What happens when the world has powerful AI chips ready to go, but not enough electricity to run them? That question is becoming increasingly important for investors. The AI boom is no longer just about who builds the fastest processors or develops the most advanced models. It also depends on who can secure reliable power, connect to the grid, and bring computing facilities online before competitors do. That is helping put regions such as Finland and Norway on the radar, while creating potential opportunities for companies across the infrastructure chain.

Nvidia (NVDA) and AMD (AMD) supply essential computing hardware, while CoreWeave (CRWV), Crusoe, Applied Digital (APLD), IREN (IREN), and Bitzero represent different parts of the expanding infrastructure ecosystem. But a promising AI connection does not automatically make a stock a good investment. Planned capacity still needs to become operational, customers must pay enough to support the costs, and financing can make or break a project's returns. The real opportunity may lie with businesses that can turn scarce electricity and infrastructure into dependable revenue—not simply announce the biggest projects. ⚡

What if the next AI winner isn't the company making the chips, but the one that can actually turn the power on? 👀

We’ll explore why Nordic data centers are attracting attention, how several infrastructure companies are positioning themselves, and the crucial differences between announced capacity, signed contracts, and a business that can consistently generate cash.

Imagine setting aside $500 a month for $ENLT ( ▲ 0.45% ) month after month, buying a fixed dollar amount on a schedule, a method called dollar-cost averaging. In our example the buy happens on the first trading day on or after the 9th of each month, from March 2023, ENLT’s first full month on Nasdaq, through September 2026: 43 buys, $21,500 in total.

ENLT’s first full Nasdaq close was $17.75 on Feb 10, 2023, and it closed at $67.41 on Oct 9, 2026, a price gain of about 280%, or roughly 44% a year compounded over those 3.7 years (the steady yearly rate that would give the same total gain). In this example the 43 buys would have been worth about $67,500 at the Oct 9, 2026 close, roughly 3.1 times the $21,500 put in, and a single $10,000 invested on Feb 10, 2023 would have become about $38,000. Price only: no dividends (ENLT pays none), fees or taxes, and these are examples, not forecasts. (Figures use ENLT’s Nasdaq closes from Yahoo Finance.)

The highest close of the past 52 weeks was $107.42 on May 29, 2026, so the Oct 9, 2026 close sat about 37% below it. The lowest close of those 52 weeks was $32.82 on Oct 21, 2025. The lowest close since the US listing was $12.19 on Oct 26, 2023. To keep the headline in perspective, about 3.1 times the money since March 2023 is a strong result but still sensitive to near-term trading moves.

Caution: past pace rarely continues. ENLT’s release for the quarter ended June 30, 2026 showed total revenues and income up 55% and adjusted EBITDA (operating profit before interest, taxes, depreciation and amortization, with special items left out) up 67%, with a raised outlook (its own forecast). The stock has since moved well below its 52-week high, a reminder that prices can pause even when results are strong.

The artificial intelligence boom has created an unusual problem for the technology industry: companies can spend billions on advanced chips and data centers, yet still struggle to find enough electricity to put that equipment to work.

For investors, this changes the way the AI opportunity should be evaluated. The next phase of growth is not simply about which company develops the most powerful chip or which cloud provider attracts the most customers. It is also about who can secure electricity, obtain permits, connect to the grid, and deliver usable computing capacity without years of delays.

That challenge is pushing technology companies to look beyond their traditional U.S. data center locations. Finland and Norway are attracting growing attention because they combine established power infrastructure, relatively abundant electricity in certain regions, and experience hosting energy-intensive facilities.

But there is an important distinction between identifying a promising location and identifying a winning investment. The companies that benefit most may not necessarily be the ones making the biggest announcements. They may be the ones that can turn land, electricity, permits, and infrastructure into revenue-generating assets on schedule.

For years, the AI investment story centered on processors, model development, and cloud computing. Those remain essential, but the physical requirements of AI are becoming harder to ignore.

Advanced chips consume substantial amounts of electricity, and the servers running them generate significant heat. A large AI data center therefore needs more than a building and a collection of GPUs. It needs a dependable power supply, electrical connections capable of supporting its load, cooling systems designed for high-density computing, and the regulatory approvals required to operate.

In the United States, the rapid expansion of data centers has collided with the time required to expand electricity infrastructure. New transmission lines, substations, generation capacity, and grid connections cannot always be delivered as quickly as technology companies want to install their equipment. Local opposition can add another obstacle, particularly when residents worry about electricity prices, water consumption, noise, or pressure on existing infrastructure.

That creates a potential mismatch: computing equipment may be available, customer demand may be strong, and financing may be in place, but a project can still be delayed because the electricity is not ready.

For an investor, this is where the economics become interesting. A data center that cannot access sufficient power cannot generate its intended computing revenue, regardless of how expensive its equipment is. Conversely, a site with secured power and completed infrastructure can become more valuable when competing projects face delays.

The opportunity, however, depends on whether that advantage translates into profitable contracts rather than simply a higher valuation for the underlying property.

Finland and Norway offer several advantages that matter to energy-intensive computing operations.

Both countries have established electricity infrastructure, and Norway generates a substantial share of its electricity from hydropower. Finland has a diverse generation mix that includes nuclear power. Certain regions also have favorable access to existing power connections and generation facilities.

That does not mean every site has unlimited electricity or that every project can connect immediately. Available capacity varies by location, and new facilities still need to satisfy permitting, grid-connection, construction, and operational requirements.

The attraction is the possibility of developing data centers where the supporting infrastructure already exists instead of relying entirely on new power generation and transmission projects.

The Nordic region also has a cooler climate, which can help reduce some cooling-related energy requirements. However, the savings should not be overstated. Modern AI facilities increasingly use liquid cooling, including direct-to-chip systems, to manage the heat produced by high-performance processors. These systems are designed around the equipment's requirements, so a colder climate does not eliminate the need for sophisticated cooling infrastructure.

The more important advantage is the combination of power availability, established infrastructure, and the ability to develop projects within a relatively predictable regulatory framework.

The underlying thesis is straightforward: when demand for AI computing grows faster than electricity infrastructure can be expanded in one market, companies have an incentive to diversify where they build.

That does not mean the United States is losing its importance. The country remains central to the AI industry, with major technology companies, research organizations, cloud platforms, and customers concentrated there. Nordic facilities can complement U.S. infrastructure by serving computing demand from companies operating across borders while additional power capacity is developed elsewhere.

For investors, geographic diversification within the AI infrastructure supply chain may become increasingly important. A company with access to electricity in a constrained market can offer a different advantage from a company that simply owns servers.

Two companies remain particularly important to the computing side of this equation: Nvidia and Advanced Micro Devices, or AMD.

Nvidia $NVDA ( ▼ 0.52% ) supplies GPUs and related computing infrastructure used extensively in AI data centers. AMD also develops processors and accelerators competing for workloads in the growing AI computing market.

Their products influence how quickly new computing capacity can be deployed, how much performance customers receive, and how efficiently facilities use available resources.

But chip production is only one part of the supply chain. Once the equipment arrives, operators still need a site with adequate electricity, appropriate cooling, network connectivity, and sufficient infrastructure to bring the systems online.

This creates a chain of dependencies. Chip availability affects deployment schedules, electricity availability constrains how much equipment can operate, and customer demand determines whether the resulting capacity earns an attractive return.

If Nvidia and AMD $AMD ( ▼ 2.03% ) increase shipments rapidly, more facilities may be needed. Yet if the industry builds too much data center capacity relative to actual demand, operators could eventually face lower pricing or weaker utilization. If chip supply remains constrained, meanwhile, some newly built facilities could struggle to fill their available capacity.

Neither outcome is guaranteed.

The key is to avoid assuming that rising demand for AI automatically benefits every company connected to the industry. Each business occupies a different position in the chain, with its own capital requirements, competitive pressures, and execution risks.

Bitzero $AIBZ ( ▲ 3.56% ) provides a useful example of the less visible work involved in preparing a site for AI computing.

The company began in 2021 and initially used its infrastructure for Bitcoin mining. Its strategy involved acquiring land and developing supporting electrical infrastructure, including high-voltage connections, rather than relying entirely on leased capacity.

That distinction matters because preparing a site for large-scale computing can require years of work before a data center is ready to serve customers. Land acquisition is only the beginning. Developers may need to secure permits, design electrical systems, build connections, coordinate with utilities, and install equipment with long lead times.

Bitzero has said it is developing company-owned sites in Norway and Finland and plans to add more capacity there over the next few years.

These plans should be treated as company-stated development targets, not as proof that the capacity has already been delivered, contracted, or converted into revenue. A project's potential capacity is different from operating capacity, and operating capacity is different from profitable utilization.

That distinction is essential when evaluating infrastructure businesses. A company can own valuable land and have a promising power position while still facing substantial construction costs, financing requirements, customer negotiations, and execution risks.

Bitzero's earlier Bitcoin mining operations also illustrate how infrastructure developed for one use can potentially be repurposed for another. Bitcoin mining and AI computing both require substantial electricity, but an existing mining site is not automatically suitable for advanced AI workloads. High-density GPU deployments can demand different cooling systems, building specifications, networking, reliability standards, and customer arrangements.

The advantage comes from what can genuinely be reused: secured power, land, grid connections, permits, and completed electrical infrastructure. The remaining conversion costs still need to be accounted for.

The Nordic expansion is not only about traditional technology giants constructing their own facilities. Specialized computing providers also play a role.

CoreWeave $CRWV ( ▲ 0.61% ) operates cloud infrastructure designed for demanding computing workloads, including AI. Crusoe has developed infrastructure connected to energy-intensive computing and AI data centers. Both companies are relevant to the broader shift toward specialized computing capacity.

Their presence in the region, illustrates how demand can reach beyond the largest established cloud platforms. AI companies need access to computing resources, but they do not all need to own every facility or piece of infrastructure themselves.

This creates opportunities for operators that can provide computing capacity under commercial agreements. It also introduces another layer of risk: a facility's economics can depend heavily on the financial strength of its customers, the terms of its contracts, and the cost of delivering the contracted service.

For investors, a reported customer relationship is only the beginning of the analysis. The more important questions are how much capacity is committed, when the contract begins, how long it lasts, what pricing and payment terms apply, and whether the operator can earn an acceptable return after electricity, financing, maintenance, and equipment expenses.

A recognizable customer can improve confidence in a project, but it does not eliminate execution or counterparty risk.

The transition from energy-intensive computing toward AI infrastructure is also relevant to Applied Digital (APLD) and IREN $IREN ( ▼ 1.46% ).

Applied Digital develops data center infrastructure designed to support high-performance computing and AI workloads. IREN, formerly known as Iris Energy, has roots in Bitcoin mining and has been expanding its focus on AI cloud and data center infrastructure.

These companies help illustrate why investors are paying attention to businesses that already possess some combination of land, electrical capacity, facilities, and experience operating power-intensive computing sites.

An existing infrastructure base can potentially shorten the path to serving AI customers compared with a project starting from scratch. But the size of that advantage depends on the condition of the assets, the availability of power, the required upgrades, and the company's ability to secure financing and customers.

The same caution applies to Bitzero. A Bitcoin mining background may provide relevant experience with electricity procurement and facility operations, but AI infrastructure is not simply Bitcoin mining with different machines installed. The technical specifications, customer expectations, capital requirements, and operating economics can be substantially different.

The investment case should therefore be based on demonstrated execution and contracted economics, not just the promise of a business transition.

One of the least visible parts of the AI infrastructure boom may also be one of the most important: regulatory preparation.

Investors naturally focus on announcements involving billions of dollars, hundreds of megawatts, and major technology customers. Yet those announcements do not necessarily reveal how much work remains before a project becomes operational.

Permits, environmental reviews, electrical connections, construction planning, and equipment procurement can take years. A developer that has already completed significant portions of this work may have an advantage over a competitor whose proposed capacity exists mainly on paper.

Bitzero has said it holds a facility license for high-voltage lines serving its own Norwegian site. That is an example of the kind of preparation that can matter when electricity access is a major constraint.

Still, permits and infrastructure do not automatically guarantee commercial success. Projects can encounter cost overruns, changing customer requirements, delays, and financing challenges even after significant regulatory progress.

The strongest position is one in which several pieces align: the site has the necessary approvals, electricity can be delivered on schedule, construction is progressing, financing is available, and credible customers are prepared to use the capacity under commercially attractive terms.

That combination is much more meaningful than a headline announcing a large theoretical development pipeline.

The AI infrastructure opportunity also carries a less comfortable possibility: too much capacity could eventually be built.

Today's electricity constraints and computing demand do not guarantee that shortages will persist indefinitely. Developers are investing in new facilities, utilities are planning additional generation and transmission capacity, and chip manufacturers are working to increase supply. At the same time, AI models, applications, and workloads are evolving quickly, making future demand difficult to forecast.

If new computing capacity grows faster than paying customer demand, operators could face lower prices, weaker utilization, or delays in filling facilities. Projects that appeared attractive when capacity was scarce may produce disappointing returns if competition increases.

The opposite is possible, too. If AI adoption continues expanding and power infrastructure remains slow to develop, access to ready-to-use capacity could remain valuable.

That uncertainty is why the industry deserves a more careful approach than treating every stock associated with AI the same way.

A large project pipeline is not the same as contracted revenue. Contracted revenue is not the same as profit. And profit is not the same as an attractive stock return if the market price already assumes years of exceptional growth.

For an overwhelmed investor, following every data center announcement can become a full-time job. A more useful approach is to focus on a small set of indicators that reveal whether the business is turning its infrastructure into economic value.

First, look for secured power rather than theoretical capacity. A site may have enough land for a large facility, but its development prospects depend on whether the necessary electricity and grid connections can actually be delivered.

Second, separate planned capacity from operational capacity. Announced megawatts, construction targets, commissioned facilities, and capacity generating revenue are different stages of development. Understanding those distinctions helps prevent an ambitious pipeline from being mistaken for current business performance.

Third, examine customer contracts. The identity of a customer matters, but so do the contract duration, pricing, payment obligations, deployment schedule, and terms governing cancellation or delays.

Fourth, pay attention to financing and capital intensity. Data centers and power infrastructure require substantial upfront investment. Debt, construction expenses, equipment commitments, and potential shareholder dilution can materially affect the returns available to existing investors.

Finally, consider valuation. Even a company with useful infrastructure and genuine demand can be a poor investment if its stock price already assumes flawless execution. The opportunity lies in finding a gap between what the market expects and what the business can realistically deliver.

Tip: When evaluating AI infrastructure stocks, start with the electricity, the permits, the construction timeline, and the contracts. The most exciting technology story is not always the strongest investment, but a company that can reliably deliver scarce computing capacity may possess an advantage that is difficult for competitors to reproduce quickly.

The AI boom is increasingly a story about physical infrastructure, not just software and chips. Nvidia and AMD influence the availability of computing hardware, while companies such as CoreWeave and Crusoe help deliver specialized computing capacity. Applied Digital, IREN, and Bitzero illustrate different ways businesses are trying to turn land, power access, and data center infrastructure into opportunities tied to AI demand.

Finland and Norway are attracting attention because parts of the region offer established electricity infrastructure, relatively favorable power conditions, and a potential alternative to constrained locations elsewhere. That does not make every Nordic project a winner, nor does it remove the risks of construction delays, oversupply, financing pressure, or changing demand.

For your portfolio, the central lesson is to look beyond the AI label. The next phase of the industry may reward companies that can turn electricity and infrastructure into reliable, contracted computing services. But the difference between a promising asset and a profitable investment will still come down to execution, cash flow, balance-sheet strength, and the price paid for the opportunity.

The companies that ultimately stand out may not be those promising the most capacity. They may be those that deliver the right capacity, at the right time, for customers willing to pay enough to make the investment worthwhile.

Want a simple next step for the $500 habit?

Our Fast Track Blueprint walks through a trusted online broker setup and a clear portfolio starting path.

Thanks for reading. This format is built to be fast to open, clear to understand, and useful enough to act on, without pretending past returns continue forever.

Caution: Past pace rarely continues. All figures here are approximate and are shown as examples, using price only (no dividends, fees, or taxes). Past performance is not a forecast; this is education, not advice.

Disclaimer: This newsletter is for informational purposes only and is not financial advice. Consult a qualified advisor before investing.