A field report on a moving system is, by its nature, a snapshot. These are the figures and claims this volume holds most lightly.
Russian losses claimed by Ukraine — forty-one aircraft hit at Spider's Web, a Kilo-class submarine critically damaged, ninety-six per cent of monthly Russian casualties from drones — are generally credible but disputed by Moscow and partially unverifiable. Where multiple sources conflict, this volume reports the conservative figure and notes the contest. Spider's Web: SBU says 41; US officials say ~20. Sub Sea Baby: UK MoD says probable serious damage; Russia denies.
As Rob Lee notes, the seventy-to-ninety-six per cent casualty figure conflates direct UAV kills with drone-spotted artillery kills. Drones mediate, but do not always cause, the fatal effect. The Atlas reports the figure in Sheet 00 with this qualifier intact.
Ukrainian and Russian monthly drone production figures span thirty per cent or more between sources. Bronk's six to seven thousand per month Shahed estimate is the high end; Bendett-cited ranges are fifty to seventy thousand annually, or about four to six thousand a month. The Atlas reports the lower number and notes the range in Sheet XI.
Statements about NATO drone walls, Taiwanese Hellscape, and 2027 autonomy horizons are projections, not facts. Zaluzhnyi's 2027 swarming warning, in particular, is informed speculation. Sheet IX is the only sheet drawn with a warm-coloured border, marking the projection.
The Atlantic Council, Carnegie, and IISS analysis from late 2025 onward all conclude that Russia has caught up or pulled ahead in several niches — Rubicon, Shahed volume, fibre-optic FPV scale. Early-war narratives of permanent Ukrainian drone superiority are obsolete. This is the contested point most consequential for Western readers: the second mover has, in several niches, become the first.
Every figure in this volume is a snapshot in May of 2026 of a rapidly moving system. The six-week cycle ensures the specifics will be wrong within months. The structural points — cost asymmetry, transparent battlefield, autonomy curve, industrial base — will be correct for years.
Read in primary, not coverage. The single best long-read on the war's structural meaning is Carnegie's April 2026 paper The New Revolution in Military Affairs; supplement with the OSW Centre for Eastern Studies' Game of Drones (October 2025), the CSIS Spider's Web analysis (June 2025), and the CNAS Hellscape for Taiwan (February 2026). Follow Rob Lee, Michael Kofman, Franz-Stefan Gady, and Samuel Bendett on X for analyst-grade running commentary. Read this volume as one synthesis of those sources, drawn at a particular date, with the contestations noted.
What this Atlas hopes you carry away is not the numbers — they will be wrong — but the shape of the network those numbers describe.