Phillips P. O'Brien, the professor of strategic studies at St. Andrews whose book War and Power has become a touchstone for readers trying to make sense of the Ukraine conflict, opens his Weekend Update #179 with a number that should reshape how the war is discussed. In March 2026, roughly 96 percent of Russian casualties were caused by Ukrainian unmanned aerial vehicles. The remaining four percent came from artillery, tanks, and small arms combined. If the figure holds up, it is the most dramatic shift in casualty causation in modern industrial warfare.
The Numbers Under The Number
O'Brien builds his case carefully. He quotes President Zelensky directly: "Russian losses this March have reached their highest level since the start of the war: our drone strikes alone resulted in 33,988 Russian servicemembers killed or seriously wounded." Defense Minister Fedorov adds the broader frame. "March was a historic month for the Army of Drones," he says. "35,300+ enemy casualties and 151,200+ targets hit—a record-breaking performance." And on sustainability: "For 4 months in a row, Russian losses have exceeded their replenishment rates. We are on track to our strategic goal: 50,000+ per month."
O'Brien admits that the 50,000 figure is one he previously considered unrealistic, and his willingness to revise publicly is one of the more useful features of the newsletter. His production numbers compound the story: Ukraine is on track to manufacture 7 million military drones in 2026, up from 4 million in 2025 and 2.2 million in 2024. The scaling curve is the kind of thing that turns a battlefield innovation into a structural condition.
Systems Over Platforms
The strategic frame O'Brien returns to repeatedly is the distinction between systems warfare and battle-centric warfare.
"Systems warfare is more complex intellectually but also far more effective than battle-centric warfare."
The observation is drawn straight from his book, and he uses it to explain why Ukraine's campaign against Russian oil infrastructure is producing cascading effects disproportionate to the raw strike counts. "Russia has to stop pumping and refining oil that they could handle," he writes, "because there is no place to put the product." The metaphor he reaches for is clogging: refineries, ports, and storage tanks damaged simultaneously create downstream failures that don't show up in any single target assessment. The bombs count less than the network topology.
O'Brien frames this as the practical vindication of his long-standing argument. "It's not about the number of tanks, aircraft or artillery pieces," he writes. "It's whether you can integrate the actions of all of these." Russia, with more of everything, is losing the integration contest. Ukraine's advantage is doctrinal rather than material.
What The Newsletter Skips
The analysis is persuasive, but it leans hard on Ukrainian-source data that O'Brien doesn't independently verify. He mentions the ePoints verification system in passing without describing how it works or how vulnerable it is to the inflation incentives that every party in every war has. A 96 percent drone-casualty figure is the kind of number that deserves more methodological scrutiny than it gets here, if only because getting it wrong would invalidate a lot of downstream analysis.
The production trajectory is similarly under-examined. Ukraine plans 7 million drones in 2026, but plans are not deliveries. The newsletter doesn't address supply-chain vulnerabilities, component sourcing bottlenecks, or the effect of ongoing Russian strikes on Ukrainian manufacturing. A single successful attack on a motor or chip supplier could compress the curve substantially.
Russian adaptation gets even less attention. The analysis essentially freezes Russian capabilities at their current state, but electronic warfare, hardened dispersal, and expanded air defense are all plausible responses that would change the math. If drones account for 96 percent of casualties today, the figure could look very different in six months for reasons that have nothing to do with Ukrainian effort.
O'Brien also mostly avoids the question of whether the casualty ratio reflects a tactical shift on the Russian side rather than purely technological dominance. If Russian forces are concentrating dismounted infantry in ways that present easier drone targets, the ratio measures a battlefield choice as much as a weapons breakthrough. The newsletter treats this possibility as residual.
Finally, Ukrainian casualty dynamics get a single acknowledgment that Ukrainians face "even worse" demographic pressures, and then the analysis moves on. The asymmetry matters. A war in which both sides are running out of soldiers but one side is running out faster is a different war from one in which only the adversary is degrading.
Bottom Line
O'Brien has identified what may be the most important single statistic of the current phase of the war, and his systems-warfare frame gives readers a genuinely useful way to think about why Ukrainian strikes are producing disproportionate effects. The 96 percent figure, if it holds, marks a transition that will eventually be taught in war colleges. Readers should also notice what the newsletter doesn't do: verify the casualty data independently, stress-test the production projections, or seriously engage with Russian adaptation. The analysis is sharp on direction and weak on uncertainty, which is a common failure mode for commentary written during wartime. Take the thesis seriously. Hold the numbers loosely.