Defense becomes a manufacturing race
Everyone is funding drones and autonomy. The constraint is production: factories, shipyards, rocket motors and cheap interceptors. That is where I would put the next check.
Neil Gaikwad
The size of the prize
What the consensus misses
The consensus bet is autonomy: drones, swarms, AI pilots. It is right about the direction and late on the price. Valuations across the category roughly doubled in a year, and Anduril's single $5B round was a third of all defense-tech dollars in the second quarter of 2026. PitchBook put it plainly: the dollars are concentrating, not broadening.
What I think is underpriced is the boring part. Ukraine changed the math to millions of cheap systems a year, and the US cannot build at that rate. It built 0.1% of the world's commercial ships in 2024. The constraint is machining capacity, energetics and solid rocket motors, test ranges and certification. Hadrian, which builds automated factories for defense parts, went from a $1.6B to a $7.87B valuation in seven months. That is the market telling you where the scarcity is.
The second underpriced idea is cost-per-kill. When a drone costs a few hundred dollars and the missile that stops it costs much more, whoever makes defense cheap wins every budget cycle. Counter-drone sensing and low-cost interceptors are a better seed market than another drone.
Why now
The money moved. US defense spending for 2026 is $839.2B plus reconciliation funds, the 2027 request is $1.5T, NATO allies pledged 5% of GDP by 2035, and the EU's €150B defense loan program started paying out in mid-2026.
The rules moved. An April 2025 executive order set a first preference for commercial solutions and faster contracting, and put programs more than 15% behind schedule or over budget up for cancellation.
The investors moved. European rules were clarified to stop excluding defense from sustainable funds, and new companies now raise at scale: Castelion at $13B, Shield AI at $12.7B, Saronic at $9.25B.
Where the value pools
| Layer | My call | Why |
|---|---|---|
| Autonomy and drones | Concentrates | A few leaders absorb most of the capital; late entrants pay up for a crowded race. |
| Advanced manufacturing | Most value | Factories, machining and additive production are the binding constraint on everything else. |
| Munitions, motors and energetics | Most value | Mass-producible weapons and propulsion, where supply is thin and demand is large. |
| Counter-drone | Underpriced | The cost-exchange problem every military now has, with programs being funded. |
| Legacy primes | Incumbents keep it | Programs of record still run through them; start-ups win about 1% of contract dollars. |
Who is winning so far
| Company | What it does | Stage and last round |
|---|---|---|
| Anduril | Software-defined prime: autonomy, Lattice command software, Arsenal-1 factory | $5B at $61B, led by Thrive and a16z May 2026 · thenextweb.com |
| Castelion | Low-cost, mass-producible hypersonic missiles | $1B Series C at $13B Aug 2026 · orrick.com |
| Shield AI | Autonomy software and unmanned aircraft | $2B at $12.7B Mar 2026 · capitalbrief.com |
| Saronic | Autonomous surface vessels and new shipyards | $1.75B at $9.25B, led by Kleiner Perkins Mar 2026 · aap.com.au |
| Hadrian | Automated precision factories for aerospace and defense | $1.37B at $7.87B Aug 2026 · thenextweb.com |
| Helsing | European defense AI and autonomous systems | $1.8B Series E at $18B Jul 2026 · euro-sd.com |
| Chaos Industries | Radar to detect drones, missiles and aircraft | $510M at $4.5B, led by Valor Nov 2025 · dronexl.co |
| Mach Industries | Drones, strike systems and solid rocket motors | $600M extension at $3.7B Sep 2026 · techcrunch.com |
| Ursa Major | Solid rocket motors and hypersonic propulsion | Going public in a $2.3B deal Aug 2026 · breakingdefense.com |
Where I would write a seed check
Manufacturing-focused defense deals nearly doubled in 2025. Every new prime needs suppliers that can produce faster.
Early examples: Firestorm Labs, DivergentThe threat now comes by the million. Programs are funded and the cost-exchange problem is unsolved.
Early examples: MaraRocket motors and propellants are a known shortage, and late-stage valuations show buyers will pay for capacity.
Start-ups stall between prototype and program of record. Shorter qualification is worth paying for.
How it plays out
- 2027
Congress decides on the 2027 budget and a second reconciliation package, which determines the drone and missile defense lines. Ursa Major's listing tests public appetite.
- 2028
Anduril's Arsenal-1 runs in volume, the first proof that a venture-backed prime can manufacture at scale.
- 2031
My expectation: start-ups hold several programs of record, and their share of Pentagon contract dollars is a multiple of today's 1.3%.
- 2036
My expectation: the industrial base looks different, with more of its factories, shipyards and motor plants built since 2025 by companies that did not exist in 2020.
As of Oct 2026
Risks
- Budget instability: fiscal 2026 opened with a 42-day funding lapse, and 2027 again faces stopgap bills. everycrsreport.com
- Reliance on one-time money: about 97% of missile defense's 2027 request depends on a reconciliation package Congress has been reluctant to pass. defensescoop.com
- The valley of death: start-ups won 1.3% of Pentagon contract dollars in 2025. cnas.org
- Concentration: one company took a third of second-quarter 2026 defense-tech funding. tectonicdefense.com
- Execution gaps: the Replicator drone initiative aimed for thousands of systems by August 2025 and fielded hundreds. everycrsreport.com
Sources
- CRS, FY2026 defense appropriations
- KSWO, Pentagon $1.5T budget for 2027
- NATO, The Hague summit declaration (Jun 2025)
- European Commission, SAFE defense loans
- Defense News, Defense tech funding 2025
- CSIS, China's dominance in shipbuilding
- Executive Order 14265, Modernizing defense acquisitions
- CNAS, From innovation ecosystem to industrial base
- DefenseScoop, Golden Dome funding (Aug 2026)
- The Next Web, Anduril at $61B (May 2026)
- The Next Web, Hadrian Series D (Aug 2026)
- CRS, Replicator initiative
Nothing here is investment advice: these are my own notes on public information, written to sharpen judgment, not to recommend buying or selling anything.