependent on new construction.
  • Delayed or cancelled projects could weigh on construction-linked suppliers, while the transcript offered a more favorable view of companies serving the existing infrastructure base.

OpenAI and Anthropic

  • The guests agreed that leading AI models may have a temporary advantage because talent and advanced model development are concentrated at a small number of companies, including OpenAI, Anthropic, and Google DeepMind.
  • They also said models could become more interchangeable over time as open-weight and lower-cost alternatives improve, potentially leading to price competition.
  • Anthropic was described as pivoting toward enterprise customers, but the guests said companies may use “control plane” or orchestration software to switch among models and keep proprietary data from being exposed to model providers.
  • One guest cautioned that Anthropic’s market opportunity could narrow as companies diversify their AI spending. The discussion said the case for investing in an Anthropic IPO would depend on its growth prospects and offering price; no price was available.
  • OpenAI and Anthropic are private companies in the transcript, not publicly traded stocks.

Takeaways

  • The discussion favors looking beyond the models themselves to enterprise tools that help companies use AI while controlling data and model choice.
  • Any investment view on a future Anthropic IPO would depend on its valuation, growth, and ability to keep its market expanding despite competition.

Bitcoin (BTC) and Bitcoin Miners

  • Bitcoin was mentioned only as an analogy for neocloud economics: the guests said Bitcoin miners can prosper when the underlying asset is rising quickly and struggle when conditions turn less favorable.
  • The comparison was used to illustrate how neoclouds could be vulnerable if compute prices stop rising.

Takeaways

  • The transcript did not make a direct investment case for Bitcoin or provide a price target.
  • The relevant point was the cyclical risk to businesses whose profitability depends on rising prices for what they sell.

Agentic AI and Potentially Disrupted Businesses

  • The guests said agentic AI could eventually comparison-shop and complete tasks for consumers, potentially reducing the value of brand or distribution advantages in areas such as travel, insurance, and financial services.
  • Booking Holdings (BKNG), Progressive (PGR), and Geico were mentioned as examples of businesses that could face pressure if AI agents steer consumers toward the lowest-cost or best-rated option.
  • JPMorgan Chase (JPM) was mentioned in the context of banks restricting access by AI agents. The guests said regulation and consumer trust could delay widespread adoption; one described normalization as still a few years away.
  • The discussion also suggested that companies offering good products at competitive prices could benefit as AI reduces friction for consumers.

Takeaways

  • Treat the potential impact on travel, insurance, and financial-services companies as a longer-term competitive risk, not an immediate forecast: the guests emphasized regulatory and trust barriers.
  • The discussion was more positive for consumers and companies that can compete effectively on value than for businesses dependent mainly on brand or customer friction.

Overall Market View

  • One guest expected technology stocks to perform well over the next six to nine months, as investors distinguish winners from losers.
  • Another described a “Goldilocks” scenario in which slower, more disciplined AI investment extends the cycle and reduces reliance on debt.
  • The guests’ broad positive thesis was that AI use cases are spreading and that companies with established infrastructure, strong cash generation, or clear enterprise value may benefit.
  • Key risks raised in the discussion included political resistance to data centers, rising financing costs, possible AI overbuilding, pressure on traditional search advertising, and competition from increasingly capable AI models.

Takeaways

  • The conversation favored selectivity over buying the entire AI theme: the guests saw meaningful differences between cash-generative hyperscalers, AI-benefiting software companies, and highly leveraged infrastructure operators.
  • Timelines and valuation comments were guest views from the discussion, not guaranteed outcomes or individualized investment advice.
Ask about this postAnswers are grounded in this post's content.
Episode Description
Sign up for The Real Eisman Playbook Premium at http://www.steveeisman.com On this episode of The Real Eisman Playbook, Steve Eisman welcomes back Dan Ives of Yorkville and Gil Luria of Davidson the day after the midterms to untangle how the AI narrative has shifted. They also discuss why Google's slowing search growth was the quarter investors feared, why Anthropic and OpenAI's slice of the market may shrink, and why CoreWeave is essentially an Nvidia special purpose vehicle.   Connect with Steve Eisman and access all things The Real Eisman Playbook: 🌐 https://linktr.ee/realeismanplaybook → Follow on socials, watch episodes, and get the latest updates — all in one place.     Disclaimer: The financial opinions expressed are for information purposes only. The opinions expressed by the hosts and participants are not an attempt to influence specific trading behavior, investments, or strategies. Past performance does not necessarily predict future outcomes. No specific results or profits are assured when relying on this content.   Before making any investment or trade, evaluate its suitability for your circumstances and consider consulting your own financial or investment advisor. The financial products discussed in 'The Eisman Playbook' carry a high level of risk and may not be appropriate for many investors. If you have uncertainties, it's advisable to seek professional advice. Remember that trading involves a risk to your capital, so only invest money you can afford to lose. Derivatives are unsuitable for all investors and involve the risk of losing more than the amount originally deposited and any profit you might have made. This communication is not a recommendation or offer to buy, sell, or retain any specific investment or service.   Copyright ©2026 Steve Eisman
About The Real Eisman Playbook
The Real Eisman Playbook

The Real Eisman Playbook

By Steve Eisman

The Real Eisman Playbook is your front-row seat to the insights, strategies, and perspectives of legendary investor Steve Eisman. Best known for predicting the 2008 financial crisis, Steve brings his sharp analysis and no-nonsense approach to dissecting the markets, global economy, and investment trends shaping the future. Whether you’re a seasoned investor or just curious about how the financial world really works, The Eisman Playbook delivers the knowledge you need to stay ahead. Tune in for expert commentary, candid conversations, and actionable takeaways from one of Wall Street’s most influential minds. Follow Us on Social Media!