America’s gas abundance still needs a delivery system
An AI campus can have financing, land and a plan for thousands of chips—and still struggle to secure the fuel needed to switch them on.
Anthropic’s IPO filing identifies natural gas supply, pipeline capacity and connections as potential obstacles to its planned data centers. Failure to secure firm fuel supply and transportation could delay a facility or cause it to be abandoned.
The distinction is between gas available somewhere in the country and gas that can reliably reach a particular power plant.
Developers are increasingly pursuing behind-the-meter generation, producing electricity at or near their facilities to reduce dependence on grid expansion. That approach brings fuel procurement, generation equipment and permitting directly into the development plan.
Our thesis is that AI’s energy opportunity increasingly centers on dependable delivery. Companies that can connect fuel, equipment and customers under workable contracts may capture value from constraints that are becoming expensive for AI developers.
GE Vernova’s results show the spending reaching suppliers
GE Vernova (NYSE: GEV) provided fresh evidence Wednesday.
Its Power segment booked $21.8 billion in third-quarter orders, up 178% organically from a year earlier, led by gas power equipment. Segment revenue increased 25% to $6.1 billion, while its EBITDA margin expanded to 21.4% from 13.4%.
Those figures show demand translating into sales and improved profitability, although the segment serves utilities and other customers alongside data centers.
The company’s combined gas equipment backlog and slot reservation agreements reached 122 gigawatts, with at least 130 GW expected by year-end.
That headline contains two different stages of business. The total comprises 64 GW of equipment backlog and 58 GW of reserved production slots. During the quarter, GE Vernova converted 13 GW of existing reservations into orders.
For investors, that conversion is worth following. Reserved manufacturing capacity establishes a potential pipeline of demand; equipment orders, deliveries and service work bring that demand closer to earnings.
Get Our Next AI Energy Market Breakdown in Your Inbox
Energy companies are moving closer to the customer
The opportunity also extends beyond selling gas into the broader market.
Chevron (NYSE: CVX) has signed a 20-year power agreement with Microsoft for a proposed West Texas facility. Project Kilby is expected to provide approximately 2.67 GW, with first power anticipated in 2028. Its final investment decision is targeted by the end of 2026, subject to remaining conditions.
That places Chevron further along the value chain, developing a system that turns gas into contracted electricity.
Williams (NYSE: WMB) is building a similar connection between its infrastructure capabilities and large power customers. In August, it confirmed completion of the first phase of its Socrates power project, with the second phase then on track for completion in the fourth quarter.
Williams also secured $5.34 billion of committed capital from Blackstone and partners in exchange for a 49% noncontrolling interest in five power projects, covering more than 2.6 GW of announced capacity.
These arrangements make construction execution, financing costs and contractual returns central to the investment case. A growing customer need creates an opportunity; delivering the infrastructure determines when it can contribute.
The contract decides who carries the energy bill
Anthropic’s disclosures add another important detail: certain arrangements pass fuel and related costs through to the company without cost caps. Higher gas prices or transportation charges could therefore increase operating costs even after a facility opens.
That creates different investment exposures across the chain.
Gas producers depend on realized selling prices, production costs and market access. Infrastructure providers depend on contract terms, utilization and construction performance. Equipment manufacturers must convert demand into profitable deliveries while expanding their own capacity.
For AI developers, more expensive electricity can raise the cost of serving customers and training models. Their ability to absorb those increases will depend partly on computing efficiency and monetization.
The strongest evidence for the energy thesis will be projects entering service at attractive contracted returns. Equipment deliveries, completed pipeline connections and dependable power production are the milestones that can turn AI’s demand for energy into a lasting earnings stream.
Disclaimer
Wall Street Access is an independent financial publisher. Wall Street Access is not a financial advisor. No company mentioned has compensated Wall Street Access for the preparation, publication or distribution of this article. This content is provided for informational and educational purposes only and is not investment advice or a recommendation to buy, sell or hold any security. Investing involves risk, including the possible loss of principal.
Prices, market capitalizations, financial figures and other market data are based on information available as of the date and time of publication and may change without notice. Information is drawn from public sources believed to be reliable but is not guaranteed to be complete or accurate. Opinions and expectations expressed are our own and may change as new information becomes available.
Readers should conduct their own due diligence and contact a licensed financial professional before making any investment decision. Past performance does not guarantee future results. Any material ownership interest or conflict of interest, where applicable, will be disclosed separately.
