The Carlin trend contains the largest concentration of Carlin-type gold deposits in the world. Two major controversies about these giant gold deposits have been their age, which is now firmly established as Eocene, and the source of heat, fluids, and metals, which remains debated. We present data that demonstrate an intense period of Eocene magmatism coincided in time and space with deposit formation and was arguably the primary heat source. Geologic studies over the last 40 years have emphasized the stratigraphy and structure of Paleozoic sedimentary rocks, which are the major ore hosts. However, four igneous episodes affected the Carlin trend, in the Jurassic, Cretaceous, Eocene, and Miocene. A Jurassic diorite-granodiorite laccolith and related dikes were emplaced at about 158 Ma in the northern Carlin trend. A Cretaceous granite intruded the north-central part of the trend at 112 Ma. Abundant Eocene dikes intruded along most of the trend and were accompanied by lavas in a large volcanic field along the southwest edge of the trend between ~40 and 36 Ma. Miocene rhyolite lavas erupted just west of and across the southern part of the trend at ~15 Ma. Exposed Eocene rocks consist predominantly of silicic to intermediate dikes, lavas, and epizonal intrusions, which we interpret to be sourced from a large Eocene plutonic complex underlying the Carlin trend. Eocene dikes are present in most deposits, are generally altered, but, with a few exceptions, were poor ore hosts. Distinct Eocene igneous suites, which are restricted to specific areas of the trend, are from north to south: (1) 40.3 to 39.0 Ma rhyolite to dacite dikes in the northern Carlin trend, centered approximately on the Betze-Post deposit and the richest part of the trend; (2) 37.6 Ma porphyritic rhyolite dikes including those that host ore at the Beast deposit; (3) 38.6 Ma intermediate to silicic intrusions of Welches Canyon; (4) 38.1 to 37.4 Ma andesite to dacite lavas and shallow intrusions of the Emigrant Pass volcanic field; (5) 36.2 Ma rhyolite dikes in the Emigrant Pass field that are indistinguishable from the 37.6 Ma suite except for age and location; and (6) ~37.5 Ma rhyolite to dacite intrusions and lavas of the southern Carlin trend (Rain subdistrict). Additionally, a few basaltic andesite dikes were emplaced at 37.8 Ma near Dee in the northernmost part of the trend and at 38.2 Ma near Rain. The petrography, distribution, and age of the Eocene igneous suites and aeromagnetic data indicate that each suite is underlain by a major, silicic pluton. The longer lived suites require either multiple plutons or long-lived magma chambers. All Eocene dikes cannot have come from any single magma chamber, for example, from a chamber beneath the Welches Canyon intrusions as proposed by some. In this case, some igneous suites would have been emplaced only 12 to 15 km north of the northern edge of the source pluton, would not have been emplaced above or symmetrically around the source pluton, and would be distinct in age from the proposed source pluton. These requirements are not consistent with the distribution and age of the igneous suites. The combined data for the Eocene igneous rocks require a plutonic complex about 50 km long (north-south), essentially coincident with the northern and central Carlin trend, and between 12 and 23 km across, underlying an area of ~1,000 km2. This complex was emplaced over ~4 million years that coincided with the formation of the Carlin-type deposits of the Carlin trend. Although many factors contributed to the formation of the deposits and the Carlin trend, magmatic heat was abundant in the right place and at the right time to generate the deposits. The Carlin trend may be the largest concentration of Carlin-type deposits because the Eocene igneous episode there was the largest and longest lived of the Great Basin.