Anduril wins $20 billion counter-drone deal

Shield AI closed a $1.5 billion Series G on 26 March at a $12.7 billion valuation, up 140% from $5.6 billion.¹ Advent International and a JPMorgan Chase national security initiative co-led the round. Blackstone separately committed $500 million in non-dilutive preferred equity, bringing total fresh capital to $2 billion.

The money backs Hivemind, Shield AI's autonomy software selected by the US Air Force for the Collaborative Combat Aircraft prototype programme in February 2026. Hivemind has already flown on Anduril's YFQ-44A Fury in a confirmed mid-air software switch , meaning two companies competing for the same CCA contract can run their autonomy stacks on each other's hardware.

Shield AI projects $540 million or more in 2026 revenue, representing over 80% growth. At roughly 23 times projected revenue, the valuation prices Shield AI as a software platform rather than a defence contractor. The company is also acquiring Aechelon Technology, a flight simulation firm, to scale Hivemind training environments. Its V-BAT drone is already deployed in Ukraine.

Volodymyr Zelenskyy disclosed on 28 March that approximately 10 drone factories have been built abroad to circumvent Ukraine's wartime export ban.¹ One company sold 1,000 interceptor drones to a foreign buyer for $3.5 million while simultaneously holding a EUR 300 million state production contract. At least one European country purchased drones without warheads and then requested Ukrainian operators to accompany them.

The economics explain the leakage. Ukrainian interceptors cost $2,500 to $5,000 per unit.² A PAC-3 Patriot interceptor costs $13.5 million. That price ratio makes Ukrainian systems irresistible to any buyer facing drone threats, and demand from 11 nations remains blocked by the export ban . Manufacturers with excess capacity and uncertain state payment timelines have rational incentives to seek foreign hard-currency buyers.

Ukraine can technically build 1,000 interceptors per day but is budget-limited to roughly half that. The funding gap, perhaps $5 million to $10 million daily, represents the difference between a cottage industry and a global export platform. Zelenskyy warned the "window of opportunity" is narrowing: private manufacturers are outpacing state coordination.

Ukraine signed 10-year defence deals with Saudi Arabia and Qatar in March 2026, deploying 228 counter-drone specialists across five Gulf states.¹ A deal with the United Arab Emirates is expected imminently.

The personnel deployments are the most commercially significant element. Zelenskyy proposed drone combat hubs to the White House in August 2025; the US dismissed the offer . The Gulf states did not. At $2,500 to $5,000 per interceptor versus $13.5 million for a PAC-3 Patriot, Ukrainian systems offer a cost advantage of roughly 3,000 to 1. Gulf buyers gain operational counter-drone capability years faster than domestic development would allow. For Ukraine, these deals create a revenue stream independent of Western military aid.

In March 2026, Shield AI and Swiss manufacturer Destinus completed a two-month Hivemind autonomy integration campaign in Segovia, Spain, flying the software on the Hornet combat drone.¹ The tests demonstrated real-time autonomous route adaptation, a capability the Air Force considers essential for contested environments where GPS and communications links are degraded.

The European test matters for market positioning. Hivemind's proven interoperability with Anduril's YFQ-44A Fury already demonstrated hardware-agnostic operation on a US platform. Running on a Swiss airframe now extends that proof to NATO-aligned European manufacturers. If CCA doctrine evolves toward mandated interoperability between autonomy stacks, Hivemind's multi-platform record becomes a competitive advantage that single-vendor systems cannot match.

FAR clause 52.240-1 took effect on 13 March 2026, prohibiting drones manufactured or assembled by American Security Drone Act covered foreign entities from any federal contract.¹ Every federal contractor must now certify compliance.

The clause is the enforcement mechanism for the FCC's December 2025 decision to add foreign-manufactured drones and components to its Covered List . Where the FCC action blocked new product certifications, FAR 52.240-1 reaches into existing supply chains. Contractors using DJI, Autel Robotics, or any other covered manufacturer's components in federally funded work must now find alternatives or lose eligibility. The compliance burden falls hardest on smaller integrators who lack the procurement teams to audit complex component supply chains.

The US Army awarded AeroVironment two contracts totalling $135 million in March: $17.58 million for Red Dragon strike UAS on 12 March and $117.3 million for P550 long-range reconnaissance drones on 20 March.¹

Red Dragon is the more consequential platform. It operates at 400 km range with GPS-denied autonomous navigation and uses SPOTR-Edge for target classification when communications are degraded.² No public Department of Defense policy addresses autonomous target classification rules in that scenario. The Army is buying the capability before the doctrine exists to govern it.

The Section 232 investigation into UAS imports, opened in July 2025, reached its 270-day statutory deadline around late March 2026 with no public tariff decision announced.¹

No announcement does not mean no action. If tariffs are imposed, the 38% of Ukrainian drones still built with Chinese parts face an immediate cost shock. So do US integrators sourcing motors, flight controllers, and sensors from Shenzhen. Combined with FAR 52.240-1's procurement ban , the regulatory environment is tightening on two fronts simultaneously: procurement exclusion and potential import duties. Manufacturers have a narrowing window to secure alternative suppliers.

Zipline closed an additional $200 million on 23 March, extending its Series H to $800 million at an unchanged $7.6 billion valuation .¹ The company has passed two million cumulative deliveries. US volumes have grown 15% week-over-week for seven consecutive months.

New US markets include Houston, Phoenix, and Seattle. Internationally, Rwanda signed a national-scale Platform 2 contract. The 15% weekly compounding rate is the critical metric: sustained over seven months, it implies roughly 20x volume growth in that period. Zipline remains dependent on individual FAA waivers until Part 108 BVLOS rules are finalised , but the growth rate suggests demand is not waiting for regulation to catch up.

DroneShield posted FY2025 revenue of AUD $216.5 million on 20 March, up 276% year-on-year, and secured an AUD $49.6 million European military contract, its second-largest single order.¹ The company is scaling EU manufacturing capacity from AUD $500 million to AUD $2.4 billion annually by end-2026, a 4.8x expansion. That capacity build follows the opening of its first EU manufacturing facility .

The growth trajectory reflects a broader pattern. European defence procurement budgets have shifted from research funding to production contracts. DroneShield's bet is that European militaries will prefer locally manufactured counter-drone systems for competitive contract bidding, a calculation that makes geographic presence as important as technical performance.

Red Cat Holdings gained NATO NSPA catalogue approval for its Black Widow reconnaissance drone in March 2026, alongside reporting 161% full-year revenue growth.¹ The NATO listing opens allied procurement channels beyond the US market where Red Cat's stock rose over 60% after the Chinese drone ban .

The counterpoint comes from Fuzzy Panda Research, a short-seller that obtained FOIA documents showing the Army's Short Range Reconnaissance low-rate initial production contract at $12.9 million, against the $55 million value the company had implied. The discrepancy has not been resolved. For a company whose investment thesis rests on filling the gap left by banned Chinese competitors, the accuracy of contract disclosures is material.

AeroVironment acquired ESAero for $200 million ($160 million in stock, remainder cash), closing 16 March 2026.¹ The deal adds 300 employees and counter-UAS drone capacity at San Luis Obispo, California.

This is AeroVironment's second major acquisition in 12 months, following the $4.1 billion BlueHalo purchase in 2025. The pace raises integration questions. BlueHalo brought electronic warfare and space capabilities; ESAero adds airframe design and rapid prototyping. Together they position AeroVironment as a vertically integrated drone manufacturer, but absorbing $4.3 billion in acquisitions while executing $135 million in new Army contracts demands management bandwidth that few defence mid-caps have tested.

The Pentagon confirmed Gauntlet II for August 2026 at an undisclosed location described as "brutally hot."¹ The exercise will target procurement of 50,000 to 60,000 additional drones, with JIATF-401 running a live counter-UAS red team using GPS jamming, communications denial, and electronic warfare.

Phase I delivered 30,000 one-way attack drones from 11 vendors at roughly $5,000 per unit . Phase II raises the stakes: drones that cannot survive contested electromagnetic environments will be eliminated from the programme. The total target remains 300,000 or more drones by 2027 under a $1.1 billion budget . Two additional Gauntlets are planned beyond Phase II. Skycutter's 99.3/100 Gauntlet I score will mean nothing if its design cannot withstand active jamming.