Watt's left in the overcrowded power trade?

The AI boom has been a series of bottlenecks. First GPUs, then data center equipment, and currently its memory chips that are in short supply. Electricity demand has been on the radar for a while, but it’s likely to become a bottleneck over the next year or two.

This is the one everyone’s seen coming, and a lot of stock prices reflect that fact. But the ripple effects could affect a growing list of industries, and that’s where the uncrowded opportunities might be…

What Happened In Markets This Week

Here’s a quick summary of what’s been going on:

📈 The Fed signals rates could stay higher for longer (CNBC)

• What happened: The Federal Reserve left interest rates unchanged at 3.5% to 3.75%, but policymakers raised their median projection for 2026 to 3.8%, up from 3.4% in March. Nine of 18 officials now expect at least one rate hike by year end.
• How it impacts investors: Markets may need to adjust to the higher for longer narrative. That could weigh on interest-rate-sensitive sectors such as housing and growth stocks, while supporting financial companies that benefit from higher rates.
• Next steps: Use the Simply Wall St portfolio tool to see which of your holdings may be exposed to higher rates.

🇯🇵 Japan pushes ahead with rate hikes as inflation pressures build (BBC)

• What happened: The Bank of Japan raised its policy rate to 1.00% from 0.75% (the highest in 31 years) and reduced its bond-buying program as inflationary pressures continued to build. The move was widely expected by economists and comes as higher energy costs and a weaker yen contribute to rising prices.
• How it impacts investors: Japan’s shift away from ultra-low interest rates has been policy for a while now. But it still reinforces the view that inflation remains a global concern, which may keep interest rates elevated across major economies.
• Next steps: Explore our Japan market page to find companies that could benefit, or face challenges, as the country moves into a higher-rate environment.

💸 OpenAI’s growth comes with a US$20.92 billion price tag (Fortune)

• What happened: Leaked 2025 financials show OpenAI generated US$13.07 billion in revenue while recording a US$20.92 billion operating loss. Revenue more than tripled year over year, while sales and marketing spending surged to US$5.73 billion and research and development costs reached US$19.18 billion.
• How it impacts investors: OpenAI spent over 40% of its revenue on sales and marketing, reflecting how much these companies need to spend to defend market share and drive adoption. This is a side of the industry where Alphabet in particular, as well as SpaceX and Meta have an advantage: a large captive audience.
• Next steps: Check the U.S. Interactive Media and Services Industry on the markets page to find companies that benefit from ad spend from AI startups.

🐉 Enterprises are rethinking closed AI models after access and pricing shocks (CNBC)

• What happened: Anthropic suspended access to its Fable 5 and Mythos 5 models to comply with a US export-control directive, highlighting the risk of relying on closed AI platforms. At the same time, rising usage-based AI costs are pushing companies to evaluate downloadable open-source models that can be self-hosted and controlled internally.
• How it impacts investors: The shift toward open-source AI could create opportunities for infrastructure providers, hosting platforms, and companies enabling enterprises to run models independently. It also raises questions about the long-term pricing power and competitive advantages of closed-model providers as customers seek lower-cost alternatives.
• Next steps: Have a look at the AI Enablers and Infrastructure Software watchlist to find companies that stand to win regardless of the models being used.

🕶️ Snap doubles down on AR with its biggest hardware bet yet (Reuters)

• What happened: Snap launched Specs, its first consumer augmented reality glasses, priced at US$2,195 and scheduled to begin shipping later this year. The company is facing criticism over the US$3.5 billion spent developing the product (a drop in the ocean compared to Meta’s Reality Labs money pit.)
• How it impacts investors: Snap is making a high-risk bet that augmented reality can become a major computing platform beyond smartphones. Success could open a new growth avenue for the company, but the investment is likely to remain a drag on profitability until consumer adoption proves itself.
• Next steps: Check the Virtual and Augmented Reality Stocks watchlist for other companies betting on AR/VR.

Overloaded: The AI Power Trade Gets Crowded

Prior to the launch of ChatGPT in 2022, the ‘electrification ’ of the global economy was already a thing. Electrification requires more power generation, and it also requires more transmission via the power grid.

With the AI boom, that trend has accelerated, and it’s only just getting started. Between 2022 and 2025, the forecast increase in peak US power demand over the next five years jumped from 24 to 166 GW (gigawatts). More than half of that growth is expected to come from data centers. Reindustrialization is also contributing to growing demand from manufacturing industries.

Getting more power to data centers, and everyone else who needs it, requires more power generation and transmission. Solving this problem isn’t straightforward:

• Older power plants (coal, gas and nuclear) are being decommissioned, so new capacity needs to both replace that capacity and meet new demand.
• The grid itself is already constrained, and more so in some regions than others. Lack of investment in the past, is now being exacerbated by a shortage of transformers and lead times of up to four years.
• Extreme weather events are leading to supply interruptions and spikes in demand.
• The complex regulatory environment is creating a massive backlog of installed power waiting to be connected to the grid (2,060 gigawatts as of 2025). New projects also face multi-year permitting timelines.

When it comes to generation, renewables and nuclear come with their own downsides:

• Solar and wind projects are relatively fast to build, but don’t deliver the 24/7 baseload that most end users require.
• Nuclear is gaining recognition as the pragmatic, long-term solution, but new reactors take 7 to 10 years to build. Restarting old reactors like Three Mile Island seems to be feasible, but there aren’t many of them. Nuclear fusion and SMRs (small modular reactors) are still in the experimental phase.

That leaves natural gas as the realistic solution in the near term. However, the wait time for gas turbines is now 5 years. GE Vernova’s backlog is 110 GW, compared to the 29 GW in new generation capacity currently under construction.

Are the ‘obvious’ winners now too obvious?

The AI power theme is now well known, and some of the clear winners have racked up big gains for shareholders. A few examples over the last three years:

Utilities: Talen Energy (740%) and Constellation Energy (190%)

Grid and generation: Quanta Services (290%), Eaton (109%), GE Vernova (770% since unbundling from GE)

Datacenter power and cooling: Vertiv (1,300%) and Fix (1,120%)

It’s fair to say these are now crowded trades, and a lot of the good news is already reflected in the price. As an example, the P/E ratio for Quanta Services is approaching 100x, which isn’t something you often see in the industrial sector.

The fact that a thesis is consensus doesn’t mean there’s no upside left. In fact, the following narratives suggest there could be more upside ahead:

• Quanta: The Picks-and-Shovels Leader of the Grid Supercycle
• Vertiv: The Infrastructure AI Cannot Be Built Without
• GE Vernova will shine with a projected profit margin boost of 23.8% (For transparency I hold GEV)

If a thesis continues to play out, the valuation can continue to rise. But, keep in mind that companies producing physical goods can face limitations on what they can produce, unlike cloud and software companies.

The USA Power Infrastructure screener includes most of the leading companies involved in generating and transmitting power to data centers, and keeping them cool.

The second and third-order winners (and losers)

When a theme becomes crowded, it’s worth thinking about second and third order effects.

• First-order effects of a trend are the obvious ones:
  • Rising power demand means utilities need to generate more power, the grid needs more capacity, electricity prices will rise.
• Second-order effects are the consequences of the first-order effects:
  • Demand for natural gas increases, electricity prices rise, orderbooks for equipment grow.
• Third-order effects are the consequences of the second-order effects:
  • Copper demand rises, the grid becomes less stable, low margin industries must compete with data centers for electricity.

These ripple effects can continue beyond the third order, and can eventually affect the entire economy. Farnam Street has a more detailed explanation of second-order thinking.

In the case of the power bottleneck, many of the second-order effects have also become the consensus view, so maybe we are now looking at third and fourth-order effects.

With something as complex as electricity generation and transmission, the potential knock on effects are vast. They could produce winners and losers across the global economy. Here are just a few examples to consider:

Possible Winners

Battery Storage

Renewable energy may be out of favor (compared to natural gas), but battery storage is still needed:

• With energy grids under pressure, battery energy storage systems (BESS) become critical for managing grid stability and dealing with sudden spikes in demand.
• Data centers also need UPS and emergency backup systems.
• UK investors may want to check this small cap narrative: Invinity Energy Systems: All About That BESS.

Beyond the battery companies, the entire battery supply chain benefits too: lithium, nickel, rare earths etc.

Copper and aluminum producers and electrical cable manufacturers

Every transformer, transmission line, and data center switchgear installation is copper-intensive. Aluminum cable is also seeing increased demand. Put simply, growing AI demand cannot be met without massive amounts of copper.

Reciprocating engines and gensets

Diesel generators aren’t the most efficient means of generating power. But they are one of the most reliable ways to generate energy when other sources fail.

If a data center wants to ensure continuous uptime, it needs either a battery solution, or a generator, or both. Caterpillar and Cummins have both seen sharp increases in orders for these engines.

GOES: The transformer bottleneck

Grain-oriented electrical steel (GOES) is amongst the most critical inputs in transformer production. Cleveland-Cliffs operates the only US facilities capable of producing the material.

Real estate close to abundant power sources

Regions with cheap, abundant power are attracting the bulk of new data center investment. Industrial REIT landowners and real estate developers in these corridors benefit disproportionately.

Regions with abundant energy

AI data centers should ideally be located close to the users. But, as electricity becomes more difficult to source, some compute can be handed off to data centers elsewhere.

Possible Losers

Energy-intensive industries

The aluminum, steel, chemical, and cement industries (which operate with thin margins) are now competing with data centers (with wider margins) in the market for power. This is an advantage for other countries, but complicates America’s reshoring efforts.

The construction industry

Electricians are in demand in the US, both by the construction industry, and the AI and power infrastructure industry. It seems likely that AI will outbid the construction industry on this one.

Smaller data center operators

Suppliers naturally prioritize their largest customers. When backlogs are running at four to five years, small operators are going to struggle to get the equipment they need.

💡 The Insight: Think Differently To Avoid The Crowd

Finding an investment opportunity before the crowd can be very rewarding. If you get it right, you get rewarded twice; first, as the market catches on, and again if the thesis continues to play out.

By definition, you are unlikely to find these opportunities from CNBC or Wall Street analysts. You need to research the problems that need to be solved, the relevant industries, and then do some creative thinking.

• AI chatbots can be very useful for this type of brainstorming. Ask an initial question, and then keep probing for deeper answers.
• Research reports written for industry professionals rather than for investors can help you understand the underlying problems and potential solutions. As an example, if you want to get into the weeds on the US grid, Grid Strategies is very comprehensive.
• Check the community page for narratives written by members of the Simply Wall St community.
• Search for companies that might not pop up elsewhere by using the keyword search function on the Simply Wall St screener.

Key Events Next Week

Monday

• 🇨🇦 Canada CPI YoY
  • 📈 Forecast: 2.9%, Previous: 2.8%
  • ➡️ Why it matters: Oil prices stayed elevated in May; a headline beat pushes BoC rate cut bets further out.

Wednesday

• 🇦🇺 Australia CPI YoY
  • 📈 Forecast: 4.9%, Previous: 4.2%
  • ➡️ Why it matters: RBA forecasts CPI peaking at 4.8% mid-2026; a beat keeps another rate hike firmly in play.

Thursday

• 🇦🇺 Australia Unemployment Rate
  • 📉 Forecast: 4.4%, Previous: 4.5%
  • ➡️ Why it matters: April's jump partly reflected a statistical distortion; a bounce steadies the RBA's hand.
• 🇺🇸 US Core PCE Price Index YoY
  • 📈 Forecast: 3.3%, Previous: 3.3%
  • ➡️ Why it matters: The Fed's preferred inflation gauge — any further tick up supports the Fed’s hawkish shift..

Friday

• 🇯🇵 Japan Tokyo CPI YoY
  • 📈 Forecast: Previous: 1.4%
  • ➡️ Why it matters: Iran oil shock filtering through; a pickup adds urgency to BoJ rate hike expectations.

Second quarter earnings season of FY26 kicks off in two weeks time. In the meantime, there are a few large caps due to report. This week it’s:

Micron

FedEx

Paychex

KB Home

Disclosure

Simply Wall St analyst Richard Bowman has positions in NYSE:GEV, ARCA:URA, BATS:PAVE. Simply Wall St have no position in any of the companies mentioned. This article is general in nature. Any comments below from SWS employees are their opinions only, should not be taken as financial advice and may not represent the views of Simply Wall St. Unless otherwise advised, SWS employees providing commentary do not own a position in any company mentioned in the article or in their comments.We provide analysis based on historical data and analyst forecasts only using an unbiased methodology and our articles are not intended to be financial advice. It does not constitute a recommendation to buy or sell any stock and does not take account of your objectives, or your financial situation. We aim to bring you long-term focused analysis driven by fundamental data. Note that our analysis may not factor in the latest price-sensitive company announcements or qualitative material.

Have feedback on this article? Concerned about the content? Get in touch with us directly. Alternatively, email editorial-team@simplywallst.com