Weekly review, week to 12 July 2026

WEEKLY ANALYTICAL MEMO: July 06–12, 2026

Theater: Ukraine / Europe. Synthesis of seven daily briefs and primary technical materials.

---

Key structural shifts

1. Breaking the "distance factor" as a paradigm of rear-area security. The strike on the Omsk refinery (the largest in Russia, processing 21–22 million t/year), with the FP-1's claimed range of up to 3400 km, eliminated the last refuge of Russian oil refining. The week's statistics record not an episode but a system: 32 of 44 refineries (73%) have been attacked, the fuel crisis has been officially acknowledged by Novak, and 7+ regions have introduced gasoline rationing by number. The key engineering-intelligence conclusion was formulated by Shtilerman: the trajectory to Omsk "was designed for more than a week" — direct evidence of the static nature of Russian air defense positional areas, which can be modeled in advance (source). [TECH] An analysis of the route confirms: two groups of UAVs flew along the border with Kazakhstan and through Bashkiria, joining up near Chelyabinsk — a deliberate bypass of stationary radars (source).

2. The opening of a maritime theater in the Sea of Azov as a "second front of attrition". During the week of 06–12.07, the USF claimed hits on ~90 shadow fleet vessels, at a pace of "every 112 minutes". Russian sources acknowledge not a tactical but a systemic failure: there is no unified shipping protection loop, and vessels travel in convoys without EW, radars or small-arms crews (source, source). The logic of the strikes has shifted from the front to the economics of supplying the peninsula.

3. The institutionalization of autonomy and "short-range air defense" on both sides. The AFU created a Long-Range Strike Command; UAV interception has taken shape as a separate branch of service with industrial-scale crew training (Sting/Taras-P). At the same time, a conceptual shift in the philosophy of C-UAS (counter-UAV) has taken shape — from the linear "detect–shoot down" chain to a "kill web" with economic selection of threats (source).

---

Technology dynamics and trends

Vector 1: Autonomy is displacing the radio link — the degradation of classical EW.

The main qualitative transition of the week: strike UAVs are losing their radio signature, depriving the adversary of a target for suppression.

• [TECH] The Russian Hornet DE-2 ("Marsianin-2", Swift Beat LLC) is an example of a mature transition. Three modes (Starlink mini / FPV / autonomous AI on a Qualcomm QCS5430 processor), but the key is inertial-optical odometry: when GPS is jammed, the drone orients itself using video from two cameras (day/night), matching the terrain against a map. The electric motor gives a low thermal signature; the airframe is EPP+plastic. The only vulnerability is slowing down to 80–90 km/h at the moment of target identification (source, source).

• The "Molniya" (a Russian fixed-wing FPV) was recorded attacking with no control antenna at all — only a camera and an onboard computer; the neural network was ported from the V2U base. The "Geran Seeker" has received automatic guidance on a Raspberry Pi. Against such targets a spectrum scanner is blind.

Vector 2: Fiber optics have taken to the air at range.

[TECH] OKB "Alfa" (Russia) has announced the first fixed-wing fiber-optic UAV — an analog of the "Molniya", range up to 50 km, warhead 6.5–8 kg. The main engineering problem has been solved — retaining payload capacity given the weight of the spool. Flash confirms: the goal is the same as with AI drones — to make the strike impossible to intercept via the radio link, while the image quality allows flying around obstacles at low altitudes, reducing radar visibility (source).

Vector 3: The "cheap drone → fast interceptor" race.

• [TECH] Griffen (Firebolt Engineering, British-Ukrainian) — the first confirmed combat use of a jet-powered interceptor against a "Geran": speed >350 km/h, altitude >7500 m, range up to 120 km, catapult launch. A new class of C-UAS has taken shape — an intermediate link between propeller-driven copters and expensive SAMs (source).

• [TECH] K100XR (Nordic Air Defence, Sweden) — kinetic hit-to-kill without a warhead, carbon fiber, ~0.5 kg, >350 km/h, IR seeker + onboard AI, operation in complete radio silence under GNSS jamming. The cost is a few thousand $ versus $480,000 for a Stinger; it is going to Ukraine for testing (source).

• Europe is purposefully building a high-speed base: the Quantum Systems Apex Recordhunter — 699 km/h (electric propulsion, V4Smart/Porsche batteries).

Vector 4: The failure of the "single button" — a transition to a network philosophy.

The key methodological shift of the week (source):

• An interceptor = a cheap guided munition, three stages (orientation by external cueing → coarse approach by autopilot → final lock-on to the motor's heat/noise/EM trace).

• Economic selection: at a 30% interception probability — launch 5–6 airframes per target (a "junk flock") and save the "Pantsir". Rogozin confirms with figures: a drone costing ~200 thousand rubles intercepts a target worth tens of thousands of $ (source).

• Removing the human from the time-critical loop. The task of EW is shifting from suppression to passive reconnaissance (acoustic profile of the propeller, triangulation).

Vector 5: Long-range weapons are moving from cottage-industry assembly to serial ballistic missiles.

[TECH] An analysis of the Fire Point plant (source, source): composite casings wound from polyacrylonitrile filament (baked at 120–180°C), radio-transparent fiberglass fairings, distributed production of "crude simplicity" — hundreds of "Flamingos" per month. The "Flamingo" fleet has grown from 64 (early June) to 90; booster stages are being tested to increase range from 2500–3000 to 4000–4500 km (source). [TECH] An in-house two-spool turbofan is being developed for the FP-5 for a low-altitude profile (unlike the borrowed AI-25, optimal for 6–10 km) — 10 test units are being assembled and the machine-tool base is being developed (source). Ballistic missiles are no longer an announcement: the FP-9 (up to 7500 km/h, ~1000 km), the "Hrim-2" and the American Rusty Dagger were shot down over Russia on June 30 (source).

Countertrend (degradation): the heavy multicopter "Baba Yaga" (Ukrainian) is losing effectiveness as a strike platform amid the growth of interceptors (2–3 interceptors are directed at a single airframe) and is being reborn as an aerial supply system — delivering large-caliber weapons, anti-tank weapons and Starlink to assault groups. "Logistics has taken to the air".

---

Weak signals (Under the radar)

1. The S-400 as an ersatz ballistic missile against ground targets — forced substitution. [TECH] Kyiv is being shelled with S-300/400 surface-to-air missiles (the 5V28 from the S-200V — up to 600 km), not "Iskanders". An indirect sign: the air raid alert is switched on only after the impacts, because preparation for a SAM launch cannot be tracked in time (source, source). This is an indicator of strain in ballistic missile production — "ramping up ballistic missiles is no easier than Patriot interceptors". For the engineering community — a signal of shifting launch signatures for early warning systems.

2. Acoustics against "invisible" drones. As targets move away from the radio link (fiber optics, autonomous AI), sound is becoming the only working detection layer. "Dronlok" ("RMS-Algoritm") — a passive acoustic sensor that sees an FPV with a payload at up to 2 km and a "Baba Yaga" at up to 3–5 km without a radio link. In parallel — a concept of distributed acoustic buoys as a long-range maritime C-UAS line for Crimea. This is the answer to Vectors 1 and 2.

3. FPV frequencies moving above 8 GHz. [TECH] An intercept at 8160 MHz with signal inversion has been recorded (previously — 6370/7540/7940). Detectors "stuck" at 5.8 GHz are going blind; "NABAT-3" urgently expanded its range to 8800 MHz. In parallel — the problem of VTX rules (channel changes every 3–5 s) and the Walksnail Ascent digital link (4.9–6.4 GHz), which many detectors do not see.

4. The physical layer of communications is "moving" into AI weights. [TECH] An analysis of NVIDIA AI Aerial: a neural network replaces PUSCH channel estimation in 5G, retraining for a specific city/interference environment after deployment. The strategic shift — training for a specific theater is becoming more important than the algorithm itself.

---

Vector projection and implications

Confirmed dynamics (high confidence):

• Classical wideband EW at 800/2.4/5.8 GHz is exhausting itself. Against the "FPV + relay + Starlink" combination, fiber optics and autonomous AI, suppressing the radio link is physically pointless. Countermeasures required in the coming weeks: mass deployment of passive multi-sensor systems (acoustic triangulation + electro-optical stations + compact radars for small RCS) and a transition to kinetic interception on the principle of "5–6 cheap airframes per target".

• Layered point air defense for refineries will become unavoidable. Omsk exposed a moral dilemma: kinetic interception of "Gerans" over residential neighborhoods vs. letting a strike on a refinery through (source). Forecast: placement of medium interception lines (Griffen-class, "Yolki") around fuel infrastructure + an attempt at a "digital dome" — shutting down the internet within a radius of up to 100 km of refineries (under discussion by the Ministry of Digital Development/FSB, source). The latter hits civilian infrastructure but is useless against autonomous drones — a dead-end measure.

• Maritime air defense for shipping requires an immediate systemic overhaul. A unified convoy protection loop (mobile groups, small radars, electro-optical systems, UAV interceptors, dispersal) — without it the Sea of Azov remains a "front". Ersatz measures (cables over the wheelhouse, shooting with MR-155s) are ineffective.

Analytical hypothesis (moderate confidence):

• The race for GPS-free autonomous navigation will define the next 12–18 months. Trident (UA, optical terminal navigation), DARPA (silicon levitation / the rubidium-based Perrseus quantum compass), V2U with Chinese AI — all point to a single vector: computer vision + new inertial sensors as the standard answer to all-pervasive EW. Whoever first makes an inertial chip cheap enough for expendable drones will remove the dependence on satellites.

• Satellite communications are opening up as a weapons domain. The declared Russia-China "anti-Starlink" coalition (lasers, HPM, the "Volna Kupol Garant" EW over ~20 km²) + the acceleration of the Russian "Rassvet"/"Bureau 1440" (22 satellites, a plan of 156 in 2026) + Fire Point's sovereign constellation (160 satellites, $200 million). Starlink as "Koschei's needle" (source) will become a priority target for physical and cyber action. The required countermeasure is redundancy of control channels and decentralized mesh architectures.

• The strategic vulnerability of C-UAS is supply chains via Taiwan/Japan. The Western drone industry is critically dependent on Taiwanese electronics and optics; a crisis around the island = an industrial shock for US allies.