🛡️ Iran's Radar-Hunting Drones Are Exposing a Multi-Billion Dollar Vulnerability

💣 Pentagon Primes Agree to Quadruple Munitions Output as Stockpile Anxiety Grows

⚔️ Nine Days In: What Operation Epic Fury Has Actually Accomplished and Cost

🤖 The Pentagon Needs Military-Grade AI. These Startups Are Building It.

📌 Plus 22 more stories across Air, Land, Sea, Space & Cyber

Iran has been running a systematic campaign to blind American missile defense networks across the Middle East, using cheap kamikaze drones to target the most expensive and irreplaceable links in the kill chain: the radars. The results are sobering. An AN/TPY-2 radar (Terminal High Altitude Area Defense radar, part of the THAAD system) at Muwaffaq Salti Air Base in Jordan was destroyed or severely damaged, confirmed by commercial satellite imagery. At least one array of the AN/FPS-132 phased-array radar in Qatar, which provides 360-degree ballistic missile early warning across the region, sustained fire damage. Additional AN/TPY-2 sites in the UAE and Saudi Arabia were targeted. A small drone with a grenade-sized charge is sufficient to achieve a mission kill on equipment costing hundreds of millions of dollars.

The arithmetic is brutal for the defense industry and the procurement community. Each AN/TPY-2 costs between $250 and $300 million, and only 16 have ever been produced. The AN/FPS-132 package sold to Qatar in 2013 ran to roughly $2.1 billion in today's dollars. These systems are static, their locations are publicly known, and they take years to replace. The War Zone had flagged this vulnerability as far back as 2019, citing drone incursions at Guam's THAAD site. What was a warning is now a documented operational result. For Raytheon (RTX), Northrop Grumman, and the broader ground-based air defense community, the message is direct: point-defense capabilities for radar sites are no longer optional. Taiwan's Leshan radar, protected by ground-based Phalanx CIWS (Close-In Weapon System) batteries, is the model that planners are now referencing. The Golden Dome initiative, championed by the Trump administration, also takes on new urgency; space-based missile-tracking constellations would reduce reliance on these fragile terrestrial nodes.

The deeper strategic concern is how Iran's approach affects decision-making timelines. Degraded early warning coverage means less time between detection and intercept. For nuclear-armed adversaries or their proxies, that compression matters enormously. Defense investors and procurement officials should watch two tracks: first, what the Pentagon's emergency procurement posture looks like for replacement radars and expanded point-defense coverage at existing sites; second, whether the space-based tracking layer the Army and Space Force have been slowly developing accelerates on the back of this combat data. The case for moving faster just became a lot harder to argue against.

One week into Operation Epic Fury, the White House convened the defense industry's heavyweights — BAE Systems, Boeing, Honeywell Aerospace, L3Harris, Lockheed Martin, Northrop Grumman, and RTX — and extracted commitments to quadruple production of "exquisite class" munitions. No specific quantities, weapon types, or contract timelines were disclosed, which tells you this is as much about signaling as supply chain. Some of this was already in motion: Lockheed had previously committed to quadrupling THAAD output and tripling Patriot production; RTX had announced agreements in February to accelerate AMRAAM, SM-3, SM-6, and Tomahawk lines. The meeting formalized the pressure and broadened it.

The phrase "exquisite class" is doing significant work here. These are precision-guided munitions, cruise missiles, and interceptors, the weapons that consume stockpiles fastest in a high-intensity air campaign. The mismatch between consumption rates in combat and industrial production rates has been the central procurement anxiety since Ukraine began burning through Western inventories in 2022. That problem is now live for the US military, not just a planning scenario. Watch for supplemental appropriations requests from the Pentagon in coming weeks that put specific numbers and contract vehicles behind this meeting's announcements.

CENTCOM has struck more than 3,000 Iranian targets over nine days of Operation Epic Fury, the joint US-Israel air campaign. The assets involved span the full spectrum of American airpower: F-15s, F-16s, F/A-18s, F-22s, F-35s, B-1 and B-2 bombers, Tomahawks fired from destroyers, and LUCAS (Low-Cost Unmanned Combat Attack System) one-way attack drones making what appears to be their first confirmed combat employment. Iran's Supreme Leader Ali Khamenei has been killed. Iranian retaliatory strikes have hit targets across Saudi Arabia, the UAE, Qatar, Jordan, and Iraq, including a rocket attack on the US Embassy in Baghdad and a drone strike on a US logistics center in Kuwait that killed six American troops. Three F-15s were lost to friendly fire from Kuwaiti air defenses, all six aircrew recovered. The political track has not moved: Trump is demanding unconditional surrender; Iran's President Pezeshkian is refusing while signaling some willingness to halt regional strikes.

The LUCAS deployment deserves attention independent of the broader campaign. One-way attack drones were central to the Drone Dominance Program's pitch, and their actual combat employment — at scale, in a contested environment — is the data point the whole industry has been waiting for. Whether the performance matches the program's promises will shape procurement conversations for years. Analysts should also watch the Iran friendly-fire incident closely; the fog of war in a high-density airspace with multiple coalition participants operating different IFF (Identification Friend or Foe) systems is a known risk that just produced real casualties.

The Anthropic-Pentagon standoff crystallized a problem the defense tech community has been navigating for two years: frontier AI models built for mass consumer use are functionally incompatible with military operations. The core issues are not philosophical. Commercial AI requires persistent cloud connectivity, which is unavailable at the tactical edge and at significant risk in contested environments, as Iran's drone strike on AWS data centers in Bahrain demonstrated. These models hallucinate, they tell users what they want to hear, and they lack domain-specific military training. The Army is now running Project Aria to develop operational AI models internally. Meanwhile, a cohort of startups is building around the gap. Smack Technologies, founded by former Marine Andrew Markoff, has raised $32 million and is building what it describes as a frontier lab for national security, with separate product suites for generative AI tasks and for denied/degraded network environments. Jason Rathje's webAI is focused on data and model sovereignty, with edge-deployable systems. IntelliGenesis, led by retired Navy commander Pete Walker, is developing a counterfactual-thinking framework specifically designed to prevent AI sycophancy in operational contexts.

The investment thesis here is simple and increasingly well-supported by events. As AIN Ventures' Sherman Williams put it, a model that is 85% as capable as GPT-5 but runs on a DDIL (denied, disrupted, intermittent, and limited) network beats GPT-5 in a data center you can't reach. Iran just demonstrated exactly that attack vector. SOCOM and INDOPACOM are reportedly driving strong pull for these purpose-built systems. The question for defense tech investors is which of these early-stage players has the distribution relationships, security clearance infrastructure, and program office access to convert that demand signal into durable contracts before the larger primes absorb the problem.

✈️ Air

🛡️ Land

🌊 Sea

🧠 C2 & Cyber

🏭 Industry & Policy

That's the brief. If this landed in the right inbox, forward it on.