Utilities face a rapidly shifting demand outlook as artificial intelligence and data-center buildouts strain existing power systems. Industry observers expect the sector to invest up to $240 billion in 2026 to meet AI-driven electricity needs, a scale of spending that is intensifying scrutiny of how costs will be recovered through regulated rates.
At the same time, higher power demand is pushing up electricity costs, contributing to pushback from parts of the AI value chain and raising questions for investors about which companies can monetize load growth—either through traditional grid expansion or through supply models that sit outside the regulated framework.
Key takeaways
- Price move: The article cites strong recent momentum in several AI-linked power names, including a roughly 35% rise over the past year for Brookfield Renewable Partners and about 25% for NextEra Energy.
- Catalyst: Expected utility capex of up to $240 billion in 2026 tied to AI power demand is the central driver behind renewed investor focus on power infrastructure.
- Key implication: Investors may prefer business models with contracts or supply outside regulated rate-setting, where cost recovery risk can be lower.
- Valuation matters: The article highlights that Bloom Energy’s valuation is already rich despite a long backlog and limited profit track record.
What drove the AI power investment thesis
According to the article, electricity demand increased 10% between 2005 and 2025 and is projected to rise by 60% between 2025 and 2045. While multiple factors contribute to the forecast, artificial intelligence and the data centers that support AI workloads are described as key drivers of the step-change in consumption.
The buildout underway for AI computing is translating into higher power requirements, which in turn is forcing utilities to expand generation, transmission, and grid capacity. The article frames this as a direct growth catalyst for regulated electric utilities, but it also flags the practical bottleneck: regulatory approval for capital spending and the timing of rate recovery.
Why regulated utilities face cost-recovery pressure
The article’s core risk assessment centers on the rate-setting process. Regulated utilities typically pass investment costs to consumers through rate increases that require approval by regulators. With inflation high and electricity prices already moving higher, the article says there has been pushback against AI-driven data-center expansion—introducing uncertainty over whether utilities can fully recover capital expenditures.
For investors, that creates a split between (1) grid-focused companies where earnings depend partly on regulatory outcomes and (2) power providers that can sell electricity under long-term arrangements or deliver power via infrastructure that is less dependent on traditional rate base mechanics.
How investors are weighing select company models
The article points to three company approaches to the AI power demand theme.
Bloom Energy is presented as a play on dedicated power supply for data centers using hydrogen fuel power cells. The article says Bloom began 2026 with a $6 billion backlog for its fuel cells, up 2.5x year over year, and notes that total backlog stands at $20 billion when including services. It also emphasizes that each new fuel cell is paired with a service contract, implying repeatable revenue support.
However, the article highlights a valuation challenge: Bloom Energy shares have risen more than 1,000% over the past year, and it points to a price-to-sales ratio of 29x. With the company described as not yet producing sustainable profits, the article argues that conventional earnings-based valuation may be less informative for investors assessing forward returns.
Brookfield Renewable Partners is framed as a dividend-oriented alternative. The article describes a globally diversified portfolio across solar, wind, hydroelectric, storage, and nuclear, with power sold to other companies under long-term contracts. It also says AI demand is already embedded in the demand outlook, citing contracts to sell power to Microsoft and Google.
On return characteristics, the article cites a distribution yield of roughly 4.5% and a goal of increasing distributions by 5% to 9% annually. It contrasts the valuation profile with a price-to-sales ratio of 1.6x, and notes that units are up around 35% over the past year.
NextEra Energy is offered as a “middle ground.” The article describes NextEra as one of the world’s largest utilities with a large and growing renewables business that sells power under long-term contracts, similar to Brookfield. It states NextEra shares are up 25% over the past year and cites a 22.5x price-to-earnings ratio, below its five-year average of 27x. The article also describes a valuation reference of 6.6x for price-to-sales.
For forward growth, the article points to NextEra’s proposed acquisition of Dominion Energy, noting Dominion’s presence in a major data-center market in Virginia. It says NextEra expects earnings growth of around 9% per year after the acquisition and dividend growth of around 6% per year, with yield around 2.8%.
Bigger picture: what to watch as AI demand meets the grid
The article’s investment message is that AI will likely continue to increase demand for electricity, but the path from demand to shareholder returns depends on how power is supplied and how quickly costs can be recovered. For utilities, that hinges on regulatory approvals and the political and consumer response to rate increases. For contract-based power providers and companies supplying dedicated capacity, monetization may be more closely tied to pre-agreed terms rather than ongoing rate cases.
Investors will likely watch for signals on regulatory timelines, electricity demand forecasts from utility regulators and grid operators, and progress on major AI-heavy data-center deployments. In the near term, company disclosures—especially guidance updates and any milestones tied to capacity additions or acquisitions—may determine whether the market’s emphasis on AI power growth continues to translate into earnings visibility.
