The next LNG cycle will be decided by availability of transformers, not a gas price

North America intends to more than double its liquefied natural gas export capacity by 2029. The gas is there, the buyers are there, and the final investment decisions are being taken. The electrical equipment is not there. A large gas turbine ordered today arrives in about five years, around 2031. For any project that has not already reserved that equipment, the delivery slot, not the gas, now sets the date of first cargo.

The US Energy Information Administration projects export capacity rising from 11.4 billion cubic feet a day at the start of 2024 to 28.7 in 2029, most of it on the Gulf Coast. The forecast carries a condition: if projects under construction begin operations as planned. Hitachi Energy’s white paper is a careful examination of that condition.

Its central observation is that liquefaction is electrically intensive, and that LNG developers now compete for generation, grid connections, transformers, switchgear and commissioning labour with data centres and general electrification. Hitachi puts the position plainly. An industry built to export energy is bidding against the digital economy for the electricity to make it.

The independent evidence sharpens the diagnosis. US electricity demand rose about five percent between 2019 and 2024, reversing a decade of flat consumption, and the EIA forecast in January 2026 the strongest four-year demand growth since 2000. Wood Mackenzie supply chain data puts lead times for power transformers and generator step-up units at two to three years. Gas turbines are worse: the Electric Power Research Institute puts a large unit ordered today at around five years, S&P Global has reported waits of up to seven, and Wood Mackenzie expects turbine prices near 600 dollars per kilowatt by end-2027, a 195 percent increase on 2019. The Department of Energy has separately found 55 percent of in-service distribution transformers, some 40 million units, are over 33 years old, so the replacement queue competes with the growth queue.

The clearest evidence sits in a manufacturer’s order book. GE Vernova reported on 28 October a combined gas power backlog and slot reservation position of 122 gigawatts, being 64 gigawatts of equipment backlog and 58 gigawatts of slot reservations, with at least 130 expected by year end. Power orders rose 178 percent organically. In those three months it contracted 11 gigawatts of new gas equipment and shipped five. The queue is lengthening, not clearing.

None of this is a reason to be bearish on LNG. The resource, market access and industrial base are real. It is a reason to be precise about what determines whether a project hits its dates.

Hitachi’s recommendation is integrated delivery, with an equipment manufacturer or integrator owning the design basis and key interfaces. The company sells exactly that, which readers can weigh, but the point stands and the paper states it well: the critical delivery question is not who supplies each component, it is who owns the performance of the complete subsystem.

What I would add sits one level up, in the financing. Large LNG projects divide electrical scope into packages so work proceeds in parallel. Every boundary creates an interface, and an interface with no owner surfaces at commissioning, the most expensive possible moment to find one. A disputed test responsibility does not appear in a cost report. It appears as a delay to energisation, with cranes, crews and milestone payments already committed.

Project finance has not caught up. Final investment decision is taken on feed gas, offtake, liquefaction technology and marine access. The electrical scope is treated as a downstream utility package, procured after the commercial concept is fixed. Yet slippage at energisation flows into capitalised interest, delayed first cargo and debt service cover in the early operating years, where the structure is tightest.

The industry has already invented the right instrument. A slot reservation agreement is a dated, priced claim on manufacturing capacity. Lenders should diligence those the way they diligence a gas supply agreement, because that is what they now are.

Three further changes follow. Put long-lead electrical equipment reservations on the critical path in the base case model rather than in the procurement annexe. Require the independent engineer to sign off the energisation date against an actual delivery slot, with evidence. And price interconnection and curtailment risk explicitly, because a grid connection on a map is not a grid connection.

Behind-the-meter generation is not the escape route it is presented as. As Hitachi concedes, it changes the boundary of the problem rather than removing it. The project still needs fuel, permits, generating equipment, transformers, protection and control, and the same scarce electricians. Reuters reported rising Gulf Coast labour costs well before the current squeeze.

Whether North America reaches the capacity the EIA projects is now a question of equipment lead times and interface ownership. Capacity that cannot be energised on schedule is not capacity. It is capitalised interest with a flare stack attached.

About the author

Paul Gardner Brook has spent more than thirty years in global financial markets, including as a Managing Director at Nomura and a Senior Vice President in Lehman Brothers’ Asia Special Situations Group. He writes on capital markets, energy and technology at paulgardnerbrook.com.

This article is general commentary on energy markets and project finance. It is not investment, legal, taxation or accounting advice, not a recommendation, and not an offer or solicitation in respect of any security or financial instrument. Figures were current at 12 November 2026 and are sourced to the organisations named in the text. Third-party figures are reported as their sources state them.