Old jet engines are being scavenged to power data centers, The New York Times reported on 3 November, in a report that opens among the retired aircraft parked in the “boneyards” of Arizona’s Sonoran desert. Techmeme’s summary of the report says US data center operators and power developers are repurposing jet engines to secure power in as little as two years, bypassing multi-year grid-connection delays.
GE Vernova’s latest quarterly filing sizes the demand for gas turbines built from jet engines, which the industry calls aeroderivatives. Customers ordered 57 of them in the three months to 30 September, up from 9 a year earlier, while GE Vernova sold 13. That is 4.4 units ordered for every one sold, by our arithmetic.
Business Pill · Grid Cost Allocation
An evergreen Business Engineer explainer on grid cost allocation: who pays when new power lines and capacity are built for large loads such as data centres. It explains the general idea only and does not discuss GE Vernova, Crusoe or jet engines.
The key insight: By our arithmetic, GE Vernova’s customers ordered 4.4 aeroderivative turbines for every one it sold in the third quarter (57 against 13), up from 9 orders a year earlier, as The New York Times reports that old jet engines are being scavenged to power data centers.
GE Vernova’s 10-Q, filed on 28 October, reports orders and sales in units for its Power segment. Aeroderivative orders were 57 in the third quarter against 9 a year earlier, and 127 in the first nine months of 2026 against 45. Aeroderivative sales were 13 in the quarter against 6, and 39 in the nine months against 16.
Orders ran ahead of sales in both periods: 4.4 orders per unit sold in the quarter and 3.3 in the nine months, by our arithmetic (57 divided by 13, and 127 divided by 39). The filing does not say who placed the orders or how many are for data centers.
The filing says the Power segment’s revenue rose $1.2 billion, or 25%, organically in the quarter, “primarily at Gas Power equipment due to higher Aeroderivative and Heavy- Duty Gas Turbine deliveries and favorable pricing.” Over nine months, it says, customer down payments on orders and slot reservation agreements at Power helped bring in $19.5 billion of contract liabilities and deferred income, a cash flow we covered in Down Payments Gave GE Vernova $19.5B of Cash in Nine Months.
Aeroderivative gas turbines (turbines derived from jet engines): orders and sales in units, from GE Vernova’s 10-Q for the quarter to 30 September 2026. The ratios of orders to sales are our arithmetic.
What an AI Builder Has Ordered
Crusoe, an AI data center builder that says it builds and operates power plants alongside its data centers, has had several turbine orders announced by its suppliers. In July 2025 GE Vernova announced a 29-unit deal for its LM2500XPRESS aeroderivative packages for Crusoe AI data centers: 10 units ordered in December 2024 and 19 booked in June 2025, which combined are “expected to provide nearly 1GW of electricity,” GE Vernova said.
GE Vernova’s release says these aero units provide “the flexibility and reliability of modern, efficient jet engines used by airlines around the world thousands of times per day.” A GE Vernova executive put a single aero unit at about 35 megawatts in the same release.
In April 2026 ProEnergy, based in Sedalia, Missouri, announced a contract to supply Crusoe with 13 PE6000 aeroderivative gas turbine generator sets, each capable of generating 50 MW, for upcoming hyperscale datacenter projects. That is 650 MW, by our arithmetic. ProEnergy says the PE6000’s parts are made to be interchangeable with the LM6000.
A third supplier is no longer launching with Crusoe. In December 2025 Boom Supersonic announced a 1.21 gigawatt order from Crusoe for 29 Superpower units, a 42 megawatt turbine built on the core of the Symphony engine Boom is developing for its airliner. On 25 September 2026 Boom’s chief executive, Blake Scholl, wrote on X that “Crusoe and Boom are no longer launching our Superpower turbine together,” and that Boom would deliver about 250MW of the turbines to other sites in 2027.
Lawrence Berkeley National Laboratory tracks how long new power takes to reach the grid. Its Queued Up report, 2026 edition, says the median power project built in 2025 took 61 months from the interconnection request to commercial operations, compared with 36 months in 2015 and 22 months in 2008.
The report counts power plants waiting to connect to the transmission grid, not data centers waiting for supply. It also notes that the time from interconnection agreement to operation has fallen for gas plants, which now have the fastest timeline across resource types.
As we read the filing, the gap between orders and sales is the signal: units booked in the quarter outnumbered units shipped by more than four to one, and the 10-Q does not say how long the rest will take to deliver.
As we read the Crusoe orders, the buyer did not rely on one supplier: it split its jet-derived turbines across GE Vernova and ProEnergy, and Boom, a newer supplier, said in September 2026 that the two were no longer launching its turbine together.
As we read Berkeley Lab’s numbers, the comparison the turbines face is a median of 61 months for a power project to go from grid request to operation; a turbine on site sits outside that queue and runs on gas burned at the site.
GE Vernova, July 2025
“These aero units provide the flexibility and reliability of modern, efficient jet engines used by airlines around the world thousands of times per day.”
Three Implications
FOR GE VERNOVA Its 10-Q shows 57 aeroderivative orders in the quarter against 13 sold, and down payments on orders and slot reservation agreements at Power.
FOR AI BUILDERS Crusoe’s announced orders, nearly 1GW from GE Vernova and 650 MW from ProEnergy by our arithmetic, show the scale one builder has bought.
FOR NEWER SUPPLIERS Boom says it will deliver about 250MW of its Superpower turbines to other sites in 2027, after saying it and Crusoe were no longer launching the turbine together.
The framework states: “The structural bottleneck isn’t computing, it’s electrons.”
As we read it, the turbine orders are builders paying for electrons on their own sites rather than waiting for them from the grid. The framework was published in December 2025 and does not discuss GE Vernova, Crusoe or jet engines.
What Is Not Established
We read The New York Times report only as far as its headline, byline, opening sentence and Techmeme’s summary; the rest is behind its paywall. The two-year figure is from that summary.
GE Vernova’s 10-Q does not name aeroderivative customers or split orders by end market, so the share of the 57 orders bound for data centers is not disclosed. The ratio of orders to sales is ours and is not a measure of backlog or delivery times, which the filing does not give for aeroderivatives. Crusoe’s orders are those its suppliers announced; Crusoe may hold others. We did not contact GE Vernova, Crusoe, ProEnergy or Boom.
This Week In The AI Supercycle
Anthropic’s Reality Check, The Capital Squeeze, & The New World Order
This week’s issue: for two years the AI build-out was priced as a growthstory. This week it started being priced as a credit story.
The New York Times reports that old jet engines are being scavenged to power data centers. GE Vernova’s filing shows customers ordered 57 aeroderivative turbines in the third quarter, up from 9, while it sold 13, or 4.4 orders per unit sold by our arithmetic. The grid alternative, by Lawrence Berkeley National Laboratory’s count, took a median 61 months for power projects built in 2025.
A note on sourcing. This piece is based on The New York Times report of 3 November 2026, of which we read the headline, byline, opening sentence and Techmeme’s summary; GE Vernova’s 10-Q for the quarter to 30 September 2026; releases from GE Vernova, ProEnergy and Boom Supersonic; a post on X by Boom’s chief executive; and Lawrence Berkeley National Laboratory’s Queued Up report. Ratios and totals labelled as ours are our arithmetic. We did not contact the companies.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
