Sub-system
Behind every announcement about generating power at a data centre is a machine with a delivery date. Gas turbines, reciprocating engines and fuel cells are the three practical options at this scale, and the first of them is sold out years ahead.
In one sentence
Prime power equipment is the generation machinery installed to serve a load directly — gas turbines, reciprocating engines and fuel cells — as distinct from standby generators sized only for outages.
The three options trade differently. Large gas turbines are the most efficient at scale, especially in combined cycle, and have the longest order books. Reciprocating engines are smaller, modular, quick to install and start fast, at lower efficiency. Fuel cells convert gas electrochemically with no combustion, which is quieter and cleaner at the point of use and costs more per unit of energy.
What has changed is that data-centre demand arrived at the same time as a wider power build-out, and turbine manufacturing capacity did not expand in anticipation. Slots for large frames are now booked years out, which converts a technology choice into a scheduling one: the right machine may not be the one available.
Split by the machine that makes the electricity — and each class is built by a different set of manufacturers.
A standby generator runs a few hours a year and is sized and maintained accordingly. A prime power machine runs continuously, which changes the duty rating, the maintenance interval, the emissions permit and the fuel supply arrangement. Equipment is specified and priced differently for the two, and conflating them understates what continuous generation requires.
Because they convert fuel electrochemically rather than burning it, fuel cells emit far less at the point of use, which can make a site permittable where a combustion source would not be. That is frequently the deciding argument rather than efficiency, and it is why they appear at urban and constrained sites disproportionately.
The common pattern is to install generation to energise a site while the grid connection is completed, then retain it for backup or peak shaving. That reframes the purchase as buying schedule rather than buying energy, and it changes which machine makes sense: fast to install beats most efficient.
2 of these are marked as a chokepoint: a handful of qualified suppliers, a multi-year lead time, or a single geography.
Turbine rotors and casings are heavy forgings from the same constrained supplier base that serves aerospace and industry.
Titanium and forgingsIndustrial turbines use the same superalloy and coating technology as aero engines, and share its supply chain.
Hot-section materialsContinuous generation needs a contracted gas connection, which has its own queue and its own constraints.
Permitting a large combustion source is frequently the binding constraint rather than the equipment itself.
A machine that cannot be stepped up, protected and synchronised produces nothing, and that equipment sits on the same multi-year lead times as the grid connection it is bought to substitute for.
TransformersEvery company named on a step below this page, ordered by how many of those steps it appears at. Compiled from the pages themselves rather than written separately, so the two cannot disagree. Not a ranking and not a recommendation.
9 more companies appear at a single step each; they are named on the pages for those steps.
Whether turbine order books ease enough to make on-site generation a fast option again.
Whether fuel cell deployment scales beyond current installation sizes.
Whether emissions permitting, rather than equipment, remains the binding constraint at constrained sites.
Each page explains one technology in plain language, states what it depends on, and names companies by what they supply at that step. Company roles are described qualitatively and deliberately carry no market shares, revenue figures or rankings — those change faster than an explainer can, and a stale number is worse than none. Ticker links point at company pages on this site and are provided for reference only.
Nothing here is investment advice, a recommendation, or a forecast. A company named on a page about a technology is not thereby a good investment, and the chokepoints described are structural facts about supply chains rather than predictions about prices. Technology moves; where a page describes something as unresolved or in development, that was true when it was written.