Helsing
B2G (business-to-government). Sells AI software and increasingly vertically-integrated hardware to democratic-nation militaries under fixed-price product contracts and multi-year framework agreements, deliberately targeting software-like ~40-50% gross margins vs. the 5-10% of traditional cost-plus primes.
Chronological priced primary rounds. The 2023-2025 points are closed and well-documented; the 2026-05 point is a reported/in-progress round and is not confirmed closed - treat as directional, not booked. USD valuations approximate EUR marks at prevailing rates.
Earnings, margins, COGS & capex
Helsing is a scaling pre-profitability defense-AI company whose disclosed financials are minimal and dated: ~EUR 9.6M revenue in FY2023 (+721% YoY, per German annual accounts). It does not disclose current revenue, margins, or profitability. The investment case rests on very large European rearmament framework contracts (German Bundeswehr HX-2 loitering-munition award; EUR 580M CFSN 'Nucleus' combat-cloud software; Ukraine strike-drone orders) whose ceilings are potential, not booked. Margins are a thesis (software-like 40-50% target) not a reported result, and the recent pivot into owned hardware manufacturing pressures that thesis.
Income statement — where each revenue dollar goes
% of revenueOf every $1 of revenue, ~60¢ is cost of goods and ~40¢ operating expense, leaving ~0¢ of operating profit.
Revenue trend
Margins
at risk as owned-hardware mix rises
presumed negative / reinvesting
presumed negative (scaling + manufacturing buildout)
COGS structure
Not disclosed. For the software business COGS is low (compute/inference, integration engineering). The move into owned drone production (HX-2, HF-1) and combat-aircraft development (CA-1 Europa) adds airframes, propulsion, seekers/sensors, and electronics as bill-of-materials COGS, structurally diluting the software-margin profile.
Capex
Not disclosed. Rising: standing up distributed drone 'resilience factories' for HX-2/HF-1 at scale, plus R&D-heavy hardware programs (CA-1 Europa uncrewed combat aircraft, SG-1 Fathom / Lura underwater). Funded from the large equity war chest rather than operating cash flow.
Latest earnings
not applicable
no public guidance
- Employees
- ~761 (May 2026, per Sacra/German accounts), growing
- Cumulative equity raised
- ~EUR 1.37B through Series D (~$2.7B if the reported in-progress round closes)
- Last closed valuation
- ~EUR 12B / ~$14B (Series D, Jun 2025)
- German HX-2 award
- ~EUR 269M initial Helsing order (Feb 2026); two-supplier framework capped at ~EUR 1B per producer
Growth drivers
- European rearmament — Germany's Zeitenwende / EU 'ReArm Europe' spending and NATO members lifting defense budgets toward and beyond 2% (in cases 3.5%+) of GDP
- German Bundeswehr HX-2 loitering-munition award (initial ~EUR 269M Helsing order approved Feb 2026 within a combined ~EUR 536M Helsing+Stark package; the multi-year two-supplier framework was cut from ~EUR 4.3B to ~EUR 2B and capped at ~EUR 1B per producer)
- EUR 580M German contract for the Combat Fighter System Nucleus (CFSN) 'Nucleus' combat-cloud / autonomy software (national program salvaged from FCAS; includes two experimental UCAVs + reference architecture)
- Ukraine as demand engine and battlefield proving ground — orders for thousands of HF-1 and HX-2 strike drones (Helsing says 1,000+ HX-2 ordered by Ukraine's army)
- Platform land-and-expand — Altra battlefield-management software pulls through Helsing hardware (HX-2 swarms) and adjacent products (Cirra EW, Centaur AI pilot, Lura/SG-1 underwater)
- Sovereign-AI positioning — Mistral partnership (Vision-Language-Action models) and European-owned stack appeal to governments wary of US-vendor dependence
Bull & bear
Helsing is the European champion of a once-in-a-generation, government-funded rearmament supercycle - a software-first, AI-native player attacking cost-plus incumbents with product margins and real battlefield feedback, backed by enough capital to build a vertically-integrated defense-autonomy platform.
- Structural demand: European defense spending is rising for years, not quarters, and Helsing is a designated national/EU champion positioned for sovereign-AI mandates
- Contract proof points already landed: EUR 580M CFSN 'Nucleus' software, the German HX-2 award (~EUR 269M initial, ~EUR 1B-per-producer framework), and large Ukraine drone orders
- Software-margin ambition (40-50%) plus a platform (Altra) that pulls through owned hardware could compound into a high-margin, high-attach franchise
- Battlefield deployment in Ukraine is a moat competitors can't easily replicate - real feedback on EW resilience, swarming, autonomy
- Deep capital access (~$2.7B raised if the reported round closes, top-tier investors) funds multi-domain expansion (air, undersea, space via OHB JV) ahead of revenue
- Scarcity value: the most credible pure-play way for public-market and crossover investors to own European defense-AI, supporting a premium private valuation
A ~$14-18B valuation sits on ~EUR 10M of disclosed 2023 revenue, an unproven margin thesis, and framework ceilings that are not bookings - while the flagship HX-2 drone drew reports of battlefield stumbles and an order pause (which Helsing disputes), all against Anduril, Palantir, and rearming legacy primes.
- Valuation-to-fundamentals gap is extreme: last public revenue ~EUR 9.6M (2023) vs. ~$14-18B valuation; the multiple prices near-flawless execution
- HX-2 was reported to suffer takeoff/technical problems and lack promised AI features, prompting reports (Bloomberg) that Ukraine and Germany paused further orders in early 2026; Helsing denies a pause, but the contested episode leaves the core hardware story not de-risked
- Pivot from software into owned drone manufacturing structurally threatens the 40-50% margin thesis by adding BOM, inventory, and capex
- Framework 'ceilings' are options, not booked revenue - and the German loitering-munition framework was itself trimmed from ~EUR 4.3B to ~EUR 2B and capped at ~EUR 1B per producer
- Intense competition: US-backed Anduril (private) and Palantir (PLTR) plus European rivals Auterion, Quantum Systems, Tekever, Stark; Rheinmetall already chose a rival integrator
- Government-buyer concentration and political/budget-cycle risk; a Ukraine ceasefire or budget reprioritization removes a key near-term demand driver
- Autonomous-lethal-weapons ethics/regulation could constrain product roadmap and addressable markets
What it is worth
Private-market last-round + forward-multiple sanity check (no public price). Last closed priced round: Series D, ~EUR 12B / ~$14B (Jun 2025, led by Prima Materia/Daniel Ek). Reported in-progress round ~$1.2B at ~$18B (Dragoneer lead + Lightspeed co-lead, May 2026) not confirmed closed as of Jul 2026.
Materially below the last mark if HX-2 reliability issues persist, Ukraine/Germany orders stay soft, hardware dilutes margins, a Ukraine ceasefire cools demand, or Anduril/Palantir/Auterion take share - defense-tech multiples compress and the round premium unwinds.
~$14-18B range
real, growing government contracts and champion status support a premium private mark, but ahead of proven revenue/margins; value hinges on converting frameworks into bookings.
~$18B+ justified/exceeded
if EU rearmament accelerates, HX-2/CA-1 execution de-risks, framework options are exercised at scale, and 40-50% margins prove out - supporting an eventual IPO or continued up-rounds well above $18B.
Valuation is driven by strategic scarcity (the European defense-AI champion) and framework-contract optionality far more than disclosed fundamentals: ~$14-18B against ~EUR 9.6M of public 2023 revenue implies a multiple only justifiable if framework ceilings convert into large, high-margin bookings. The contested HX-2 order-pause reports and the hardware-margin dilution are the key downside swing factors.
SWOT
Strengths
- Europe's most valuable defense-tech startup with deep-pocketed backers (Prima Materia/Daniel Ek, General Catalyst, Lightspeed, Accel, Plural, Saab, BDT & MSD) and a ~$14B+ valuation providing a large equity cushion
- AI-software DNA and full-stack integration (Altra sensor-fusion + Cirra EW + Centaur AI pilot + owned drones) rather than a single-point product
- Anchor customer relationships with Germany, Ukraine, UK, and France, plus flagship framework contracts already awarded (CFSN 'Nucleus', HX-2)
- Battlefield deployment in Ukraine gives real combat feedback loops most Western competitors lack
- Strategically aligned with a durable, government-funded European rearmament cycle
Weaknesses
- Minimal and stale financial disclosure — last public revenue is ~EUR 9.6M (2023) against a ~$14-18B valuation implying extreme forward-revenue multiples
- HX-2 drones were reported (Bloomberg, Jan 2026) to have hit battlefield technical issues (takeoff failures, missing AI features, EW vulnerability), with Ukraine/Germany said to have paused further orders; Helsing publicly disputed the reports as an unverified internal document and says Ukraine ordered 1,000+ units - the episode is contested but dented product credibility either way
- Margin thesis (40-50%) is unproven and threatened by the shift from software into capital- and BOM-heavy owned hardware manufacturing
- Framework 'ceiling' numbers are potential, not booked revenue — the German loitering-munition framework was itself cut from ~EUR 4.3B to ~EUR 2B and capped at ~EUR 1B per producer
- Concentrated dependence on a handful of European government buyers and their political budget cycles
Opportunities
- Multi-year structural tailwind from EU/NATO defense-budget expansion and 'sovereign' European tech preference
- Expansion into new domains — uncrewed combat aircraft (CA-1 Europa, first flight targeted 2027), underwater (SG-1 Fathom / Lura), and space targeting (OHB joint venture)
- Cross-selling battlefield software as the integration layer atop other primes' hardware
- Export to additional NATO/allied and Indo-Pacific democratic militaries (e.g. reported Japan HX-2 initial agreement)
- Sovereign-AI partnerships (Mistral) to differentiate from US vendors
Threats
- Well-capitalized US competitor Anduril and software incumbent Palantir, plus European rivals (Auterion, Quantum Systems, Tekever, Stark) and legacy primes moving into autonomy
- Loss of Rheinmetall as a partner (it teamed with rival Auterion in 2024) shows prime-integration relationships are contestable
- Program cancellation / budget reprioritization risk inherent to government contracting (the German framework was already downsized)
- Reputational and order risk if further battlefield failures emerge
- Ethical/regulatory scrutiny of autonomous lethal systems could constrain features or markets; a peace/ceasefire scenario in Ukraine could soften near-term demand
Moats, dependencies & bottlenecks
Moats
EW, AI pilots, and owned drones into one kill-chain) Moderate-Strong Genuine technical depth, but Anduril/Palantir and rising rivals contest the software layer.
security clearances, and framework contracts with Germany/UK/France/Ukraine High switching costs and trust barriers once embedded in a national program; but contestable and politically exposed (framework already downsized).
Hard to replicate advantage, but reputational risk cuts both ways when systems underperform.
Large equity war chest and political favor; fades if a rival wins the champion mantle or capital markets cool.
Classic defense moat - slow, credentialed procurement keeps out pure software entrants.
Dependencies
UK, France, EU 'ReArm Europe') Nearly all revenue is B2G; a budget slowdown or peace dividend directly hits demand.
Demand / validation A ceasefire reduces near-term orders and removes the combat-feedback advantage.
Airbus, Hensoldt) Channel / integration Rheinmetall's move to rival Auterion shows these relationships are contestable.
foundation models incl. Mistral partnership) Supply / technology Depends on continued access to advanced compute and models for training/inference.
Supply / manufacturing Owned-hardware scaling exposes it to BOM sourcing, seekers, propulsion, and EW-hardened electronics.
Multi-domain buildout is funded by rounds, not operating cash flow.
Advantages
- AI-native, software-first culture attacking slow cost-plus incumbents
- European sovereign-champion positioning and political goodwill
- Top-tier capital base and premium valuation currency for hiring/M&A
- Multi-domain product breadth (air, undersea, EW, space JV) under one integration platform
- Real deployed combat experience in Ukraine
Weaknesses
- Valuation far ahead of disclosed fundamentals; opaque financials
- Hardware execution risk exposed by the HX-2 battlefield-issue reports and contested order pause
- Unproven margin structure as hardware mix rises
- Heavy concentration in a few European government buyers
- Dependence on the current geopolitical/war-driven demand spike
Bottlenecks
- Manufacturing scale-up — converting drone framework contracts into reliable, high-volume, battlefield-proven deliveries (the HX-2 reliability reports show the gap)
- Product reliability/quality control on hardware after the contested Ukraine/Germany HX-2 order-pause reports
- Talent: scarce defense-cleared AI/robotics engineers competing with commercial tech pay
- Converting framework 'ceilings' into exercised, booked orders
- Cross-border export controls and per-country certification for each national customer
Top signals & trends
Top signals
Investor conviction rising fast, round reportedly oversubscribed; but not confirmed closed as of Jul 2026 - treat as reported, not booked.
Converts political intent into a real order - but the multi-year framework was cut to ~EUR 2B and capped ~EUR 1B per producer.
Contested, but the reliability questions undercut the drone-hardware growth narrative regardless of the dispute.
Signals prime-integration relationships are contestable.
Multi-domain and sovereign-AI expansion, though early-stage.
Scale-up underway; cash-consuming ahead of disclosed revenue.
Trends
Primary structural tailwind; multi-year, government-funded.
Core to Helsing's thesis; validated by Ukraine.
Favors a European-owned defense-AI stack.
Funds Helsing but also arms rivals (Anduril, Auterion, Quantum Systems, Tekever, Stark).
Potential constraint on features, exports, and reputation.
Would soften the near-term demand spike and remove the proving-ground edge.
Ecosystem & competitor graph
Suppliers feed the company; customers pull from it. Line thickness shows the strength of each tie (supply-chain dependency, customer earnings contribution). Hover to isolate a tie.
GPUs/compute underpinning AI training and edge inference.
Strategic partner for Vision-Language-Action / generative models on defense platforms.
German sensor/radar maker; subcontractor on CFSN 'Nucleus' and partner on CA-1 Europa.
propulsion, seeker & EW-electronics vendors Bill-of-materials supply for HX-2/HF-1 owned-hardware production.
German Bundeswehr / Federal Ministry of Defence Anchor customer: CFSN 'Nucleus' EUR 580M software; HX-2 award (~EUR 269M initial).
Thousands of HF-1/HX-2 strike drones ordered (Helsing cites 1,000+ HX-2); reported to have paused further HX-2 orders in early 2026, which Helsing disputes.
European customer for defense-AI/software.
Customer / sovereign-AI defense buyer.
Closest analog - vertically-integrated AI defense (Lattice software + autonomous systems); far larger, US-anchored, well-capitalized.
Battlefield AI/data software (AIP, Gotham/Maven) - the software-layer incumbent Helsing competes with for the AI kill-chain.
Drone OS / autonomy software; won the Rheinmetall relationship - direct named rival in European drones.
German strike-drone startup (Peter Thiel-backed); co-awarded the Feb 2026 German HX-2-class loitering-munition package alongside Helsing.
German drone/ISR unicorn (~EUR 3B valuation); direct European competitor.
European UAS/ISR player deployed in Ukraine; competes for European drone budgets.
US loitering-munition maker (Switchblade) - direct product competitor to HX-2/HF-1.
US uncrewed combat aircraft (Valkyrie) - competes with CA-1 Europa ambitions.
AI pilot (Hivemind) for autonomous aircraft - competes with Centaur AI-pilot software.
Loitering munitions and EW at scale - established competitor in Helsing's hardware categories.
European prime scaling into autonomy; former potential partner now leaning to rival Auterion.