Reputable source sees huge build of gas generation. I believe that renewables will keep gaining share, as they are in every other country
We publish interesting analysis, even if we disagree, especially if they are part of the policy discussion. Christopher Louney is a Commodity Strategist at RBC Capital Markets, a respected investment firm. He’s predicting future high gas prices based on a large increase in demand. I don’t predict commodity or stock prices. I’m picking this up because of the demand prediction.
Louney’s projection is based on a large queue of projects not yet under construction. Historically, 60-80% of projects in queue never happen. They are put in queue to reserve a place.
Projects already in construction are far more likely to be renewables. I believe the trend to solar and wind will continue. They are about the same cost as gas and far more acceptable.
Here’s the summary of the March RBC research. I confirmed with the author that his views are the same in August.
Important facts
RBC now models U.S. data-center natural gas demand at roughly 6.1 Bcf/d by 2030, versus 2024 levels — a nearly 20% increase to recent annual average powerburn. RBC frames data centers, alongside LNG export growth, as one of the two most consequential drivers of U.S. gas demand through 2030, projecting data-center load at roughly 17% of 2025’s average U.S. power consumption by decade’s end.
Geography does most of the explanatory work. Texas and Virginia host about 40% of operating, planned, and under-construction data center capacity; Georgia, Ohio, Illinois, and Arizona add another 17%. Of the top six states, all but Illinois run grids that get close to half their generation from gas — RBC treats that overlap as the reason its forecast moved, not a coincidence riding alongside it. DOE figures cited in the report put 36.6 GW of data-center capacity under construction nationally and 201.5 GW in planning.
The queue-jumping story is in the behind-the-meter numbers. Developers have announced roughly 101 GW of behind-the-meter gas generation, with 57 GW already carrying disclosed equipment orders and about 7 GW under construction. Texas alone accounts for an estimated 38 GW of that. As of March 2026, ERCOT had logged approximately 356 GW of data-center interconnection requests — the scale of queue pressure RBC cites as the reason hyperscalers are self-generating rather than waiting.
On the front-of-the-meter side, planned gas capacity queued for interconnection between 2026 and 2030 grew from 23.8 GW at the start of 2025 to roughly 64 GW by early 2026, with 16 GW added in Q1 2026 alone. New projects entering the queue in 2025 averaged 590 MW, versus 509 MW for projects already there — bigger units, not just more of them. The report’s marquee examples: the 4.4 GW Homer City plant rising on a retired Pennsylvania coal site (665,000 MMBtu/day beginning 2027) and Meta’s commitment to fund seven new gas plants in Louisiana.
Two constraints get real weight. Global gas turbine orders hit an estimated 100 GW by end-2025 against 60–70 GW of annual manufacturing capacity, pushing new-order premiums to 10–20% over backlog pricing. And community opposition is no longer a rounding error: RBC cites roughly $98 billion in projects blocked or delayed in Q2 2025, at least 25 cancellations in 2025, and 300-plus state bills addressing data-center siting.
Projections
RBC’s base case is 6.1 Bcf/d of data-center gas demand by 2030, with a stated range of 6–7 Bcf/d. It calls this secondary to planned LNG export capacity additions over the same window — gas demand growth has two drivers, and data centers are the smaller one, even after this upward revision. Beyond 2030, RBC expects zero-carbon sources to gain relative share, but argues gas keeps “a significant operational role” based on current technology and infrastructure trajectories — a hedge worth noting, not a hard forecast.
Author and credits
Written by Christopher Louney, Commodity Strategist at RBC Capital Markets. Published as a public-facing summary on May 14, 2026, drawn from RBC’s fuller report “Natural Gas Strategy: Gas Is Powered Up,” published May 3, 2026 (available to RBC clients on request).
Link
https://www.rbccm.com/en/insights/2026/05/natural-gas-powers-the-data-center-boom
Summary
The piece argues gas has become the practical default fuel for data-center power because it’s the only thing available fast enough — not because it’s cheap, clean, or even the developer’s first choice, but because interconnection queues and turbine lead times foreclose the alternatives on the timelines hyperscalers are working with. RBC frames U.S. gas’s insulation from geopolitical price shocks (contrasted with a 40%+ spike in European TTF and surging Asian LNG prices) as a second, independent reason developers favor it, on top of the geography-and-speed argument.
Where it complicates its own story: the report spends real space on headwinds that could blunt its own forecast — turbine scarcity, state large-load tariffs colliding with a still-forming federal framework, and community resistance that’s already killed or delayed tens of billions in projects. RBC doesn’t resolve the tension between “gas is structurally advantaged” and “gas developers can’t get turbines or social license fast enough” — it holds both.
Conclusions
RBC lands on a real number (6.1 Bcf/d, 6–7 Bcf/d range) it’s willing to defend, and treats the behind-the-meter build as the more important structural shift than the headline forecast itself. It declines to take a side on how binding the turbine and opposition constraints actually are — those sit as risks to the forecast, not inputs that changed it.
One thing has already moved past the report. RBC (writing in May) described FERC as expected to “unveil a regulatory proposal in June” establishing nationwide interconnection rules for large loads. That’s not what happened. On June 18, 2026, FERC issued six separate Section 206 show-cause orders, one to each RTO/ISO, directing each to justify or reform its own large-load interconnection tariff rather than imposing a single federal standard. It’s the same directional bias RBC describes — toward firm generation commitments, away from case-by-case discretion — but the mechanism is regional and tariff-specific, not the nationwide rule the report anticipated. Worth keeping straight before citing this piece on what FERC actually did.
edited and authored by Dave with close collaboration by Claude