HX-2 should not be confused with the HF-1, a separate and cheaper Helsing loitering munition with a plywood fuselage that the company has also supplied to Ukraine; the two systems are frequently conflated in press coverage of German-funded drone deliveries, a confusion addressed in detail under Conflicts below.^7,17 This article covers the HX-2/Karma airframe specifically: its design, its rapid rise through German government contracts and venture funding, and its considerably more contested record on the Ukrainian front line.

HX-2 measures 112.5 cm in length and has a takeoff weight of up to 12 kg, placing it in the mini-UAS "Weight Class 1" category.³ The airframe combines four wings and four rotors in an X arrangement, giving it vertical takeoff and landing capability without a launch tube, while its electric propulsion is described by Helsing as reaching a cruise speed of about 80 km/h and a maximum speed of 250 km/h; several independent outlets, including New Atlas and the War Zone, instead cite a maximum speed of 220 km/h, a discrepancy this article notes without resolving in favor of one figure.^2,3,5,17 Disclosed range is up to 100 km, and the munition carries a payload of up to 4.5 kg.^2,3

Targeting runs through Helsing's Altra reconnaissance-and-strike software, which the company describes as controlling onboard sensors, evaluating results, and generating engagement recommendations for a human operator; Helsing states a human "stays in or on the loop for all critical decisions," with the operator confirming a target before HX-2 commits to its terminal approach.² Navigation relies on stored mapping data and landmark-based image recognition rather than satellite positioning, which Helsing markets as resistance to GNSS jamming and to loss of a continuous control link — a claim later tested, with mixed results, on the Ukrainian front line (see Conflicts).^2,4

The warhead is modular, with Helsing marketing options described as anti-armor and multi-purpose; independent reporting describes a shaped-charge configuration intended to penetrate armored vehicles alongside blast-fragmentation options for structures and personnel.^2,9 In June 2026, Helsing signed a memorandum of understanding with Eurenco, France's state-owned energetic-materials manufacturer, to co-develop a "complete high-performance Modular Explosive Warhead" for HX-2; the companies framed the deal around supply-chain sovereignty, citing 12–24 month lead times for propellants and explosives as the reason armed forces should not depend on warhead components sourced outside Europe.¹⁴ With the Eurenco agreement, Helsing states that the HX-2 kill chain — airframe, Altra autonomy software, and now explosive fill — runs entirely on European suppliers.¹⁴

Helsing was founded in Munich in March 2021 as a software-first artificial-intelligence company for defense applications, raising a €100 million Series A later that year led by Spotify founder Daniel Ek's investment vehicle Prima Materia.¹ In September 2022 the company announced a partnership with Rheinmetall to integrate Helsing's AI software into Rheinmetall's existing weapons platforms; that partnership dissolved in 2024, and Rheinmetall instead began integrating AI software from Helsing's competitor Auterion.¹ Helsing itself pivoted toward building its own hardware in 2024, raising a €450 million Series C in July of that year at a €5 billion valuation, and using the proceeds in part to build its first "Resilience Factory" production line in southern Germany, completed in December 2024 with an initial capacity of more than 1,000 HX-2 units per month.^1,6

Germany's defense minister Boris Pistorius confirmed to the newspaper Bild on 18 November 2024 that Germany would fund 4,000 HX-2 Karma loitering munitions for Ukraine, with a first batch of "hundreds" expected that December.⁴ Helsing formally unveiled the HX-2 Karma two weeks later, on 2 December 2024, the same day German Chancellor Olaf Scholz visited Kyiv.^2,17 Reporting on that original 4,000-unit pledge is contested: a later Defense Express analysis found that the original September 2024 contract for 4,000 units actually covered the cheaper HF-1 system rather than HX-2, and that a separate, additional 6,000-unit HX-2 order announced by Helsing in February 2025 — around the Munich Security Conference — was structured as an optional extension requiring further German government funding approval whose status was not publicly confirmed.^6,7 By July 2025, Ukraine had received roughly 2,200 HF-1 units under the original agreement, according to the same analysis.⁷

Helsing's corporate growth continued through 2025 and 2026 on the back of both Ukraine-related and NATO-facing procurement. The company raised a €600 million Series D in June 2025 at a €12 billion valuation led by Prima Materia, and was reported in May 2026 to be closing a $1.2 billion round led by Dragoneer Investment Group at an approximately $18 billion valuation.^1,13 On the government-contract side, Germany's Bundestag budget committee approved an initial €269 million HX-2 contract for the Bundeswehr in early 2026, with a framework option that could grow to €1.46 billion over seven years, subject to a combined €2 billion cap shared with competitor Stark Defence.¹³ That followed a roughly €900 million three-way contract reported in November 2025, split evenly among Helsing, Stark Defence, and Rheinmetall, covering up to 12,000 drones for a new German army brigade stationed in Lithuania; in the German military's October 2025 evaluation ahead of that award, HX-2 reportedly completed all 17 of its test flights successfully, in contrast to repeated failures logged by Stark Defence's competing Virtus system.¹²

HX-2's most closely watched deployment has been in Ukraine, where its record is considerably more contested than Helsing's NATO-facing trials. Field testing was carried out by Ukraine's 14th Separate Regiment of Unmanned Systems, later restructured into the 1st Separate Center for Unmanned Aviation Systems.¹¹ An internal German military presentation dated 20 November 2025, compiled by General Günter Schneider and later reported by Bloomberg, recorded that only about 25% of HX-2 units launched successfully during frontline trials, attributing most failures to catapult malfunctions, and noted that some units lacked AI-driven flight-control components the design had advertised; the same reporting described Russian electronic-warfare systems disrupting operator data links to units that did reach the field.^11,16 Separately, a Ukrainian document submitted to Germany's Ministry of Defense and reported by the German newspaper WELT put HX-2's combat accuracy at 5 hits out of 14 engagements — about 35.7%.¹⁰ Bloomberg reported on 19 January 2026 that Ukraine had paused additional HX-2 orders pending resolution of these issues, and that Germany would not fund further orders without a renewed request from Kyiv.^8,9

The Ukrainian controversy stands in contrast to HX-2's results in NATO-organized exercises away from the front line. During Project Flytrap 5.0, a US Army V Corps-led multinational counter-drone exercise at the Pabradė training area in Lithuania in June 2026 involving American, British, and Australian personnel, HX-2 recorded 15 confirmed kills and 2 near-misses across 17 engagements among more than 20 systems evaluated.^15,16 US Army Chief Technology Officer Alex Miller said HX-2 "was initially evaluated for one-way attack and counter-drone missions but demonstrated additional value as a reconnaissance and loitering platform."¹⁵ Whether the gap between the Lithuania exercise results and the Ukrainian frontline reports reflects different testing conditions, electronic-warfare intensity, unit training, or airframe batches is not established by any source reviewed for this article.

Germany is HX-2's home-market operator: the Bundeswehr's own procurement was approved by the Bundestag budget committee in early 2026, an initial €269 million contract with a framework option of up to €1.46 billion over seven years, separate from Germany's role funding deliveries to Ukraine.¹³ Ukraine has received and field-tested HX-2 units through at least one unmanned-systems unit, though the scale of deliveries is disputed: press coverage of a 2024 pledge for "4,000" units has been reported to actually describe Helsing's separate HF-1 system, a further 6,000-unit HX-2 order announced in February 2025 was structured as an optional extension pending additional German funding, and Ukraine paused further HX-2 orders in January 2026 following disputed frontline test results.^6,7,8,11 HX-2 has also been evaluated, but not confirmed as procured, by United States, United Kingdom, and Australian forces during the multinational Project Flytrap exercise series in Lithuania.^15,16

HX-2's 100 km disclosed range sets it apart from most rivals in its weight class. The Russian ZALA Lancet's standard Lancet-3 variant weighs a similar 12 kg at maximum takeoff weight but is disclosed with a considerably shorter 40 km range, a 110 km/h cruise speed, and up to 300 km/h in its terminal dive — figures broadly comparable in speed class to HX-2 but confined to a shorter operating radius.¹⁸ The AeroVironment Switchblade 600 is heavier at roughly 22.7 kg for the airframe alone (54.5 kg including launch tube and fire-control system) and trades HX-2's rotor-borne VTOL launch for tube launch, with a disclosed 80 km operational range built around a 20-minute flight-out followed by 20 minutes of loiter and a Javelin-derived anti-armor warhead.¹⁹ Anduril's Bolt-M, fielded by the US Marine Corps, is considerably lighter at roughly 6–7 kg with about 40 minutes of endurance and a smaller disclosed 20 km-class range and 1.4 kg warhead, positioning it as a shorter-range, longer-loiter complement rather than a direct HX-2 analogue.²⁰

Across this field, HX-2's combination of VTOL launch, 100 km range, and swarm-capable AI targeting is distinctive on paper, but it remains the only one of the four whose combat accuracy has been put in writing at under 40% by a national government's own internal reporting, a gap between disclosed design specification and documented frontline performance that has not been reported for Lancet, Switchblade 600, or Bolt-M in the sources reviewed for this article.^10,11

MASS Assessment

Like its German rival STARK Virtus, HX-2's MASS components are shaped as much by documented uncertainty over real-world performance as by its own disclosed specifications.

Range (50) sits well below what HX-2's own disclosed 100 km spec would suggest on paper, because this archive's scoring weighs the disputed frontline test results covered above — a German military presentation reportedly found roughly 25% launch success and a Ukrainian document reported by WELT put hit rate at 35.7%, both disputed by Helsing itself. A spec sheet is not a combat record, and that gap is the main reason range and loiter endurance (70) don't score as highly as HX-2's headline numbers alone would imply. Portability (85), by contrast, is a straightforward design fact — a 12 kg, VTOL-capable airframe is genuinely light and deployable regardless of how it performs once launched — and scores accordingly high, among the best in this batch.

Precision (80), autonomy (75), and technology (80) credit HX-2's AI-assisted navigation and image-recognition target tracking, a genuinely modern approach to human-on-the-loop terminal engagement, largely independent of the disputed field results since the underlying software concept is well documented regardless of battlefield hit rate. Lethality (60) reflects a modular warhead (shaped-charge or fragmentation) of moderate size relative to this batch's heavier anti-armor systems. Sensors (65) sits at the middle of the range for similar reasons to its peers — a capable but not fully disclosed sensor package. Counter-measure resistance (40) is the lowest of HX-2's components, reflecting both the lack of disclosed anti-jam specifics and the same frontline-reliability questions affecting range.

Cost-effectiveness (70) trails this archive's separate 85/100 overall price-performance rating shown in the sidebar for the same reason range trails its own spec: a competitive price is confirmed, but whether it is buying reliably delivered capability is not yet settled by the sourcing available.

What is the Helsing HX-2 Karma? It is an electrically powered, X-wing loitering munition built by the German AI-defense company Helsing, unveiled on 2 December 2024 as the "HX-2 Karma"; most reporting since early 2025 refers to the same system simply as HX-2.^2,3

Is HX-2 the same as the Helsing HF-1? No. HF-1 is a separate, cheaper Helsing loitering munition with a plywood fuselage. The two are frequently conflated in press reporting on German-funded Ukraine deliveries, and at least one analysis found that an early "4,000-unit" pledge widely reported as HX-2 actually covered HF-1.^7,17

What are HX-2's specifications? Helsing states a length of 112.5 cm, a takeoff weight of up to 12 kg, a payload of up to 4.5 kg, a cruise speed near 80 km/h, and a disclosed range of up to 100 km. Maximum speed is given as 250 km/h in Helsing's own materials, though some independent outlets cite 220 km/h.^2,3,5

How has HX-2 performed in Ukraine? Results are disputed. A German military presentation reported by Bloomberg recorded roughly 25% launch success in frontline trials, and a Ukrainian document reported by WELT put combat accuracy at 5 hits in 14 engagements (about 35.7%). Helsing disputes both figures and says its own trials elsewhere produced hit rates close to 100%.^10,11

Does Rheinmetall build the HX-2? No. Helsing and Rheinmetall had an AI-integration partnership announced in 2022 that dissolved in 2024, before HX-2 entered production. Rheinmetall builds and funds its own separate drone deliveries to German forces, including a share of a 2025 three-company Bundeswehr contract alongside Helsing, but does not co-manufacture HX-2.^1,12