【The Motley Fool】无人谈论的AI终极瓶颈及配置策略

【The Motley Fool】无人谈论的AI终极瓶颈及配置策略 — 浑水摸鱼清源

浑水摸鱼清源

AI-assisted summary of 【The Motley Fool】无人谈论的AI终极瓶颈及配置策略 — original source. Check the original for full context.

Source published:

Key ideas

  • Vistra Corp. (VST) is positioned to benefit from surging AI power demand through operational transmission interconnections across ERCOT and Ohio, supported by a reported $4 billion federal loan for nuclear plant uprates. (0:00)
  • GE Vernova (GEV) is leveraging massive equipment backlogs by demanding customer prepayments and hiking turbine prices, strictly avoiding speculative manufacturing capacity expansion. (3:45)
  • Solaris Energy Infrastructure (SEI) has more than doubled its revenue by deploying behind-the-meter, on-site microgrids that allow data centers to bypass multi-year utility interconnection queues. (5:40)
  • Xcel Energy (XEL) provides a low-beta bond alternative trading under 20x P/E with a ~3% dividend yield, capturing hyperscaler data center buildouts across the Upper Midwest and Mountain West. (7:25)

Detailed notes

## Background This video was published on Bilibili by creator 浑水摸鱼清源 on October 13, 2026, presenting a Motley Fool segment titled "The AI Bottleneck Nobody Is Talking About (4 Stocks to Buy)." According to the page description, the program explores how energy availability, grid transmission, and cooling have superseded compute chips as AI's primary structural bottlenecks, presenting four infrastructure and power stocks poised to benefit.

## TL;DR Moving past speculative chip plays, the discussion focuses on how physical power and grid interconnection delays have become the defining bottleneck for hyperscale AI expansion in 2026. The program analyzes four distinct energy plays: independent power producer Vistra Corp. (VST), which leverages existing grid hooks and nuclear uprates; GE Vernova (GEV), which exercises semiconductor-style pricing power over its massive turbine backlogs; Solaris Energy Infrastructure (SEI), which provides rapid behind-the-meter turnkey generation to bypass grid queues; and regulated utility Xcel Energy (XEL), which offers defensive, inflation-protected dividend growth across data-center-friendly Midwestern territories. Together, these companies demonstrate how investors can capture durable cash flows from generative AI by financing baseload power, industrial equipment, microgrids, and steady utility rate bases.

## Atmosphere & style The presentation carries a grounded, pragmatic, and analytically sober tone that deliberately cuts through speculative tech hype with hard physical realities. The hosts speak with the measured wisdom of veteran market observers, cautioning viewers against chasing volatile momentum while emphasizing capital discipline, contractual leverage, and operational moats. Listening feels like sitting in on an institutional strategy briefing, as the narrative methodically recalibrates the investor mindset from software euphoria to the gritty, capital-intensive world of turbines, switchyards, and baseload megawatts.

## Bulletin - [00:00](https://www.bilibili.com/video/BV1jepD6aEGh?t=0s) Vistra Corp. (VST) is positioned to benefit from surging AI power demand through operational transmission interconnections across ERCOT and Ohio, supported by a reported $4 billion federal loan for nuclear plant uprates. - [03:45](https://www.bilibili.com/video/BV1jepD6aEGh?t=225s) GE Vernova (GEV) is leveraging massive equipment backlogs by demanding customer prepayments and hiking turbine prices, strictly avoiding speculative manufacturing capacity expansion. - [05:40](https://www.bilibili.com/video/BV1jepD6aEGh?t=340s) Solaris Energy Infrastructure (SEI) has more than doubled its revenue by deploying behind-the-meter, on-site microgrids that allow data centers to bypass multi-year utility interconnection queues. - [07:25](https://www.bilibili.com/video/BV1jepD6aEGh?t=445s) Xcel Energy (XEL) provides a low-beta bond alternative trading under 20x P/E with a ~3% dividend yield, capturing hyperscaler data center buildouts across the Upper Midwest and Mountain West.

## Key takeaways - [00:00](https://www.bilibili.com/video/BV1jepD6aEGh?t=0s) The defining AI constraint has evolved from compute chips to fundamental physical boundaries: power generation, transmission line capacity, and grid interconnection. - [00:00](https://www.bilibili.com/video/BV1jepD6aEGh?t=0s) Established power producers with already-interconnected grid assets represent a far lower execution risk than pre-revenue advanced nuclear technology startups. - [00:00](https://www.bilibili.com/video/BV1jepD6aEGh?t=0s) Nuclear uprating enables Vistra to expand output using existing switchyards and transmission infrastructure, bypassing lengthy environmental and regulatory approval cycles. - [00:00](https://www.bilibili.com/video/BV1jepD6aEGh?t=0s) Independent Power Producers (IPPs) hold superior leverage over regulated utilities in negotiating long-term clean baseload power contracts directly with hyperscalers like AWS and Meta. - [00:00](https://www.bilibili.com/video/BV1jepD6aEGh?t=0s) Transmission line interconnection queues represent the slowest link in AI infrastructure, granting outsized pricing power to operational power generators. - [03:45](https://www.bilibili.com/video/BV1jepD6aEGh?t=225s) GE Vernova produces the heavy-duty gas and wind turbines required to build new generating plants to address massive grid deficits. - [03:45](https://www.bilibili.com/video/BV1jepD6aEGh?t=225s) Rather than overbuilding factory footprint for a potential 3-to-5-year cycle, GE Vernova is improving margins by requiring upfront customer prepayments for turbine build slots. - [03:45](https://www.bilibili.com/video/BV1jepD6aEGh?t=225s) Enterprise token usage across foundational model labs like OpenAI and Anthropic directly drives demand down into heavy industrial turbine order books. - [05:40](https://www.bilibili.com/video/BV1jepD6aEGh?t=340s) Behind-the-meter power solutions provide a critical workaround for data centers that cannot wait years for centralized utility grid connections. - [05:40](https://www.bilibili.com/video/BV1jepD6aEGh?t=340s) Solaris Energy Infrastructure has evolved from mobile equipment into turnkey microgrid management, doubling revenue on urgency from compute cluster operators. - [07:25](https://www.bilibili.com/video/BV1jepD6aEGh?t=445s) Regulated utilities offer capital-base growth with guaranteed rates of return, providing low-volatility anchors that outpace fixed-coupon bonds during inflationary periods. - [07:25](https://www.bilibili.com/video/BV1jepD6aEGh?t=445s) Hyperscale data centers are migrating to interior markets such as Minnesota, Colorado, and North Dakota due to cheap land, cooler climates, and receptive local regulators.

## Detailed walkthrough ### [00:00](https://www.bilibili.com/video/BV1jepD6aEGh?t=0s) The AI Power Bottleneck and Vistra (VST): Nuclear Uprates & IPP Tailwinds

Bottlenecks in artificial intelligence have become the defining challenge of 2026, shifting from compute hardware constraints to an even more fundamental physical barrier: electricity supply and grid transmission. Rather than chasing speculative, pre-revenue advanced nuclear startups, investors looking to capitalize on this multi-decade headwind can turn to established power producers. Vistra Corp. (ticker: VST) operates across a broad energy spectrum—running the gamut from solar and natural gas to coal and nuclear power. Operating both as a direct retail supplier and as a primary provider into wholesale regional grids such as ERCOT in Texas and markets reaching through to Ohio, Vistra holds a commanding footprint of already-interconnected generation assets.

A central catalyst for Vistra is a reported $4 billion loan from the U.S. federal government aimed at nuclear "uprating." Because constructing brand-new nuclear power plants requires vast capital expenditures and years of regulatory approvals, uprating allows Vistra to expand the power output of its existing reactors using the transmission lines and switchyards already physically connected to the grid. Hyperscale tech giants are actively hunting for these exact generation assets; upgraded power from these facilities can flow directly into massive data centers run by Meta and Amazon's AWS.

The discussion clarifies the structural difference between regulated utilities and Independent Power Producers (IPPs). In a traditional regulated utility framework, electricity rates and margins are tightly capped by public utility commissions in exchange for monopoly territories. As an unregulated IPP, Vistra takes on merchant market risks: when wholesale power prices fluctuate, profits can be uneven. However, in an environment of unprecedented AI power demand where hyperscalers are desperate for long-term clean baseload contracts—evidenced by landmark deals like Microsoft's arrangement to restart Three Mile Island—IPPs gain tremendous leverage. Because the physical interconnection of transmission lines is often the slowest link in the development chain, having operational assets already in the ground positions Vistra for a durable, multi-decade expansion.

- **Ticker:** VST (Vistra Corp.) - **Key Government Support:** Reportedly set to receive approximately $4 billion in U.S. government loans for nuclear plant uprates. - **Grid Footprint:** ERCOT in Texas extending north through the PJM footprint into Ohio. - **Target Hyperscalers:** Direct beneficiary facilities supporting Meta and Amazon Web Services (AWS) data centers. - **Core Advantage:** Interconnection and transmission lines are already operational, bypassing multi-year interconnection queues.

[03:45](https://www.bilibili.com/video/BV1jepD6aEGh?t=225s) GE Vernova (GEV): Turbine Backlogs and Heavy Industrial Pricing Power

While Vistra capitalizes on generation assets already hooked to the grid, GE Vernova represents the critical hardware bottleneck for capacity that still needs to be built. Spun off from General Electric to house its power and renewable energy businesses, GE Vernova produces the gas and wind turbines that form the backbone of modern power plant construction. With utility backlogs swelling across the country, developers simply cannot procure heavy generation equipment fast enough to meet demand.

Crucially, GE Vernova's management is not repeating the historical cyclical trap of doubling manufacturing footprint just to meet a short-term spike. Having navigated energy cycles for decades, the company refuses to risk being "left holding the bag" if demand were to normalize over a 3-to-5-year horizon. Instead, GE Vernova is handling unprecedented demand much like Micron Technology manages memory chip supply crunches: by raising prices, expanding margins, and forcing prospective buyers to provide upfront prepayments simply to secure manufacturing slots for turbines. This provides significant commercial pricing power, ensuring that as long as consumer and enterprise token usage expands across AI labs like OpenAI and Anthropic, the downstream cash flows inevitably concentrate in the turbine order books.

```mermaid flowchart LR accTitle: AI Demand to Equipment Pricing Power Pipeline accDescr: Diagram illustrating how token usage drives AI data center demand, grid power deficits, and turbine supplier pricing leverage. subgraph Compute ["AI Compute Expansion"] A["Token Usage (OpenAI / Anthropic)"] --> B["Hyperscale Data Center Buildouts"] end subgraph PowerGrid ["Energy Constraints"] B --> C["Massive Electricity Deficits"] C --> D["New Gas & Wind Plant Projects"] end subgraph Equipment ["Supplier Leverage"] D --> E["GE Vernova Heavy Turbines"] E --> F["Pricing Power & Prepayments"] end ```

- **Ticker:** GEV (GE Vernova) - **Primary Offerings:** Heavy-duty gas turbines, wind turbines, and grid infrastructure hardware. - **Capital Discipline:** Intentional choice not to rapidly overbuild manufacturing capacity, protecting long-term returns. - **Operational Strategy:** Emulating semiconductor pricing dynamics by increasing equipment prices, improving operating margins, and mandating customer advance prepayments.

[05:40](https://www.bilibili.com/video/BV1jepD6aEGh?t=340s) Solaris Energy Infrastructure (SEI): Rapid Behind-the-Meter Generation

For data center operators unable to wait several years for utility grid interconnects, the alternative is "behind-the-meter" on-site power generation. Solaris Energy Infrastructure (ticker: SEI) serves as the exact counterweight to the centralized grid thesis. When local utilities and regional transmission organizations cannot supply the requisite multi-megawatt capacity on an acceptable timeline, developers bypass the grid altogether by installing dedicated power plants directly at the data center site.

Solaris has evolved beyond its origins in mobile natural gas generators into a full-service, turnkey microgrid operator. The company designs, installs, and manages self-contained generation facilities on-site, effectively providing immediate, islanded power. Driven by the sheer urgency of data center operators to bring compute clusters online, Solaris has more than doubled its revenue over the past few years. This positions behind-the-meter engineering as a vital stopgap and permanent solution alongside centralized utility buildouts.

- **Ticker:** SEI (Solaris Energy Infrastructure) - **Deployment Model:** Behind-the-meter, on-site turnkey power generation and microgrid management. - **Financial Metric:** Revenue has more than doubled over the past several years. - **Value Proposition:** Solves immediate interconnection delays by deploying standalone natural gas generation directly adjacent to data center loads.

[07:25](https://www.bilibili.com/video/BV1jepD6aEGh?t=445s) Xcel Energy (XEL): Boring Regulated Utilities as Bond Alternatives in the Data Center Heartland

To round out the AI energy theme, Xcel Energy (ticker: XEL) represents the conservative, traditional regulated utility play. Unlike merchant generators or specialized microgrid contractors, Xcel operates under standard public utility commissions, delivering baseline residential and commercial electricity and earning guaranteed rates of return on capital invested into its system rate base. While it lacks the explosive volatility of momentum tech stocks—with the hosts cautioning that investors expecting "the next NVIDIA" should "keep dreaming"—Xcel holds unique geographic tailwinds that the broader market often overlooks.

Xcel's territory covers key interior and Midwestern states, including Colorado, Minnesota, and North Dakota. These regions are increasingly favored by hyperscale developers due to cheap, plentiful land, lower cooling costs, abundant wind and solar resources, and local communities and regulators that are far more welcoming to new generation and data center infrastructure than coastal markets. With a price-to-earnings multiple under 20 and a dividend yield hovering around 3%, Xcel provides a stable, low-beta anchor for a portfolio. The hosts emphasize that for conservative investors nearing retirement, such regulated utilities can serve as superior alternatives to fixed-income bonds: while a traditional 20-year bond pays an unvarying fixed coupon that cannot adjust to inflation, an expanding utility rate base allows companies like Xcel to steadily grow their dividend payouts over time.

- **Ticker:** XEL (Xcel Energy) - **Geographic Footprint:** Upper Midwest and Mountain West states, notably Colorado, Minnesota, and North Dakota. - **Valuation & Yield:** Price-to-earnings (P/E) multiple trading under 20x; dividend yield around ~3%. - **Investment Profile:** Low-beta asset with inflation-protected income growth driven by regulatory rate-base additions, providing a defensive alternative to fixed-coupon bonds.

## Notable quotes > [00:00](https://www.bilibili.com/video/BV1jepD6aEGh?t=0s) Bottlenecks in artificial intelligence have become the defining challenge of 2026, shifting from compute hardware constraints to an even more fundamental physical barrier: electricity supply and grid transmission.

> [00:00](https://www.bilibili.com/video/BV1jepD6aEGh?t=0s) Because constructing brand-new nuclear power plants requires vast capital expenditures and years of regulatory approvals, uprating allows Vistra to expand the power output of its existing reactors using the transmission lines and switchyards already physically connected to the grid.

> [00:00](https://www.bilibili.com/video/BV1jepD6aEGh?t=0s) Because the physical interconnection of transmission lines is often the slowest link in the development chain, having operational assets already in the ground positions Vistra for a durable, multi-decade expansion.

> [03:45](https://www.bilibili.com/video/BV1jepD6aEGh?t=225s) Crucially, GE Vernova's management is not repeating the historical cyclical trap of doubling manufacturing footprint just to meet a short-term spike.

> [03:45](https://www.bilibili.com/video/BV1jepD6aEGh?t=225s) Instead, GE Vernova is handling unprecedented demand much like Micron Technology manages memory chip supply crunches: by raising prices, expanding margins, and forcing prospective buyers to provide upfront prepayments simply to secure manufacturing slots for turbines.

> [05:40](https://www.bilibili.com/video/BV1jepD6aEGh?t=340s) For data center operators unable to wait several years for utility grid interconnects, the alternative is 'behind-the-meter' on-site power generation.

> [05:40](https://www.bilibili.com/video/BV1jepD6aEGh?t=340s) When local utilities and regional transmission organizations cannot supply the requisite multi-megawatt capacity on an acceptable timeline, developers bypass the grid altogether by installing dedicated power plants directly at the data center site.

> [05:40](https://www.bilibili.com/video/BV1jepD6aEGh?t=340s) Driven by the sheer urgency of data center operators to bring compute clusters online, Solaris has more than doubled its revenue over the past few years.

> [07:25](https://www.bilibili.com/video/BV1jepD6aEGh?t=445s) While it lacks the explosive volatility of momentum tech stocks—with the hosts cautioning that investors expecting 'the next NVIDIA' should 'keep dreaming'—Xcel holds unique geographic tailwinds that the broader market often overlooks.

> [07:25](https://www.bilibili.com/video/BV1jepD6aEGh?t=445s) With a price-to-earnings multiple under 20 and a dividend yield hovering around 3%, Xcel provides a stable, low-beta anchor for a portfolio.

## Mentioned Amazon Web Services (AWS), Anthropic, Broadcom (AVGO), Constellation Energy (CEG), ERCOT, GE Vernova (GEV), General Electric, Meta, Micron Technology, Microsoft, NVIDIA, OpenAI, PJM, Solaris Energy Infrastructure (SEI), Three Mile Island, Vertiv Holdings (VRT), Vistra Corp. (VST), Xcel Energy (XEL)

## Video description 【The Motley Fool】 无人谈论的AI终极瓶颈及配置策略 The AI Bottleneck Nobody Is Talking About (4 Stocks to Buy)

大家好，欢迎来到本期美股投资栏目。当全市场都在盯着英伟达的GPU芯片和各大科技巨头的百亿大模型军备竞赛时，一个更大的、关乎整个AI命运的隐形瓶颈正在浮出水面——那就是电力供应、电网传输以及数据中心散热。今天我们来深度剖析这个被严重忽视的结构性瓶颈，并盘点四家具备极高护城河、值得在当前节点重点布局的优质标的。

电力短缺与冷却革命：AI淘金热背后的硬约束 回顾过去两年，生成式AI的发展速度超出了所有人的预期，但随之而来的是天文数字般的能耗需求。一座现代化超大规模AI数据中心的耗电量往往高达数百兆瓦，甚至相当于一个中型城市的用电量。然而，传统的电网建设周期长达数年，电力供应严重滞后于数据中心的扩张速度。与此同时，芯片功耗激增带来的散热难题也达到了物理极限，传统风冷技术难以为继。谁能解决电力与热管理的瓶颈，谁就掌握了AI下半场的咽喉。

Constellation Energy (CEG)：锁定清洁核能的巨无霸 在电力供应方面，核能由于其全天候、零碳排放的基荷电力特性，成为了科技巨头采购AI电力的首选。Constellation Energy作为美国最大的无碳能源生产商，近期股价表现稳健，市值突破千亿美元大关，forward PE保持在28倍左右。公司通过与微软等巨头签署长期电力采购协议（PPA），将旗下核电站的电力直接锁定给数据中心。这种长单模式不仅保障了现金流的极高确定性，也使其成为AI淘金热中稳赚不赔的卖水人。

Vertiv Holdings (VRT)：数据中心热管理的绝对统治者 当GPU算力密度不断刷新纪录时，液体冷却（Liquid Cooling）技术已从可选项变为必选项。Vertiv作为数据中心热管理和基础设施解决方案的领军企业，市盈率约35-40倍，过去一年营收保持30%以上的高速增长。随着各大云厂商加速向液冷架构过渡，Vertiv的订单积压（Backlog）持续创下历史新高。其毛利率在供应链优化和高附加值产品推动下不断改善，是直击AI散热痛点的核心标的。

Broadcom (AVGO) 与 基础设施生态链 除了直接的电力与散热，定制化ASIC芯片和网络连接带宽同样是打破数据传输瓶颈的关键。博通（Broadcom）市值突破8000亿美元级别，动态市盈率约30倍，凭借其在AI加速器定制芯片（ASIC）设计以及高速网络交换芯片（Tomahawk/Jericho系列）的垄断地位，持续收割大厂自研芯片红利。其高额的股息增长率与强劲的自由现金流，为长期投资者提供了极强的安全边际。

针对当前AI基础设施瓶颈期的投资布局，特此梳理三条可执行操作指引： 采取分批建仓与网格定投策略，避免在周期波动高点一次性满仓，重点关注核能与热管理板块的回调买入机会。 动态平衡投资组合，将高成长、高估值的散热与算力设备股（如Vertiv）与现金流稳健、拥有长期电力长单的能源巨头（如CEG）进行合理配置。 密切追踪各标的的订单积压（Backlog）数据、自由现金流转化率以及各大云厂商资本开支（Capex）的指引变化。

简洁时间轴 [00:00] AI狂飙背后的电力短缺与散热危机 [03:15] 核能巨头Constellation Energy长单与基本面解析 [07:40] 液冷龙头Vertiv的订单增长与散热技术优势 [12:20] 博通在定制芯片与网络传输领域的护城河 [18:50] 投资组合配置策略与总结建议

[Source](https://www.bilibili.com/video/BV1jepD6aEGh)

中文简报

【The Motley Fool】无人谈论的AI终极瓶颈及配置策略 抛开投机性的芯片炒作，讨论重点转向实体电力供应与电网互联延迟如何成为2026年超大规模AI扩张的核心瓶颈。节目分析了四类不同的能源投资标的：独立发电生产商Vistra Corp. (VST)依托现有电网接口与核电增容获益；GE Vernova (GEV)凭借庞大的燃气轮机积压订单行使类似半导体行业的定价权；Solaris Energy Infrastructure (SEI)提供快速的“表后”交钥匙发电方案以绕过电网排队；受监管公用事业公司Xcel Energy (XEL)则在对数据中心友好的中西部地区提供防御性、抗通胀的股息增长。这些公司共同表明，投资者可通过布局基荷电力、工业设备、微电网及稳健的公用事业费率基数，从生成式AI中获取持久的现金流。 Vistra Corp. (VST) 依托在ERCOT及俄亥俄州现有的输电并网接口，并得到据报道达40亿美元核电站增容联邦贷款的支持，有望从激增的AI用电需求中获利。 (0:00) GE Vernova (GEV) 正在充分利用庞大的设备积压订单，要求客户预付款并上调燃气轮机价格，同时严格避免投机性扩张制造产能。 (3:45) Solaris Energy Infrastructure (SEI) 通过部署“表后”现场微电网，帮助数据中心绕过多达数年的公用事业并网排队，推动其营收实现翻倍以上增长。 (5:40) Xcel Energy (XEL) 市盈率低于20倍且股息率约3%，作为低贝塔值的类债券替代标的，持续受益于中西部北部和山地西部地区的超大规模数据中心建设浪潮。 (7:25)

Reading summary

As artificial intelligence bottlenecks shift from computing hardware to electricity supply and grid transmission, infrastructure investors are rotating into operational power producers, equipment manufacturers, and microgrid providers to capture long-term hyperscaler demand. Operational Generation Assets and Merchant Pricing Leverage (0:00) The primary barrier facing artificial intelligence expansion has transitioned toward power generation and grid interconnection queues. Rather than pursuing speculative, pre-revenue advanced nuclear projects, capital is flowing to established independent power producers (IPPs) like Vistra Corp. (VST). Vistra maintains a diversified generation fleet across solar, natural gas, coal, and nuclear assets, operating directly within competitive wholesale markets such as ERCOT in Texas and across the PJM footprint into Ohio. Because physical grid interconnects represent multi-year bottlenecks, Vistra's operational assets provide hyperscalers like Meta and Amazon Web Services with immediate access to power. A key operational catalyst is a reported $4 billion federal loan directed toward nuclear uprating. Expanding capacity at existing reactors enables Vistra to scale output through already-built switchyards and transmission lines, bypassing lengthy regulatory and construction cycles for greenfield plants. Operating as an unregulated IPP subjects Vistra to merchant electricity price fluctuations, but surging demand from tech companies seeking clean baseload power—reflected in deals such as Microsoft's agreement to restart Three Mile Island—grants merchant producers significant contracting leverage relative to tightly capped, regulated utilities. Turbine Manufacturing Discipline and Pricing Leverage (3:45) For new generation capacity that must be built, GE Vernova (GEV) controls the core equipment bottleneck through its production of heavy-duty gas and wind turbines. Spun off from General Electric, the company faces expansive utility order backlogs as power developers race to construct facilities to support rising token demand from AI labs like OpenAI and Anthropic. Rather than doubling its manufacturing capacity to chase current demand spikes and risking cyclical oversupply over a 3-to-5-year horizon, GE Vernova's leadership is exercising strict capital discipline. The company is adopting dynamic pricing models analogous to the memory chip sector, increasing equipment pricing, widening profit margins, and requiring upfront customer prepayments simply to secure manufacturing slots. This operational strategy concentrates downstream AI capital expenditure within the equipment manufacturer's order backlog while insulating the business from future cyclical downturns. Behind-the-Meter Microgrids to Bypass Interconnection Queues (5:40) When regional transmission organizations and centralized utilities cannot deliver multi-megawatt connections on acceptable schedules, data center operators are increasingly deploying behind-the-meter generation. Solaris Energy Infrastructure (SEI) functions as a turnkey provider in this sector, evolving from mobile natural gas generators into a developer and operator of integrated microgrids situated directly on-site at compute facilities. By building standalone natural gas generation adjacent to data centers, Solaris allows operators to bypass traditional transmission queues entirely and bring capacity online immediately. Driven by hyperscaler urgency to activate computational clusters without waiting years for utility interconnections, Solaris has more than doubled its revenue over the past several years, establishing microgrid infrastructure as both an interim stopgap and a lasting decentralized power model. Regulated Utilities as Inflation-Protected Defensive Allocations (7:25) For conservative portfolios seeking exposure to data center electricity demand without merchant market volatility, regulated utilities such as Xcel Energy (XEL) provide defensive stability. Xcel operates under traditional public utility frameworks across Upper Midwest and Mountain West states—notably Colorado, Minnesota, and North Dakota. These interior territories attract hyperscalers due to abundant wind and solar resources, lower land and cooling costs, and local regulators that are comparatively receptive to large-scale data center infrastructure. Xcel trades at a price-to-earnings multiple below 20 with an approximate 3% dividend yield. Because public utility commissions establish guaranteed returns on invested capital, Xcel's expanding system rate base allows it to steadily increase dividend payouts over time. Consequently, the discussion highlights regulated utilities not as rapid capital-appreciation assets, but as bond alternatives for conservative investors, offering low-beta characteristics alongside equity-based inflation protection that traditional fixed-coupon bonds lack. The presented strategy categorizes the AI energy bottleneck across four distinct risk profiles: merchant generators with existing interconnects, disciplined turbine manufacturers with pricing power, rapid behind-the-meter microgrid developers, and rate-regulated regional utilities serving data center corridors.

Source quotes

[0:00] Bottlenecks in artificial intelligence have become the defining challenge of 2026, shifting from compute hardware constraints to an even more fundamental physical barrier: electricity supply and grid transmission. [0:00] Because constructing brand-new nuclear power plants requires vast capital expenditures and years of regulatory approvals, uprating allows Vistra to expand the power output of its existing reactors using the transmission lines and switchyards already physically connected to the grid. [0:00] Because the physical interconnection of transmission lines is often the slowest link in the development chain, having operational assets already in the ground positions Vistra for a durable, multi-decade expansion. [3:45] Crucially, GE Vernova's management is not repeating the historical cyclical trap of doubling manufacturing footprint just to meet a short-term spike. [3:45] Instead, GE Vernova is handling unprecedented demand much like Micron Technology manages memory chip supply crunches: by raising prices, expanding margins, and forcing prospective buyers to provide upfront prepayments simply to secure manufacturing slots for turbines. [5:40] For data center operators unable to wait several years for utility grid interconnects, the alternative is 'behind-the-meter' on-site power generation. [5:40] When local utilities and regional transmission organizations cannot supply the requisite multi-megawatt capacity on an acceptable timeline, developers bypass the grid altogether by installing dedicated power plants directly at the data center site. [5:40] Driven by the sheer urgency of data center operators to bring compute clusters online, Solaris has more than doubled its revenue over the past few years. [7:25] While it lacks the explosive volatility of momentum tech stocks—with the hosts cautioning that investors expecting 'the next NVIDIA' should 'keep dreaming'—Xcel holds unique geographic tailwinds that the broader market often overlooks. [7:25] With a price-to-earnings multiple under 20 and a dividend yield hovering around 3%, Xcel provides a stable, low-beta anchor for a portfolio.

Original transcript

[0:00] Bottlenecks and AI have been the story of 2026 with the biggest bottleneck going forward is probably going to be not just energy but electricity. So we have four stocks for you today to play that AI bottleneck in the future. These are not necessarily the most exciting companies, but they might be extremely profitable for investors long-term. I'm going to start off. What's first on your list for investors looking at this AI bottleneck? First up, I want to talk about VISTRA. That is ticker symbol VST. There are so many investors who are excited about pre-revenue nuclear companies out there. And to that, I say, why not actually look at an actual nuclear company? Now, of course, Vistra doesn't only do nuclear power. It generates power from a variety of sources, including solar, coal, gas. I mean, it pretty much runs the gambit. It does sell energy. electricity to customers directly in some places. However, it also supplies major grid operators such as AirCot in Texas. And so Supplies actually a lot of customers from Texas and then on up north, I believe all the way up to Ohio. But this is a company – the reason I wanted to highlight it today was because – The U.S. government is reportedly going to lend this company about $4 billion. Now, why is it doing that? Because it has existing nuclear facilities, and rather than needing to build new facilities, it can actually increase the amount of electricity that existing facilities are able to produce. That's called upgrading. So it can actually upgrade existing facilities and use the grid infrastructure that it already has in place. or a much more efficient deployment of energy in my view. And you look at the nuclear plants that are potentially going to be uprated here from this government loan, one of them going directly to Meta, a Meta plant, data center is a big recipient of that another one would be AWS from Amazon. One of these data centers that could benefit from Vistra simply upgrading the nuclear plants that it already has. Yeah, this has been one of the really interesting trends. And just to be clear, Vistra, the piece that I think we're talking about here is what's called the IPP, the independent power producer. The way that a lot of these utilities work is there's a regulated side of the business and there's an unregulated side of the business. So the regulated side is the way that it's a regulated side of the business. in most states is going to be you pay your electricity bill. That rate is negotiated and it's kind of capped because they have a regulated monopoly. The independent power producer side is the unregulated side, and that's where they build a power plant. And if there's a customer who's willing to buy it, maybe on a long-term contract, great, you can make a margin on that. But they take more risk because if there's not a customer, if you're just selling into the wholesale market, sometimes you're going to be really profitable. Sometimes you're not going to be as profitable. So there's more risk in the market. But right now, I think the tailwind behind artificial intelligence and the demand is going to be what's really driving companies like Vistra. Thank you. I think you're right. Having those assets... in the ground that are already connected. That's, by the way, that's a huge part of the problem. It's not just building up power plant whether it's a gas power plant or nuclear power plant it's actually building the transmission to get it to the grid or to the data center that you're trying to function and operate. So they're going to find more and more of these opportunities where, Hey, here's something that if it's collecting dust right now, uh, Can we bring it back online? We've seen this, you know, since the early days, Three Mile Island. I think that was a Microsoft project, right? They signed a long-term contract, said, hey, can we get this back up and running? Because they're that desperate for power. So it's going to be a long-term, potentially, you know, multi-decade tailwind for companies like Vistra. Indeed. And to your point, I mean, again, we're talking about stuff that is already where it needs to be. We just need more power. And so it makes a lot of sense that it's going to benefit from that. Thank you. All right, let's talk about a company that has stuff that is not quite yet where it needs to be. That is GE Vernova. This is going to be the turbines behind... gas power plants behind wind power. This is the spinoff from GE and their energy business. So when we say that there's a backlog, we can't build enough power. A huge part of that backlog is coming from companies like GE Vernova. The interesting thing for GE is... They've been in this business for a really long time, and all of these companies have kind of said... Yeah, we got tons of demand today, but we're not going to go out and double our capacity and then be left holding the bag if this demand is short-lived, if it's only the next three, four, five years, because we're thinking about this as a multi-decade business. So the way that that's going to manifest in their operations is that we're going to see revenue grow. It's going to be a little bit more like Micron, where they're going to say, hey, you know what, we're just going to increase our prices. We're going to have a little bit of money. [5:00] better margins. We're going to make you prepay for these turbines that you need. And that's exactly what's happening in the market right now. So you talk about a power position in the market. It's what are the components that companies need that they can't get fast enough This is going to be a lot of the stuff coming from GE Vernova. Not necessarily a business that is all that sexy, but it is extremely powerful in the current market. whether we have demand for Anthropic and OpenAI, and those companies fight, as long as we're using more tokens, That's what's going to drive electricity demand, and that's ultimately going to funnel its way down to GE Vernova. Yeah, and you know, my... third stock here, the second stock that I'm bringing, but the third stock for this video that I'm bringing to the table Very similar thesis here with Solaris Energy Infrastructure. Ticker symbol S-E-I. This is essentially an on-site power generation kind of a thing. This is kind of the opposite thesis of Vistra. where we're saying we want to use the grid. This is when the grid simply can't supply what we need and we need it right now. GE Vernova is a beneficiary there, but Solaris Energy Infrastructure also a beneficiary there, running that behind the meter, kind of building out your own power plant right where you need it for what you need it for. But it's becoming much more common. It is because we need the power now and there's problems with getting connected to the grid. And so companies like these are actually benefiting. Solaris is actually more than doubled its revenue over just the last few years. And it's gone from... Being simply a natural gas-powered caravan, generation to really becoming the whole operator of that behind the meter system. And so essentially if you're a data center or that needs... power now and needs a lot of power and you can't get it from the grid, you can contact Solaris and it'll essentially turnkey your own mini power plant on site. Thank you. Yeah, a lot of a lot of these companies were kind of all playing the same theme, but there's a lot of different ways to play it because there is the need for those huge generators. There's the need for small energy production. You could look at solar wind developers battery lots of different players here are going to benefit over the next few years. Let's end with a stock that is... Just on the really simple, keep it simple. side of things and that is Excel energy. So Excel energy just a typical utility This is the bill that I pay every month. But the reason that I think Excel Energy is in a really interesting position today is where they operate. They're operating in a lot of these Midwestern states, so Colorado, Minnesota, North Dakota, where... data centers are starting to go. There's ample land, there's maybe a little bit more appetite for having You know, power plants in some of these areas, there's ample wind power, there's, you know, solar availability, there's land to build these kinds of things. So I think that is going to be just a natural tailwind long term, whether you're talking about building out distribution, transportation, transmission, transportation. The power plants themselves, whether it's going to increase, you know, the rate base that they're operating in the regulated business, all of those things. Not going to be huge tailwinds, right? This is not going to be a bloom energy. that's going to just go up like a rocket ship. I think... This is one of the better buys out there. Price earnings multiple is under $20. The yield for the stock right now, the dividend yield, is about 3%. That continues to grow over the next few years. If you're like, let's say closer to retirement, you just want, you know, consistent yield. This is the kind of company whether you're looking at the AI build out as a small tailwind or a big tailwind, I think is just going to be positioned extremely well. There's a lot of utilities like this. One of the reasons I like Excel is just it's just. In a good position geographically, because I think those middle of the country states are going to probably have more tailwinds than you think in the AI build out. But yeah, what do you think about owning just the boring old utilities, John? it's fine so long as you have correct expectations. If you're hoping that this will be the next NVIDIA, Please keep dreaming. It's not going to be that. Mm-hmm. I want... Low beta? I don't want a lot of volatility in my portfolio. I want something that I know is going to be fine and it'll pay me a little. Check every single month. Or maybe you're transitioning from some – at the stage of your investing life, Where you're saying, I'm going to take some risk off the table, but I still want to be invested in the market. I don't quite want to go the bond route. I mean, this would be something that, yeah, you should sleep very well at night. Yeah, ironically, you bring up Bond's Companies like this are potentially going to be even safer than bonds because that yield can go up, right? A bond payment, if you own a 20-year bond. that you're your yield doesn't change your your check from that bond stays the same every single year. [10:00] 有点这么厉害可以上升,如果行的,更加更多,也会更加更多。甚至是,更加更多。所有的主要是,Vistra, GE,Burnova,Solaris,XcelEnergy,您知道您的想法在那里。hello不忘了订阅我的频道我今天要用你You too谢谢大家 谢谢大家See you next time.字幕志愿者 杨茜茜好谢谢大家