u24.gov.ua · Official Ukrainian government platform
About the tracker
Original research. Open access.
Built alongside PhD research in War Studies at King's College London. Every entry carries source citations. All data is free to access and share under CC BY-NC 4.0.
UNMANNED SYSTEMS TRACKER · COMPILED FROM PUBLIC OSINT SOURCES · NOT AFFILIATED WITH ANY GOVERNMENT OR MILITARY ORGANISATION ·
DATA PROVIDED FOR RESEARCH AND EDUCATIONAL PURPOSES ·
Privacy Policy ·
Disclaimer 50% of advertising revenue from this site is donated to support Ukrainian military forces via United24
Sources: Ukraine MoD (DELTA-verified) · USF Cmdr Brovdi · Ukrainian Navy · DIU
July 2026
236
Boats & small vessels hit · record
205
Tankers struck · MoLoChKa to date
+74%
Strikes beyond 50 km vs June
30
RU air defence assets struck · USF
July marked a doctrinal shift in USV employment. DIU's Group 13 used Magura platforms as FPV carriers to strike a Podlyot radar command vehicle and a Parol-4 IFF interrogator in occupied Crimea, the first Magura strikes against land targets. Marines separately launched FPV drones from Barracuda USVs against positions near Zaliznyi Port, Kherson Oblast, losing one vessel to Lancet fire. USF struck a Russian naval USV base at Snizhne in Crimea, indicating both sides now field surface drones and target each other's basing. Of the 205 tankers struck since Operation MoLoChKa began, 130 were in the Sea of Azov and 75 in the Black Sea.
Black Sea Campaign · Cumulative
Ukrainian claims unless stated · Mar 2022–present · not photo-confirmed to Oryx standard
To Aug 2026
~30%
Black Sea Fleet destroyed or seriously damaged
205
Tankers struck · Op. MoLoChKa
236
Boats & small craft hit · Jul 2026 alone
18
Vessels damaged by Magura · to Aug 2024
Moskva
Flagship sunk · Apr 2022 · Neptune
Ivanovets
Corvette sunk · Jan 2024 · Magura V5
Tsezar Kunikov
Landing ship sunk · Feb 2024 · Magura V5
Sergey Kotov
Patrol ship sunk · Mar 2024 · Magura V5
The strategic effect was sea denial rather than fleet destruction. Russia withdrew major vessels from Sevastopol to Novorossiysk and further east, dismissed the Black Sea Fleet commander, and curtailed strike and logistics missions in contested waters. Estimates of the proportion of the fleet destroyed or disabled vary between 20% and 30% depending on source and date, and on whether relocated but intact vessels are counted. Russian counter-USV measures matured from March 2024, with harbour booms, helicopter patrols along approach routes and small-arms positions reducing USV survivability from Odesa.
Logistics, casualty evacuation, mine-laying, engineering and assault missions logged in the DELTA combat system.
Solid bars are exact published figures; hollow bars are MoD statements of "more than" or "almost" X and are floors, not exact counts.
Months with no published figure are omitted rather than interpolated.
Search
Side
Outcome
Year
Showing — events
Date ↕
Side ↕
UGV System ↕
Operation / Target
Location
Outcome ↕
Notes
Source
LOADING…
// NO RESULTS MATCHING FILTERS //
UAV Kill Board
USF PIDRAKHUYKA · UNMANNED SYSTEMS FORCES OF UKRAINE · VERIFIED HITS ONLY · Since USF creation: Jun 2025 – Jul 2026. Updated manually each quarter. · UPDATED: Jul 2026
⚠ DATA COVERAGE: The category table below shows the Jun–Dec 2025 annual totals; monthly figures drive the charts only. Monthly coverage is uneven because USF publishes a variable set of fields: recon flights are missing for May and June 2026, destroyed targets for April and June 2026, and personnel wounded for every month before April 2026. Gaps are left empty rather than interpolated. USF also revised its category taxonomy in 2026 — from July, figures are reported under new and expanded headings (multirotor drones, UGVs, fixed-wing UAVs, Shaheds and Gerberas, antennas, vessels, power hubs among them), so category series do not run continuously across that break. Figures are Ukrainian claims, verified in the DELTA system, covering both damaged and destroyed targets.
—damaged targets
—incl. destroyed
—strike flights
—recon flights
—enemy personnel
—including killed
—incl. wounded
CHART VIEW:Monthly total
STRIKE EFFICIENCY · DAMAGED TARGETS PER 1,000 SORTIES
CONFIRMED & ESTIMATED RUSSIAN MILITARY EQUIPMENT AND PERSONNEL LOSSES · FEBRUARY 2022 TO PRESENT · — · UPDATED: Jun 2026
⚠ DATA NOTICE: Personnel killed figures use Mediazona/BBC News Russian confirmed-by-name methodology as the primary source — the most rigorous publicly available approach. Equipment losses use Oryx photo-confirmed figures where available, supplemented by DIA and CSIS estimates. All figures represent floors, not ceilings. Sources are cited individually for each metric. Oryx ceased updating equipment loss estimates in 2025.
HEADLINE FIGURES
LOADING…
RUSSIAN AIR DEFENCE ATTRITION — UKRAINIAN STRIKE CAMPAIGN
DATA SOURCES: Oryx photo-confirmed equipment losses (2025 annual); USF Commander Brovdi official statements and USF operational releases; ISW daily assessments; Ukrainska Pravda/Oboronka mid-range strike analysis; SBU Alpha unit statements. USF figures are Ukrainian claims covering both destroyed and damaged systems, and are not photo-confirmed to the Oryx standard — the two should not be summed. Figures represent confirmed minimums. Last updated: August 2026.
77
SAM systems destroyed (2025)
Oryx photo-confirmed
23
Radar stations destroyed (2025)
Oryx photo-confirmed
174
AD assets struck Jan–May 2026
USF, Jun 2026 · ~$5.4bn claimed
248
AD units · USF first year
Brovdi, Jun 2025–Jun 2026
300+
AD systems destroyed · Crimea
Brovdi, Aug 2026 · campaign to date
DOCUMENTED STRIKES ON RUSSIAN AIR DEFENCE SYSTEMS
Date
System / Target
Type
Location
Outcome
Source
Aug 2025
GLONASS relay + RT-70 telescope
Strategic comms / SIGINT
Yevpatoriya, Crimea
Destroyed
tochnyi.info · 31st Air Defence Division HQ complex
Sep 10, 2025
40th Command & Measurement Complex HQ
C2 / deep space comms
Yevpatoriya, Crimea
Destroyed
tochnyi.info · HQ building + main comms building struck
Sep–Feb 2025/26
55Zh6U Nebo-U radar (×4 strikes)
VHF early-warning radar
NW Yevpatoriya, Crimea
Repeatedly hit
tochnyi.info · Same position struck at least 4 times
Jul 2025
S-350 Vityaz SAM system
Medium/long-range SAM
Occupied Ukraine
Destroyed
tochnyi.info · 1 of 3 confirmed S-350 strikes (Jul–Feb)
Nov 2025
S-350 Vityaz SAM systems (×7)
Medium/long-range SAM
Occupied Ukraine
Destroyed
tochnyi.info · Peak month — 7 strikes in single month
Table shows selected documented individual strikes. Full campaign: 365 mid-range drone strikes recorded Mar 2025–Mar 2026 (Oboronka/Ukrainska Pravda), with ~50% targeting air defence assets. Data from tochnyi.info ↗, Ukrainska Pravda ↗, ISW, Oryx. Updated manually — last updated March 2026.
Advertisement
Air Defence Log
RUSSIAN DRONE & MISSILE STRIKES ON UKRAINE · OCT 2022–PRESENT · SOURCE: UKRAINIAN AIR FORCE VIA PETRO IVANIUK · — · UPDATED: Daily auto-update
DATA SOURCE: Ukrainian Air Force official daily reports, compiled and verified by data scientist Petro Ivaniuk. Dataset verified by CSIS as matching official Ukrainian Air Force figures. Coverage from 27 September 2022 to present. Each row represents a single attack wave. · github.com/PetroIvaniuk ↗
Filters the stats and weapon table below. The trend chart has its own range control.
—Total launched
—Total neutralised
—Neutralisation rate
—Cruise missile rate
—Ballistic rate (excl. S-300)
—Got through
—— of which destroyed
—— of which EW-suppressed
—Attack waves logged
—Largest single wave
NEUTRALISATION RATE OVER TIME — DRONES vs MISSILES
Date range:
Oct 22
Now
BREAKDOWN BY WEAPON TYPE
ATTACK FREQUENCY — MONTHLY & ANNUAL
Low intensityMediumPeak (top 20%)Hover for details · Source: Ukrainian Air Force via Petro Ivaniuk
All weapon types combined · Hover bars for exact figures
Independent Russian media outlet maintaining the most rigorous publicly available list of Russian soldiers killed — confirmed by name using obituaries, social media, and official sources. Cross-checked against the National Probate Registry. This tracker uses Mediazona/BBC as its primary KIA source.
Independent Russian investigative outlet. Works with Mediazona to produce statistical estimates of total Russian deaths using probate registry excess mortality data — giving a broader picture beyond what's confirmed by name. Estimated ~352,000 killed to end of 2025.
The gold standard for equipment loss tracking — every entry in the Oryx database is individually photo or video confirmed. Tracks Russian and Ukrainian losses of tanks, AFVs, aircraft, and naval vessels. Ceased updating in 2025 but the historical record remains essential.
The most detailed open-source technical analysis of Ukrainian and Russian USVs. H.I. Sutton documents every known drone boat variant with specifications, operational history, and imagery. Essential reference for USV platform identification and capabilities.
Consistently the best English-language specialist defence reporting on USV and UGV developments. Well-sourced, sceptical of unverified claims, and fast to cover new platform developments. Used extensively as a secondary verification source for this tracker.
Ukrainian-language defence publication with strong English coverage of weapons systems, unmanned platforms and battlefield technology. Particularly valuable for early reporting on new Ukrainian drone and UGV developments, often sourced directly from manufacturers and the Ministry of Defence.
Ukrainian data scientist maintaining the most comprehensive structured dataset of Russian drone and missile attacks on Ukraine, compiled daily from official Ukrainian Air Force reports. Verified by CSIS as matching official figures. Covers every attack wave from October 2022 to present — the primary data source for the Air Defence tab on this tracker.
Ukrainian open-source research publication producing detailed long-form analysis of Ukraine's strike campaigns against Russian military assets. Primary source for the Air Defence Attrition section in this tracker — their March 2026 investigation documents Ukraine's systematic campaign against Russian SAM systems and radar networks across Crimea and occupied territories from 2025 onwards.
AFFILIATE NOTICE: The links below are Amazon affiliate links. If you purchase through them we receive a small commission at no extra cost to you. All books are recommended on their own merits and are directly relevant to the subject matter of this tracker.
📖
We Are Bellingcat
Eliot Higgins · 2021
The story of how a school dropout built the world's most influential open-source intelligence unit — and how citizen journalists are redefining the way we think about news, accountability, and the digital future. Essential reading for anyone using this tracker.
A concise, rigorous guide to war in the contemporary world by a Senior Lecturer in War Studies at King's College London. Examines how technology — including unmanned systems — is transforming warfare and widening the gap between military practice and international law.
By RUSI's Senior Research Fellow for Land Warfare, who has worked extensively with Ukrainian Armed Forces during Russia's invasion. An authoritative analysis of how autonomous weapons, AI and novel technologies are reshaping the battlefield — the closest thing to a field manual for understanding this tracker's subject matter.
The definitive history of the conflict by Harvard's foremost Ukraine scholar. Traces the origins of the invasion from the collapse of the Russian empire to the present day — essential context for this tracker's data.
The first book of frontline reportage from the Ukraine war, by the Guardian's award-winning correspondent. A compelling, on-the-ground chronicle of Russia's invasion and Ukraine's fight for survival.
Pulitzer Prize-winner Anne Applebaum's devastating account of Stalin's deliberate starvation of Ukraine in 1932–33. Essential background to understanding the deep roots of Ukrainian national identity and resistance.
The Ukraine War Unmanned Systems Tracker is an independent open-source research project built and maintained by Andro Mathewson, PhD Candidate in War Studies at King's College London. It documents the confirmed and claimed combat use of unmanned systems — surface, ground, and air — in the Russian invasion of Ukraine.
The tracker is built from publicly available OSINT, official Ukrainian military releases, and verified reporting. All entries carry source citations for independent verification. It is maintained alongside ongoing PhD research and updated regularly.
This tracker is not affiliated with any government, military organisation, or commercial entity. It is not affiliated with any of the organisations listed as sources above. Links are provided for research and verification purposes only.
PhD Candidate in War Studies King's College London
For research collaboration, media enquiries, data requests, or corrections to entries, please get in touch.
Corrections policy
If you believe an entry contains a factual error — wrong date, misattributed source, incorrect damage assessment — please get in touch with the relevant entry, your evidence, and a suggested correction. All corrections are reviewed and applied promptly. Significant corrections are noted on the relevant entry.
Media enquiries
Available for comment on drone warfare, unmanned systems, and the Russia-Ukraine conflict. Data requests, research collaboration, and citation queries welcome.
Original analyses using tracker data, and working findings on Ukraine's long-range drone campaign against Russia.
Incorrect password
Requesting access
This section holds original research that is still in progress and not yet peer-reviewed, so it is shared on request rather than published openly.
Access is normally granted to researchers, journalists and analysts. Get in touch with a short note on who you are and what you are working on, and I will usually reply within a few days.
Data and code underlying these analyses are available on request regardless of access to this page.
Deep Strike Research
UKRAINIAN LONG-RANGE DRONE CAMPAIGN AGAINST RUSSIA - WORKING FINDINGS - NOT PEER-REVIEWED
Original research by Andro Mathewson, PhD Candidate in War Studies, King's College London. Data: @dronbomber, @mod_russia, Open-Meteo ERA5, Jun 2024-Apr 2026. Code and data on GitHub
CLAIMED INTERCEPTS
48,615
Russian MoD propaganda
CONFIRMED STRIKES
2,480
@dronbomber OSINT
INFLATION FACTOR
20:1
Rising to 36:1 Apr 2026 · Measures the gap between Russian intercept claims and OSINT-confirmed hits — two differently-scoped datasets, not a direct ratio. Treat as a trend signal, not a precise multiplier. Gap widening is the meaningful signal regardless of cause
STRIKE DAYS
624
94% of all days
AD ATTRITION EFFECT
+84%
Severity post-campaign p=0.0000
MAX STRIKE RANGE
1,355 km
Bashkortostan, mean 510 km
RUSSIAN CLAIMED INTERCEPTS OVER TIME - PROPAGANDA
20x increase since 2023. Inflation factor rising from 2x to 36x.
CONFIRMED STRIKES BY TARGET CATEGORY
Strike volume doubled from ~30/month to ~160/month (early 2026).
TARGET EFFECTIVENESS - DAMAGE SEVERITY (0-3)
Radar/AD (2.08) and fuel depots (2.06) most damaging. Command/control 55% confirmation rate.
AD ATTRITION - CRIMEA SEVERITY (RED = AD DESTROYED)
CONFIRMATION RATE (%)
Severity +84%, confirmation rate +162% after the AD attrition campaign. p=0.0000 (time-series comparison). Caution: multiple confounders — improved drone technology, operator experience, target selection shift, and changing Russian AD posture — all coincide with the campaign period. Causality cannot be cleanly attributed to radar/SAM destruction without further controls.
STRIKE RANGE DISTRIBUTION (KM FROM FRONT LINE)
Min 92km, Mean 510km, Max 1,355km. Oil refineries deepest at 712km mean.
WEATHER EFFECT - SEVERITY BY CONDITIONS
No significant weather effect (cloud p=0.73). Ukraine strikes 94% of days regardless of conditions.
Analysis of Ukrainian long-range drone penetration rates by geographic corridor, cross-referencing Russian MoD intercept claims with @dronbomber OSINT-confirmed strikes. Penetration rate = confirmed OSINT hits ÷ Russian MoD intercept reports by oblast cluster. Higher penetration indicates Russia is under-reporting successful Ukrainian strikes.
Most heavily defended corridor. Belgorod (2,332 intercepts) and Kursk (1,658) dominate Russian AD reporting — transit zones saturated with Pantsir, Tor-M2, and EW systems. Despite this, confirmed hits are high in absolute terms. Key for: Krasnodar refineries, Rostov depots, Crimea AD.
Primary route to high-value central targets. Bryansk (1,183) and Smolensk (195) show heavy interception activity. Despite losses, Tver (293 confirmed hits) and Moscow (161) are successfully reached — saturation volume compensates for attrition. The "old route" with significant losses en route but the main path to strategic depth.
Northern / Baltic corridor
Highest penetration
Chernihiv → Bryansk → Pskov → Leningrad / St Petersburg
~61
Intercept reports
41
Confirmed hits
~67%
Penetration rate
AD density (over sea): Very low
Best penetration rate — but recently opened. Overland route through Russian territory following the western edge of Bryansk and Pskov oblasts toward Leningrad, avoiding the heavily defended Moscow air defence zone. High penetration rate (~67%) reflects sparse Russian AD coverage in Pskov and Leningrad oblasts. Drone incursions into Estonia and Finland recorded since March 2026 are best explained by Russian EW diversion near the border — there is no Baltic Sea route that avoids NATO territory, and Baltic governments have officially rejected Russian claims of tacit permission. Key targets: Ust-Luga and Primorsk oil export terminals, Leningrad refinery complex. Key limitation: only viable for northern targets.
Sparse AD, Russia rarely admits intercepts. Saratov (72), Samara (49) and Engels (43) all show confirmed dronbomber hits with minimal @mod_russia intercept reporting. AD systems thin out beyond the border corridor. Kuibyshev refinery complex and Engels bomber base have been repeatedly struck via this route. Key for: Volga industrial belt.
Russia has historically had minimal AD at this depth — though deployments of Pantsir systems to deep rear areas have been reported since 2024 and the picture is changing. Tatarstan (KAMAZ, TANECO), Bashkortostan (oil), and Siberia (airbases) remain among the least defended targets in the dataset as of early 2026. @mod_russia almost never reports intercepts in these oblasts. Key limitation: extreme range requires very long battery/fuel endurance.
Pre-positioned infiltration
Bypasses all corridors
Human intelligence networks · covert logistics · launch from within Russia
0
Intercept exposure
~20km
Final flight
6,452:1
Exchange ratio
AD exposure in flight: None
Operational complexity: Extremely high
Eliminates the corridor problem entirely. Operation Spiderweb smuggled 117 FPV drones into Russia via cargo trucks, launching within 20km of target airbases. ArdUPilot dead-reckoning navigation resists GPS jamming. Agents evacuated before launch. Requires 18+ months of preparation, HUMINT networks, and significant counterintelligence risk. Not replicable at scale and not directly comparable to corridor penetration rates — this is a covert HUMINT operation, not a drone corridor. The 6,452:1 exchange ratio reflects the unique zero-AD-exposure profile of this specific operation only.
Corridor Comparison — AD density vs Penetration rate
Higher penetration rate = more confirmed OSINT hits relative to Russian MoD intercept claims
How systematic AD attrition creates corridors
Ukraine's strategy: destroy radar/SAM systems to open gaps, then exploit them at scale
Step 1 · Radar suppression
492 verified strikes on Russian AD infrastructure Jun 2025–Mar 2026 (HartreeFock dataset). Nebo-U, Oborona-14, S-350 radars destroyed. Each loss degrades integration.
Step 2 · Gap exploitation
Ukraine routes subsequent long-range drones through radar coverage gaps created by Step 1. Crimea case study: confirmation rate 9.9% → 25.9% after intensive AD attrition (p=0.0000).
Step 3 · Volume saturation
Even on defended corridors, Ukraine launches enough drones to overwhelm point defences. 94% of all days see strikes. Pantsir and Tor have limited magazine depth — reloading takes time. Note: Russia has adapted with mobile reload vehicles and the Pantsir-SM (extended magazine); EW-based interception also reduces dependence on kinetics. Magazine depletion was a stronger constraint in 2022–23 than today.
Step 4 · Terrain / route variation
Drones fly tree-top altitude, use forested terrain masking, vary approach angles. Russian MoD confirmed flights via Belarus border corridor, circuitous routes over Black Sea, and now Baltic Sea.
Penetration rate = @dronbomber confirmed/reported hits ÷ @mod_russia intercept reports grouped by oblast cluster. Metric is imperfect: @mod_russia systematically under-reports intercepts in deep/eastern oblasts. Northern Baltic corridor penetration rate reflects land-route baseline — sea approach data insufficient for robust calculation. Damage severity scores (0–3 scale) are coded from @dronbomber imagery and reporting descriptions; inter-rater reliability not formally tested — potential for unconscious scoring bias exists and should be noted when interpreting severity trends. Working research — not peer-reviewed.
WARNING: @mod_russia figures are state propaganda. @dronbomber provides OSINT floor estimates. Working findings - not peer-reviewed.
Logistics Lockdown
NEW · MAY 2026
Ukraine's mid-strike campaign against Russian rear logistics · Announced 27 May 2026
On 27 May 2026, Ukrainian Defence Minister Mykhailo Fedorov formalised a $113 million programme — "Logistics Lockdown" — to scale medium-range drone strikes (20–200km depth) against Russian rear-area supply chains. The programme directs funding to the most effective brigades and runs centralised tenders for mid-range strike systems. Fedorov tied the push to a measurable drop in Russian assault tempo at the contact line. This section tracks the campaign's operational data and analytical findings. Not peer-reviewed — working findings only.
Primary Russian land corridor to Crimea under sustained attack
Volnovakha–Mariupol railway
ATESH partisan disruption · bypass route for Kerch Bridge-wary logistics
Analytical Findings
Finding 1
Record-pace vehicle attrition
483 vehicles destroyed in a single day (29 May 2026) — the highest daily figure since the full-scale invasion began. May's total of 7,704 vehicles represents a 4× acceleration versus the monthly average through 2025. Fedorov directly attributed the drop in Russian assault tempo to this attrition rate.
Source: Ukrainian General Staff · Euromaidan Press
Finding 2
Infrastructure targeting dominates
57% of all strikes inside Russia targeted infrastructure and logistics (electricity, telecoms, railways, fuel). Electricity strikes alone reached 31/month in Jan–Feb 2026 — a sustained winter-pressure strategy aimed at the Russian energy grid supplying occupied territories and border regions. Consistent with the systemic degradation hypothesis over point-strike approaches.
Source: Tochnyi dataset · NV.ua
Finding 3 — Caution
Assault tempo correlation
Fedorov claims a direct link between the vehicle attrition rate and a drop in Russian infantry assaults at the contact line. This is plausible but not yet independently verified — the correlation may reflect seasonal patterns, ceasefire dynamics around the May 9 period, or Russian operational pausing rather than logistics degradation alone. Treat as a working hypothesis pending further data.
Working hypothesis — not independently verified
Connections to Other Tracker Analyses
→
Analysis 5 · Geographic Dispersion
The dispersion of strikes to 8+ oblasts per attack observed in early 2026 may partly reflect the Logistics Lockdown's deliberate targeting of the southern land corridor — a geographic shift that Analysis 5 recorded as a trend without identifying a causal mechanism.
→
Analysis 6 · Targeting Doctrine
Logistics Lockdown targets vehicles and supply infrastructure — not personnel. This represents a deliberate shift toward materiel attrition at operational depth, consistent with the doctrinal asymmetry identified in Analysis 6 but at a different range band from FPV strikes.
→
Deep Strike · Southern Corridor
The Rostov–Crimea and Mariupol–Donetsk routes targeted by Logistics Lockdown overlap directly with the Southern corridor in the corridor analysis. The campaign effectively operationalises the corridor data — targeting the routes with the highest confirmed hit density.
→
Territorial Data · May 2026
May 2026 saw Russia's first net territorial loss since February. Fedorov attributes this to logistics degradation reducing Russian assault capacity. Plausible but unproven — the May territorial figure may also reflect seasonal factors and the lull around the Victory Day ceasefire period.
Sources: Fedorov / Ukrainian MoD · Tochnyi OSINT dataset · Euromaidan Press · CEPA · Militarnyi · NV.ua · Ukrainian General Staff
Working research — not peer-reviewed. Data collection ongoing. Last updated: June 2026.
Original Analyses
ORIGINAL RESEARCH USING TRACKER DATA · PHD CANDIDATE IN WAR STUDIES · KING'S COLLEGE LONDON · UPDATED: July 2026
Original research by Andro Mathewson, PhD Candidate in War Studies, King's College London. Working findings — not peer-reviewed conclusions. Data and code available on request.
Filter:
Weather and Shahed Interception Rates
April 2026 · n = 603 attack days · Shahed/Geran only · Oct 2022–Mar 2026
Air defence data: Petro Ivaniuk / Ukrainian Air Force · Weather: Open-Meteo Historical Archive API
Cross-referencing 603 Shahed/Geran attack days with night-time weather conditions from the Open-Meteo Historical Archive API. Each attack day was classified as Clear, Overcast, or Adverse based on cloud cover and precipitation.
~70% consistent across all weather
Clear nights: 70.7% (n=110) · Overcast: 72.2% (n=193) · Adverse: 68.2% (n=300). The 4-point range across conditions is not statistically significant — Ukraine's interception rate is weather-robust.
Cloud cover: r = 0.051, p = 0.21
OLS regression of night-time cloud cover against interception rate shows a near-zero relationship. Cloud cover explains less than 0.3% of variance in intercept outcomes.
Russia does not exploit adverse weather
Launch volumes are highest on clear nights (150.4/day) vs adverse nights (129.6/day). There is no evidence Russia times swarms to exploit poor weather conditions.
▸ METHODOLOGY NOTE
Weather retrieved for Kyiv (50.45°N, 30.52°E). Night-time = 20:00–06:00 local. Attacks with <10 Shaheds excluded. Clear = cloud cover <30%, Overcast = 30–79%, Adverse = ≥80% or precipitation present.
Air Defence Saturation: Interception Rate vs Launch Volume
April 2026 · n = 955 attack nights
Petro Ivaniuk / Ukrainian Air Force
Tests whether Ukrainian air defence degrades under high launch volumes — identifying the saturation threshold above which interception rates fall. Attack nights are bucketed by total weapons launched across nine volume bands.
Degradation zone: 61–150 weapons (66–67%)
The only volume band that consistently falls below the 74% overall mean. Medium-intensity attacks — not mass salvos — are most effective at degrading performance.
High-volume resilience: 83% at 600+ weapons
At the highest attack volumes (n=10), interception recovers to 83% — above the overall mean. Mass swarms do not overwhelm Ukrainian air defence in the way commonly assumed.
Non-linear relationship overall
The volume-to-interception relationship is non-linear. No attack nights fell into the overwhelmed category (<55%). The system appears structurally robust at both low and very high volumes.
Salvo Composition: Testing the 'Shahed as Decoy' Hypothesis
April 2026 · n = 1,060 attack nights
Petro Ivaniuk / Ukrainian Air Force
A widely cited claim holds that Russia uses Shaheds as decoys to saturate Ukrainian air defence and enable missile penetration. If true, mixed salvos (Shahed + missile) should produce lower overall interception rates than Shahed-only attacks.
Hypothesis not supported: 73% vs 75%
Mixed salvos (Shahed + missile, n=497) produce a 73% interception rate vs 75% for Shahed-only (n=411). The 2-point difference is within error margins — mixed salvos do not measurably degrade overall interception.
Missiles intercepted well alone: 80%
Missile-only salvos (n=82) produce the highest interception rate at 80%, suggesting Ukraine's missile defence performs well independently of Shahed saturation.
Volume may explain marginal degradation
Mixed nights use the highest mean volumes (113 weapons/night vs 101 Shahed-only). Any marginal degradation in mixed salvos likely reflects volume effects rather than a decoy function.
▸ METHODOLOGY NOTE
Salvo types classified from model field: Shahed Only (n=411), Mixed Shahed + Missile (n=497), Missile Only (n=82), Other/Unknown (n=70). Error bars represent 95% confidence intervals.
Russian Weapon Mix Evolution: From Missiles to Shaheds
April 2026 · Full campaign period
Petro Ivaniuk / Ukrainian Air Force · Aug 2022–Mar 2026
Tracks the composition of Russian aerial strikes over the full campaign — examining how the balance between Shaheds, missiles, and other UAVs has shifted since August 2022. Both absolute volume and proportional share are examined.
Shahed now 85–90% of all launches
Up from roughly 45–50% in late 2022. The campaign has become a drone-dominant operation, with missiles representing under 10% of launches proportionally by 2025–2026.
Monthly volume reached 8,000+ by early 2026
Approximately 20× the 2022 baseline. This represents both substitution away from expensive missiles and a raw expansion of strike capacity enabled by Iranian-origin drone supply chains.
Missiles declining as a share, not in absolute terms
Absolute missile numbers remain significant, but their proportional role has clearly shifted — from strategic primary weapon to supplementary precision element within Shahed-dominant salvos.
▸ METHODOLOGY NOTE
Weapon types classified by keyword matching on model field: SHAHED (Shahed-136, Geran variants), MISSILE (Kalibr, Iskander, Kh-series, Kinzhal, Zircon), OTHER_UAV, OTHER. Monthly aggregation.
Geographic Dispersion: Russia Spreading Attacks Wider Over Time
April 2026 · Oct 2022–Mar 2026
Petro Ivaniuk / Ukrainian Air Force
Tracks how many Ukrainian oblasts are affected per attack over time. Higher values indicate Russia is spreading strikes across more regions simultaneously — with implications for how Ukraine must allocate air defence resources nationally.
Mean oblasts per attack: ~3 → 8+ (mid-2025 to Mar 2026)
After falling to a low of ~3 oblasts per attack in mid-2025, geographic reach tripled to 8+ by March 2026 — a significant shift in Russia's targeting strategy.
Kharkiv most targeted: ~128 events
Kharkiv (~128) and Sumy (~115) lead by a significant margin, reflecting front-line proximity. Kyiv (~96) and Dnipropetrovsk (~98) follow as primary strategic targets.
Dispersal may be more effective than volume
Forcing Ukraine to maintain air defence coverage across 8+ oblasts simultaneously creates resourcing pressure that mass volume attacks on a single axis do not — a potential strategic adaptation.
▸ METHODOLOGY NOTE
Geographic dispersion = mean distinct oblasts per attack day, extracted from regions field. 3-month rolling mean applied. Oblast frequency covers full campaign period.
Sources: Boris Rozhin (@colonelcassad Telegram) · Veterans Notes (Telegram) · USF official statistics · via @ChrisO_wiki
Russia's Rubikon drone unit has published target breakdown statistics across 24,000+ engagements, enabling the first direct comparison of Russian and Ukrainian FPV targeting doctrine. The contrast is striking — and has prompted concern among Russian warbloggers about doctrinal mismatch.
22–30% personnel across reported periods — July 2025: 21.9% (5,134 of 23,433 targets), first half April 2026: 24.2% (4,840 of 20,022), Jun 2025–Jan 2026: 29.9% (50,238 of 168,000+). USF commander Madyar has stated a minimum 30% personnel targeting mandate.
Finding: fundamental doctrinal asymmetry
Ukraine prioritises personnel destruction (~25–30% of targets); Russia prioritises materiel and counter-drone (~37% drone-on-drone, only 6% personnel). Russian warbloggers have flagged this as alarming given the economics of war — infantry takes years to produce, vehicles can be factory-built. As of early 2026, unmanned systems account for approximately 60% of all effective strikes on Ukrainian forces.
▸ SOURCE NOTE
Russian data sourced from Rubikon's official Telegram channel via Boris Rozhin (Colonelcassad) and Veterans Notes Telegram commentary. Rozhin is a pro-Russian analyst — figures are taken as reported, framing is independent. Ukrainian data from official USF Pidrakhuyka statistics. Rubikon represents one Russian drone unit, not the full Russian USF equivalent; aggregate Russian figures may differ. Thread compiled by @ChrisO_wiki (independent military history researcher).
Drone Attribution: What Share of Russian Losses Do Drones Cause?
July 2026 · Monthly anchor points Nov 2025–May 2026
AFU General Staff daily loss reports · Syrskyi / Brovdi monthly unmanned systems reports
The General Staff publishes cumulative Russian loss figures (~1.44M personnel as of late July 2026); Ukraine's drone command publishes monthly, video-confirmed figures for personnel struck by unmanned systems. Dividing the second by the first, month by month, yields an attribution series: what proportion of claimed Russian personnel losses is drone-caused. Because both series originate from the same Ukrainian reporting ecosystem, the ratio is more internally consistent than either absolute figure.
~94–96% of monthly personnel losses now drone-attributed
Dec 2025: 33,019 drone-attributed of ~35,050 total claimed casualties (94%). Mar 2026: 96% of 35,351 total, per the AFU, with artillery and small arms accounting for the remainder. Since Dec 2025, monthly drone-attributed losses have exceeded Russian monthly recruitment; by May 2026 the cumulative 2026 gap stood at ~12,500.
A recent shift, not a wartime constant
Ukrainian army estimates through 2023–2024 put the drone-caused share at one-quarter to one-half of losses, with recon-drone-cued artillery adding a further indirect share. The climb above 90% tracks the scaling of FPV sortie volumes through 2025. This is why cumulative loss figures cannot be used for attribution: 2022–2023 losses were artillery-dominated, and a cumulative ratio would understate the current share by a wide margin.
Second-order evidence: the killed-to-wounded ratio has inverted
Modern wars typically produce roughly 1 killed per 3 wounded. Ukrainian intelligence assessments of internal Russian documents report a 62:38 killed-to-wounded ratio (Syrskyi's own figures imply 58:42). This inversion is independently consistent with drone dominance: persistent surveillance and re-strike capability convert wounded into killed at rates artillery warfare never did.
▸ METHODOLOGY & CAVEATS
Numerator and denominator use different confirmation standards: drone figures are video-confirmed strikes ("neutralized" includes wounded), while General Staff totals are broader estimates. Both are Ukrainian official claims; the ratio is indicative, not independently verified. Note that this tracker's own "USF personnel hits/mo" series (Key Trends) runs lower (~9–10k/mo) because Pidrakhuyka covers USF-branch units only, whereas Syrskyi's monthly figures aggregate all unmanned units across the Defence Forces. The drone-caused vs drone-assisted boundary is blurred where reconnaissance drones cue artillery, and near-total attribution partly reflects that drone video is now the primary confirmation mechanism for all casualty counting: strikes without drone footage are less likely to enter either count. Cross-reference Analysis 6: Russian-side reporting attributes ~60% of effective strikes on Ukrainian forces to unmanned systems, so the asymmetry runs in both directions but is steeper on the Ukrainian side.