The AI Power Trade That Has Nothing to Do With Chips

The most consequential question in AI investing right now is not which chipmaker wins the next benchmark. It is who builds the power that lets any chip run at scale. Venture Global answered part of that question on October 15, 2026, and the answer points directly to GE Vernova and Baker Hughes.

Venture Global announced it has entered into power and grid frame agreements with GE Vernova and Baker Hughes Energy Services for up to 16 new, 720 MW power island units and associated power transformer equipment, supporting the continued expansion of its existing behind-the-meter, Island-Mode industrial power network that is not reliant on the grid. These agreements include initial reservation deals and pre-negotiated purchase order terms, with potential aggregate spending of about $17.2 billion if all options are exercised.

Why This Is Different From Every Other AI Power Story

The usual power shortage conversation focuses on grid congestion, permitting delays, and utility queues stretching years into the future. Venture Global sidesteps all of it. The agreements provide a pathway to make up to 10 GW of compute capacity available over time at sites owned or controlled by the company, with the expansion placing no additional demand on the public electric grid or public water utilities.

This new capacity would build on the company’s existing 4.5 gigawatts of Island Mode modular, gas-fired generation that has and will continue to supply power to its facilities. Venture Global is not starting from zero. It is extending a proven industrial model into a new, higher-margin market. The October 15, 2026 announcement creates a framework for equipment procurement, with future orders tied to commercial needs and customer commitments, and Venture Global expects its first power-island purchase order by the end of 2026.

The Investment Case for GEV and BKR

Here is where the opportunity gets specific. Rather than picking a data center operator or an AI hyperscaler, investors can own the companies supplying the physical infrastructure every one of those players depends on, regardless of who ultimately wins the AI compute race.

GE Vernova’s numbers already show the momentum. In the second quarter of 2026, GEV reported orders of $24.2 billion, up 88% organically, with gas power equipment backlog and slot reservation agreements growing from 100 to 116 GW. Data center orders reached over $5 billion year-to-date, more than double the company’s 2025 total. The Venture Global frame agreement adds a long-duration layer on top of a backlog that was already expanding rapidly.

Baker Hughes tells a similar story from a different angle. IET orders reached $7.1 billion in the second quarter of 2026, more than doubling year-over-year, while the IET remaining performance obligation climbed to an all-time high of $37.1 billion. Baker Hughes is pivoting toward a diversified energy and industrial technology profile, with the IET segment serving as a primary growth driver through LNG infrastructure, power generation equipment, and industrial applications.

The equipment demand running through Baker Hughes and GE Vernova does not stop at the turbine. Vertiv’s power and cooling backlog shows how deep the physical infrastructure spending cycle runs inside every data center these turbines will feed.

Building Wealth Around This Idea

The structural case here is that AI’s electricity problem is not a short-term bottleneck. It compounds over years. Deliveries under the Venture Global power frame agreement have scheduled delivery dates running through 2034, which means GE Vernova and Baker Hughes have a visible, multi-year workload tied directly to the AI infrastructure build-out, secured well before a single data center tenant has been named.

For long-term portfolios, both names offer exposure to the same underlying demand surge that drives Nvidia headlines, but with a different risk profile. Equipment manufacturers operating on multi-year backlogs carry far less earnings volatility than chipmakers exposed to product cycle timing. A position split between GEV and BKR gives a portfolio the gas turbine side and the LNG compression and services side of the same trade, diversified across two strong balance sheets.

Risks to Monitor

The frame agreements are not firm purchase orders. Venture Global has disclosed that the agreements include pre-negotiated purchase order terms, and that it expects to place its first power-island purchase order by the end of 2026, but it has not disclosed a full equipment price list or a named data center customer. If AI capital spending contracts sharply, Venture Global may defer reservations, and the $17.2 billion figure stays theoretical. Regulatory approval for permitting, which Venture Global said it plans to begin pursuing immediately, adds another variable. Rising input costs could also pressure margins as the manufacturing schedule scales toward 2028 and 2029.

Wealth Builder Takeaway

When a theme as large as AI power demand produces a concrete, named, multi-year equipment framework, the best entry point is often the supplier, not the headline company. GE Vernova and Baker Hughes are not bystanders to the AI infrastructure race. They are building the engines that make it run. That is the kind of durable positioning wealth is built on.