welcome back to the INFINITE INNOVATION NEWSLETTER – Pre‑IPO Weekly, where we hunt for the next generation of mission‑critical compounders before they ever ring a bell on the NYSE or Nasdaq.
This week: why DEFENSE TECH is quietly becoming one of the most important innovation platforms of the 2020s – and how a new crop of software‑native defense primes is rewriting the rules of military power, procurement, and pre‑IPO return potential.
Why Defense Tech Is Having Its “AI iPhone Moment”
For most of the 2010s, the defense complex was the definition of “uninvestable innovation” – low growth, capped margins, glacial procurement, and incumbents that looked more like regulated utilities than tech companies. That changed after 2022 as three forces converged:
Ukraine became the world’s proving ground for cheap drones, autonomy, and electronic warfare at scale.
The Indo‑Pacific naval gap versus China forced the U.S. to look beyond $13B carriers toward unmanned, attritable systems above and below the waterline.
Drone proliferation in the Middle East and beyond made it obvious that a $10k quadcopter can threaten a $3M missile battery or a $2B ship – flipping the cost curve of warfare.
The result: U.S. defense spending has pushed past the trillion‑dollar mark, and a rising share of that incremental budget is flowing to startups solving autonomy, AI‑enabled command‑and‑control, and counter‑drone at software‑like margins rather than legacy cost‑plus contracts.
For long‑term innovation investors looking to create generational wealth, this is key: we’re not just funding more missiles and steel; we’re funding the software operating system for modern warfare – and those platforms have the potential to scale across allies, domains, and ultimately dual‑use commercial markets.
The New Primes: Anduril, Shield AI, Saronic
Below is how I’m thinking about the three flagship private names that are functionally becoming “new primes” around autonomy and software.
1. Anduril – The Software‑First Prime
Anduril is the closest thing we’ve seen to a next‑gen Lockheed or RTX built from first principles as a software company. Founded in 2017 by Palmer Luckey, the company has scaled from roughly $150M in revenue in 2021 to an estimated ~$2.1–2.2B in 2025, more than doubling year‑over‑year. Recent reporting pegs its valuation around $60–61B after a $5B raise, effectively pricing it like a high‑growth, AI‑infrastructure platform rather than a traditional prime.
Key points:
Product stack spans drones, counter‑drone, border surveillance towers, underwater systems, and other autonomous platforms.
Roughly half of revenue is tied to drone and counter‑drone operations, with large customers in U.S. Special Operations Command plus programs with the Army, Air Force, Navy, and Customs & Border Protection.
The crown jewel is Lattice, an AI‑powered command‑and‑control system that networks sensors, drones, missiles, and crewed platforms into a single software layer – shifting the business toward recurring “platform plus hardware” economics rather than one‑off defense hardware sales.
At a ~60B+ valuation on a little over $2B of revenue, Anduril screens expensive against primes trading ~1–2x sales – but it’s growing >100% versus low‑single‑digit growth for the legacy complex and indexing aggressively to the fastest‑growing buckets in the DoD budget (autonomy, counter‑UAS, AI C2). If they hit the implied ~$4B+ forward revenue discussed in private markets and continue to win large enterprise contracts (e.g., a decade‑long U.S. Army contract with a $20B ceiling), the multiple compresses into high‑growth SaaS territory rather than pure hype.
On IPO timing, Luckey has publicly indicated that he’s “definitely” pro‑IPO, and with funding rounds now sized more like pre‑IPO balance sheet fortification than early‑stage venture, a 2026–2027 listing window is realistic.
2. Shield AI – The AI Pilot for Airpower
If Anduril is trying to be the new prime across domains, Shield AI is building the operating system for autonomous military aviation. The company’s core product, Hivemind, is an AI “pilot” that allows drones and jets to perceive, decide, and fly missions autonomously in GPS‑denied and comms‑degraded environments.
What stands out:
Platforms: Nova quadcopter, V‑BAT VTOL drone, and X‑BAT, a VTOL stealth fighter drone piloted entirely by Hivemind.
Operational proof: Hivemind‑powered systems have flown combat sorties and participated in DARPA‑backed simulated F‑16 dogfights, reportedly beating human pilots – a huge signaling event for autonomy in air combat.
Scale and valuation: third‑party research pegs 2024 revenue around ~$260–270M, with investors and insiders estimating ~$300–400M run‑rate into 2025. A recent $2B financing package values the company around $12.7B, more than doubling its March 2025 valuation of ~$5.3B.
Strategically, Shield AI has a slightly different go‑to‑market versus Anduril: in addition to selling directly into the Pentagon, it has partnered with primes like Lockheed and RTX to license Hivemind as the autonomy layer inside their experimental aircraft and drone programs. That positions Shield as a potential “arms dealer to the arms dealers,” in the same way Nvidia is infrastructure for hyperscalers rather than a direct competitor.
3. Saronic – Autonomous Ships for a New Naval Order
On the maritime side, Saronic is rapidly becoming the default bet on unmanned surface vessels (USVs) at scale. With the U.S. shipbuilding base unable to match China ship‑for‑ship on $2B+ destroyers and $13B supercarriers, the Navy increasingly needs attritable, autonomous vessels to extend its reach, conduct ISR, lay and clear mines, and serve as weapons platforms at a fraction of the cost and risk.
Saronic is building a family of autonomous boats from small six‑foot scout craft (Spyglass) to 40‑metric‑ton surface combatants, with aggressive production targets. The company:
Was founded in 2022 and ramped from essentially zero to hundreds of millions in contracted revenue within a few years, including a ~$392M Navy contract.
Recently closed a $1.75B Series D at a $9.25B valuation, more than doubling its prior ~$4B mark, and is investing heavily in supply chain and shipyard capacity to hit WWII‑style production rates by 2027.
Where Anduril sells “brains + bodies” across domains and Shield AI sells “brains for anything that flies,” Saronic is the purest play on autonomous surface vessels – directly addressing the Indo‑Pacific naval gap and the risk of sending $2B crewed ships into constrained waterways like the Strait of Hormuz.
Quick Snapshot: Three Core Private Names
Company | Domain focus | 2025E revenue | Latest valuation | Core edge |
|---|---|---|---|---|
Anduril | Multi‑domain (air, land, sea, border) | ~$2.1–2.2B | ~$60–61B | Lattice AI C2 + full‑stack autonomous systems |
Shield AI | Air / aviation | ~300–400M est. | $12.7B | Hivemind AI pilot for drones and jets |
Saronic | Maritime (surface) | Ramp from near‑zero to $200M+ 2025 implied | $9.25B | Attritable autonomous ships at scale |
(All figures are approximate, based on private‑market and media estimates.)
Beyond the Big Three: Drones, Counter‑Drone, and Global Players
The ecosystem is much broader than three U.S. names. A few segments are particularly interesting for Pre‑IPO Weekly readers:
European and Ukrainian platforms – Companies like Germany’s Quantum Systems have become deeply embedded in Ukraine, iterating rapidly on fixed‑wing reconnaissance drones under real combat conditions and building dual‑use businesses that can sell into agriculture and infrastructure once the war ends. Ukraine‑based players such as Swarmer recently went public, with their IPO rally underscoring retail and institutional appetite for “pure play” drone exposure linked to real‑world battlefield performance.
U.S. domestic drones – With Chinese DJI drones now banned from U.S. government and military procurement, domestic players like Skydio (referred to as Skyo in the transcript) have seen revenue jump from roughly $100M to ~$180M in a year as federal and state buyers replatform to U.S. systems.
Counter‑drone / directed energy – Perhaps the most under‑owned theme: high‑power microwave and laser systems that can take out swarms of cheap drones at a cost per engagement lower than the drones themselves. Startups like Epirus are building high‑power microwave solutions that can fry drone electronics over an area, flipping the economics from “$3M missile vs. $10k drone” to “$1k pulse vs. $10k drone.” Counter‑UAS may ultimately be a larger budget line than drones themselves in rich democracies that must defend fixed infrastructure.
At a macro level, global venture funding into defense tech has surged into the tens of billions per year, with Anduril, Shield AI, Saronic, and others capturing outsized rounds and valuations as investors search for scarce exposure to real revenue growth backed by government budgets rather than consumer demand cycles.
How This Plays for Pre‑IPO Investors
Let’s talk positioning as an innovation‑focused, long‑duration investor.
1. Structural tailwinds vs. political risk
We are likely in a “Goldilocks” window where geopolitical conflict (Ukraine, Middle East, Indo‑Pacific), a defense‑friendly administration, and a renewed focus on industrial policy all align in favor of higher defense R&D and procurement, especially for autonomy, AI, and space. That said:
A change in administration or congressional sentiment can slow DoD budget growth and, more importantly, stall reforms that make it easier for startups to win and scale programs of record.
Congressional inertia still favors legacy primes that employ tens of thousands of workers across many districts, meaning startups must either become primes themselves (Anduril), sell into primes (Shield AI), or carve out “net new” budget lines (Saronic, Quantum Systems).
2. Business model and margin structure
Legacy primes operate under cost‑plus contracts with margins intentionally capped around ~10%. The most interesting defense tech startups are explicitly not playing that game:
They build on their own balance sheet and sell products at a fixed price with software‑like gross margins in the 30–40% range.
They attach recurring software (Lattice, Hivemind Enterprise) to hardware platforms, creating “razor‑and‑blades” economics where the software layer can scale far beyond any single platform.
In return, they accept higher risk: more capital intensity upfront, higher dependence on continued program renewals, and exposure to political swings. But if you believe that autonomy and AI‑first C2 will be the dominant spend buckets for the next 10–20 years, these business models look much closer to enterprise software than to traditional defense manufacturing.
3. Access: how do we actually get exposure?
Right now, most of the pure‑play upside is locked in the private markets:
Direct secondary exposure to names like Anduril, Shield AI, and Saronic sits inside specialist funds, SPVs, and newer listed “venture for all” vehicles (e.g., certain public funds whose underlying holdings include high‑profile defense tech stakes).
On the public side, Palantir remains the closest proxy for a software‑first defense data and AI platform, with meaningful DoD and IC exposure and a business model that rhymes with what Anduril is building.
A handful of defense‑tilted ETFs and listed venture portfolios (including newer vehicles like VCX, which saw a rapid post‑listing repricing) can offer basket exposure but come with their own liquidity and valuation dynamics to underwrite carefully.
My base‑case roadmap as Coach Kyle:
Track Anduril’s march toward IPO and treat it as a potential anchor position in the “New Prime” bucket if pricing is rational.
Selectively use secondaries or structured products to gain small, high‑conviction positions in Shield AI and Saronic where governance, information rights, and pricing make sense relative to their growth and contract backlogs.
Complement private defense tech with public liquid proxies (Palantir, defense‑AI ETFs, and select primes leaning hardest into autonomy and C2), while keeping position sizes aligned with policy and concentration risk.
