Construction / Aerospace Manufacturing · Deep dive

Hadrian Automation

Software-defined precision-parts factories for aerospace and defense — Torrance, Mesa and a 2.2M-sq-ft Navy submarine plant in Alabama — now valued at $7.87B on a Series D that only prices well if the factories-as-a-service model actually runs at prime scale.

emerging

The question that decides it: Does Hadrian's software-defined-factory model — Opus plus multi-axis CNC cells routed and inspected under a single control plane — hit the throughput and gross-margin curves needed to turn one Torrance line into a network (Mesa live in Jan 2026, a 2.2M-sq-ft Alabama submarine plant announced Mar 2026, more to come on the Series D) that serves prime-contractor demand at per-part cost the primes cannot beat in-house? Or do the primes vertically integrate — Anduril already builds much of its own hardware; RTX and Lockheed have deep in-house machining; a Navy-cofunded Alabama plant is de facto a captive shipbuilding node — while the ~15,000 traditional shops adopt just-enough automation (pallet changers, off-the-shelf CAM, cobots) to close the gap before Hadrian reaches network economics? Answer conditions: (a) at least one publicly disclosed prime contract at unit-price parity with in-house machining plus a two-year utilization ramp at Mesa above 70%; (b) Opus deployed inside a third-party prime's factory (Hadrian's stated factories-as-a-service pitch) with a named customer; (c) gross margin disclosure that supports a $7.87B mark on manufacturing multiples, not software multiples.

My take ★★ Very interesting ★ Interesting − Maybe ✗ Not interesting

Snapshot

Hadrian Automation makes precision machined parts for rockets, missiles, satellites, jets and — as of 2026 — Virginia- and Columbia-class submarines. It runs an original ~100,000-sq-ft factory in Torrance, California, a new 270,000-sq-ft plant in Mesa, Arizona (Factory 3, opened January 29, 2026) and a 2.2M-sq-ft Navy-cofunded plant in Cherokee, Alabama (Factory 4, opened March 20, 2026) whose stand-up carries a $2.4B public-private price tag. On August 6, 2026 Hadrian closed a $1.37B Series D at a $7.87B post-money — co-led by WCM Investment Management, Washington Harbour Partners, Valor Equity Partners, 137 Ventures, Baillie Gifford, with JPMorgan Chase’s Strategic Investment Group as anchor co-lead and a16z American Dynamism, Lux Capital and Altimeter following on. The pitch that made those marks work is not that Hadrian is a good machine shop; it is that its proprietary software stack, Opus, turns high-mix, low-volume aerospace machining into a repeatable industrial process that a prime cannot beat with its own captive shop.

Founding story

Chris Power arrived in the US from Australia in 2019 at 28, a self-taught college dropout with no manufacturing background, and — as he tells it in interviews with Not Boring, Sourcery and Forbes — spent his first months touring old machine shops in Southern California. He concluded the US precision-machining base was 30-40 years behind, geriatric on labor, and structurally incapable of supplying the ramp aerospace and defense customers needed. Hadrian was incorporated in Torrance in 2020 on that thesis: rebuild the sub-tier as a software company that happens to own CNC cells.

The founding cap table set the tone. Katherine Boyle (a16z, American Dynamism) and Brandon Reeves (Lux Capital) took the pitch early and later took board seats when the Series A/A-Prime closed in March 2022 — Boyle’s American Dynamism thesis and Founders Fund’s defense-and-industrial orientation gave Hadrian a policy narrative before it had a second factory. Power’s biggest operational hire, Chris Baker, ran Dragon capsule production at SpaceX; Power has said publicly it took six months to recruit him. Hadrian has since layered on nearly a dozen former SpaceX operators. No academic or engineering co-founder has been publicly credited; the story is Power the CEO plus an ex-SpaceX manufacturing bench.

The frame Power leans on is binary. In a June 2025 Sourcery interview he described Hadrian’s outcome distribution as “$10B or zero” — Series D at $7.87B has effectively closed the distance to the first number on paper without answering the underlying operational question.

How it works

A Hadrian factory looks less like a job shop than a warehouse of five-axis CNC machining centers (bought from third-party OEMs — Hadrian is explicit that it does not build machines), Kuka robotic arms, palletized fixturing, and in-line coordinate-measuring inspection, all orchestrated by an internal software layer called Opus. Opus is the thing being sold. When a customer request lands, a designed-for-manufacturing (DFM) module reviews the CAD, quotes a price and lead time, decides whether Hadrian should accept the job, and — if yes — auto-generates the CAM tool paths, routes the raw stock to a specific cell, sequences fixturing and tool changes, and pushes each finished part through robotic inspection before packing.

The claims Hadrian makes for that stack are unusually specific. CAM programming time cut by roughly 90%, machine cycle time cut by roughly 60% versus a traditional shop (Contrary Research, 2024). Parts produced 10 times faster and 40% more efficiently than legacy manufacturers (Forbes, August 2024). Around 90% of the physical labor per part is done by machines; the remaining human role is closer to a system operator loading raw stock and pressing go than a journeyman machinist writing G-code. That last part is central to the labor pitch: Opus is designed to train a factory technician in 30 days or less, letting Hadrian hire from a labor pool ten times larger than the shrinking pool of skilled machinists whose median age (Bureau of Labor Statistics, 2024) is 45.7 and whose over-55 cohort is roughly 31% of the ~309,000 US machinists working today.

The differentiation, in other words, is not the machining physics — five-axis CNC, TRISO fixturing, KUKA robotic loading are all off-the-shelf — but the control plane on top. Whether Opus produces the promised throughput and margin curve at the scale of Mesa and Alabama, not just Torrance, is the operational question the Series D is priced against.

Product and business overview

Three product lines and, since July 2025, a fourth strategic vector.

Contract manufacturing is the original business: Hadrian quotes and ships aerospace- and defense-grade machined parts to customers who send it a print. Named customers include Anduril (formal strategic partnership June 2023), SpaceX, Relativity Space, RTX and Lockheed Martin; the customer list on the Hadrian site and in Forbes’ 2024 profile skews toward primes and neo-primes rather than commercial aerospace.

Full product manufacturing, announced with the Series C in July 2025, extends Hadrian from parts to assemblies and finished products — precision welding, casting and additive processes stitched onto the machining base. This is the wedge that lets Hadrian book more revenue per program without adding a new customer.

Factories-as-a-service is the newest wedge and the most strategically interesting: Hadrian will stand up Opus-run production inside a prime’s own facility. It is a way to convert the primes from customers into distribution — but it also concedes that the physical footprint of the primes’ existing plants is the constraint, not the software.

The Alabama submarine plant, announced March 20, 2026, is the strategic-vector step. A 2.2M-sq-ft facility in Cherokee, Alabama, built out under a $2.4B public-private structure (~$1.5B private, ~$900M federal) to mass-produce Virginia- and Columbia-class components. It will not reach full-rate production for roughly two years. This is a program-of-record commitment on a scale Hadrian did not have before, and it locks a significant slice of its capital to a single Navy program.

Business model and pricing

Two revenue streams sit inside the customer contract: a per-part price for parts shipped and, on the factories-as-a-service side, what Hadrian frames publicly as a software-plus-services relationship — Opus deployed inside a third-party plant with associated implementation, licensing and outcome-sharing. Public per-part pricing does not exist; the DFM module produces bespoke quotes on request. Revenue is booked on shipment for the contract business and on a program milestone / cost-share basis for the Alabama-scale work.

Contrary Research’s 2024 breakdown tracks Hadrian’s stated ARR from roughly $1M in 2022, when it began collecting recurring revenue, to more than $35M in 2024. Hadrian itself told Fortune in October 2024 it expected to hit roughly $30M in revenue by year-end 2024, and has publicly claimed 10x year-over-year revenue growth since the February 2024 Series B. Full-year 2025 and 2026 revenue is not disclosed. The company has not published gross margin, factory utilization, or per-part unit economics — which is the diligence gap the Series D valuation makes load-bearing. At $7.87B post-money on the last plausible ARR datapoint (~$35M in 2024) the multiple only reconciles as a software-company mark on the Opus layer plus a bet on the factories-as-a-service and program-of-record vectors. On aerospace-manufacturing comps (Xometry trades near 4x revenue; Protolabs near 2x) the mark implies revenue several times higher than the last disclosure.

Traction over time

Parts shipped: “thousands per month, hundreds of unique parts” (Tracxn, 2026). Utilization by factory: not disclosed.

Market analysis

The addressable pool is enormous and the alternatives are structurally weak, which is the entire investor thesis. The US machining and precision-parts industry is a ~$40B/year cottage business run through roughly 15,000 job shops of a few dozen employees each, most of them second- or third-generation and struggling with succession. Layered on top: a defense industrial base that DoD, CSIS and every congressional study since 2018 has called structurally under-capacitized — submarine build cadence half of what the Navy’s shipbuilding plan requires, missile-tube stocks depleted after Ukraine, tactical airframe parts sole-sourced through single tier-two shops. The July 2025 Trump-administration executive orders on defense industrial policy, the CHIPS-and-Science-adjacent industrial-strategy funding, and the March 2026 Navy co-funding of Hadrian’s Alabama plant are all in the same direction of travel.

The bull case, articulated by a16z American Dynamism and Founders Fund and reflected in Not Boring’s 2022 Hadrian piece, is that the mix of retirement, tariffs, and program-of-record demand creates a once-in-a-generation opening for a consolidator that combines modern software with physical plant. The bear case is that the same demand pulls the primes into deeper vertical integration (Anduril bringing infrared imaging in-house via the AIRS acquisition is the template), that DoD program-of-record work moves slowly and is priced on cost-plus, and that traditional shops adopting off-the-shelf CAM and pallet-changers close most of the productivity gap without a software rewrite.

Competitive intel

The competitive set sorts into three tiers, and Hadrian’s angle against each is different. See the frontmatter for the enumerated list; the shape of the fight:

The fragmented incumbent. The ~15,000-shop machining base is the market Hadrian is trying to consolidate. Its structural advantages against Hadrian are proximity and price for one-offs. Its structural weakness is the age curve: with roughly 31% of US machinists over 55 (industry data, 2025) and ~10% annual attrition, retirement alone shrinks the sub-tier faster than replacement hiring. Hadrian’s answer is that Opus makes a 30-day-trained technician competitive with a 20-year machinist on repeatable production runs. That is the operational claim the Mesa and Alabama ramps have to prove.

The public marketplaces and prototype shops. Xometry (NASDAQ: XMTR, ~$604M TTM revenue and ~$2.4B market cap as of October 2025) and Protolabs (NYSE: PRLB) are the public comps and the alternative every buyer will price against. Xometry attacks Hadrian on breadth and speed of quote via an asset-light network; Protolabs on prototype-and-short-run turnaround via owned capacity. Neither is chasing the flight-hardware, ITAR/AS9100-certified, program-of-record wedge that is Hadrian’s home. Fictiv adds a global (China-inclusive) network that DoD work cannot use.

The strategic risk group — vertical integration by primes. Anduril is the template: its acquisition of cooled-IR maker AIRS, its extensive in-house sensors and propulsion work, and its explicit vertical-integration playbook mean that every part class Anduril insources caps Hadrian’s TAM at that customer. RTX and Lockheed run large captive machining and forging operations that they will not readily hand over. The factories-as-a-service pitch — Opus deployed inside a prime’s plant — is Hadrian’s answer, but it converts Hadrian into a software-and-services vendor sitting inside a customer whose incentives are to internalize the IP as fast as possible.

Where Hadrian actually beats each: on aerospace/defense specificity against Xometry/Protolabs; on modern software and labor economics against traditional shops; on capital scale and Navy backing against the machine-shop base. Where it loses: on breadth (Xometry), on price for commercial parts (Fictiv), and on strategic control at any prime willing to insource.

History and evolution

• 2020 — Hadrian founded in Torrance by Chris Power.
• Apr 2021 — $9.5M seed led by Founders Fund and Lux Capital; Chris Baker (ex-SpaceX Dragon) recruited over a six-month period.
• Mar 2022 — $90M Series A/A-Prime, split between Lux Capital (A) and Andreessen Horowitz (A-Prime); Katherine Boyle and Brandon Reeves to board. Torrance factory expanded from 20,000 sq ft to 100,000 sq ft.
• Jun 2023 — Strategic partnership with Anduril announced.
• Feb 2024 — $117M Series B (a16z, Lux Capital, Founders Fund).
• Aug 2024 — Forbes names Hadrian a “Next Billion-Dollar Startup”; company estimates ~$30M revenue for full-year 2024 and cites ARR growth from ~$1M in 2022 to $35M+ in 2024.
• Oct 2024 — Hires former Blue Origin / Rocket Lab-affiliated finance leader as CFO (Fortune, Oct 29, 2024).
• Jul 17, 2025 — $260M Series C (Founders Fund + Lux, a16z, Altimeter). Announces Mesa Arizona Factory 3 (~270k sq ft, $200M invested, up to 350 jobs) and full-product manufacturing (welding, casting, additive). Factories-as-a-service explicitly framed.
• Jan 29, 2026 — Mesa Factory 3 opens (ribbon cutting); capacity framed at up to ~$200M in revenue.
• Mar 20, 2026 — Cherokee, Alabama Factory 4 opens: 2.2M sq ft, $2.4B public-private ($1.5B private + $900M federal), Virginia/Columbia-class submarine components; ~1,000 jobs at full run rate; ramp to full production ~2 years.
• Aug 6, 2026 — $1.37B Series D at $7.87B post-money.

The stumbles that are not in the deck: revenue and gross margin disclosure has narrowed as valuation has expanded (nothing between the ~$35M ARR 2024 datapoint and the $7.87B mark); Glassdoor’s ~3.7-of-5 rating on 15+ reviews (accessed 2026) surfaces a recurring theme of an assertive, high-turnover culture with organizational whiplash — one review characterized the CEO as micromanaging and the executive team as “unbelievably unorganized” with poor communication; and the Alabama plant, while a marquee national-security win, is de facto a captive submarine node whose ramp is on the Navy’s timeline, not Hadrian’s.

What people say

The case for. Investor and trade-press framing is unusually consistent: Hadrian is the physical-AI thesis in cleanest form. Not Boring’s “Ex Machina Ad Lunam” (2022), a16z’s American Dynamism note, Forbes’ Next Billion-Dollar profile (August 2024), Sourcery’s “$10B or zero” interview (2025), Breaking Defense’s factories-as-a-service coverage (July 2025), and Contrary Research’s 2024 breakdown all tell the same story — a founder with unusual clarity about the industrial-base problem, an operations bench recruited out of SpaceX, a software layer with specific and testable throughput claims (90% CAM time reduction; 60% cycle time reduction), and named customers spanning Anduril, SpaceX, Relativity, RTX and Lockheed. The Navy co-funding of the Alabama plant is the strongest single external endorsement any private manufacturing startup has secured in this cycle. Employees who thrive at Hadrian — Glassdoor’s positive minority — describe an “assertive culture” with real ownership and mission clarity.

The complaints. Glassdoor (accessed 2026, 15+ reviews, 3.7/5, 58% would recommend) surfaces a consistent bear read on culture: micromanaging CEO, an executive layer characterized as unorganized and communication-poor, no published performance metrics or deliverables inside the org, long hours as the default, and a pattern of employees either thriving or churning fast. On the business, Washington Technology and Manufacturing Today (both August 2026) note that the Series D was priced with revenue, margins and utilization all undisclosed — an unusual gap at a $7.87B mark. Aerospace-analyst commentary consistently flags that scaling from new product introduction to certified high-yield volume is where capital-intensive manufacturing startups historically die. And the strategic risk from vertical integration is real and named: Anduril’s AIRS acquisition and the primes’ in-house machining have material overlap with Hadrian’s product surface.

Outlook: the open question

The question is whether Hadrian’s Opus-driven, software-defined-factory model produces the throughput and gross-margin curve needed to turn one Torrance line into a network — Mesa live, Alabama ramping, more sites priced into the Series D — that primes cannot beat by insourcing, before either (a) the primes vertically integrate the part classes Hadrian is best at (Anduril’s template) or (b) the ~15,000-shop traditional base adopts just-enough automation to close the gap on retirement-driven consolidation.

Bull case holds if: Mesa reaches sustained utilization above 70% within 24 months of the January 2026 ribbon-cut; Alabama hits its 2-year ramp to full-rate production on the Navy’s timeline without cost overruns Hadrian has to absorb; at least one major prime signs a firm-priced contract at or under its in-house make-cost on a program-of-record part; Opus is deployed inside a third-party prime’s factory as the factories-as-a-service pitch requires; and gross margin disclosure at any point in 2027 supports the $7.87B mark on a plausible blend of hardware and software multiples. Under those conditions, Hadrian is the anchor consolidator of the defense industrial base and a straight line to a public listing.

Bear case holds if: revenue, margin and utilization stay undisclosed through 2027; Anduril insources more part classes and RTX/Lockheed continue to prefer their captive shops; the Alabama plant’s ramp slips; the ~$35M-ARR-in-2024 datapoint is not followed by a public number that supports the mark on hardware multiples; and the traditional shop base adopts off-the-shelf CAM and cobots fast enough to keep price parity on the mid-mix. Under that scenario, Hadrian is a well-capitalized specialty manufacturer with a defense-industrial-base narrative priced multiple turns ahead of what any manufacturing comp — Xometry at ~4x revenue, Protolabs closer to 2x, Divergent private — supports.

Every input Hadrian controls it has moved on aggressively: four sites across three states in six years, a Navy co-funded plant, ARR from $1M to $35M+ in two years, and now a $7.87B mark. Every input it does not — prime insourcing decisions, DoD program timelines, machinist retirement’s actual pace, gross-margin discipline at scale — is on someone else’s clock. At $7.87B, the mark demands both software-company margins and factory-network scale, which is a combination no aerospace-manufacturing incumbent has ever demonstrated. That is the investment question.

How a challenger would attack it

Attack the capital model, not the software. Hadrian has sunk $1.85B+ into owned factories, one of which — the 2.2M-sq-ft Alabama plant — is a two-year ramp chained to a single Navy program’s timeline. A challenger builds Opus-without-the-buildings: an orchestration layer (auto-CAM, routing, robotic inspection integration) licensed to the existing 15,000-shop base, whose hardware — five-axis CNC, pallet changers, cobots — is the same off-the-shelf gear Hadrian itself buys. Hadrian’s own thesis concedes the physics aren’t proprietary; a software-only rival turns the fragmented incumbent into its distribution instead of its competition, arming retiring shop owners rather than replacing them. The second wedge is disclosure: Hadrian priced a $7.87B round with revenue, margin and utilization all withheld — a challenger that publishes per-part prices, lead times and audited quality rates makes primes’ procurement teams ask why the $8B vendor won’t. Third, poach the churn: Glassdoor describes a micromanaged, high-turnover culture with “unbelievably unorganized” leadership — the trained ex-SpaceX operations bench that walks out is the exact talent a rival needs. And the factories-as-a-service pitch is attackable at its seam: any prime that wants Opus inside its plant would rather buy it from a vendor that doesn’t also compete for its parts volume.

Same playbook, new buyer

Run the software-defined machine shop for buyers Hadrian’s cap table forbids. Hadrian is structurally locked into US defense: Founders Fund/a16z American Dynamism narrative, ITAR work, Navy co-funding, prime customers. That leaves the entire commercial precision-machining market — medical devices, semiconductor equipment, energy turbomachinery, robotics OEMs — facing the same geriatric shop base (median machinist age 45.7, 31% over 55) with no consolidator, and these buyers don’t need AS9100 program-of-record overhead, just tight tolerances at predictable lead times. Fictiv attacks it from a China-inclusive network; nobody attacks it with owned automated US capacity. The second shift is allied-nation defense: Australia (Power’s home market, AUKUS submarine demand), the UK, Japan and Korea all face worse machinist demographics and have no Hadrian equivalent — and Hadrian, with its capital committed to Torrance, Mesa and Cherokee and its ramp obligations to the US Navy, cannot open a fifth factory abroad this decade. Either buyer lets a new entrant prove the Opus-style model on faster commercial qualification cycles, then arrive at defense with margins already demonstrated — the disclosure Hadrian still hasn’t made.

Sources and further reading

• Hadrian raises $260M to build out automated factories for space and defense parts (TechCrunch, July 17, 2025)
• Defense manufacturing startup Hadrian closes $260 million funding round led by Peter Thiel’s Founders Fund (CNBC, July 17, 2025)
• Exclusive: Hadrian raises $1.37B amid surging defense-production demand (Axios, August 6, 2026)
• Hadrian valued at nearly $8 billion after fresh funding as money pours into defense tech (CNBC, August 6, 2026)
• Hadrian Raises $1.37B Series D to Build Highly Automated Factories to Accelerate America’s Industrial Renewal (PR Newswire, August 2026)
• Next Billion-Dollar Startups: Hadrian’s High-Tech Factory Makes Parts For Rockets And Fighter Jets 10 Times Faster (Forbes, August 15, 2024)
• How startup Hadrian plans to take over the defense manufacturing world (Breaking Defense, August 2024)
• Manufacturing startup Hadrian to expand to Arizona, and into defense primes’ own factories (Breaking Defense, July 2025)
• Hadrian Opens Advanced Manufacturing Facility in Alabama to Support Columbia- and Virginia-Class Submarine Production (PR Newswire, March 20, 2026)
• Hadrian ribbon cutting — Mesa (Greater Phoenix Economic Council, January 29, 2026)
• Report: Hadrian Business Breakdown & Founding Story (Contrary Research, 2024)
• Hadrian: Ex Machina Ad Lunam (Not Boring, 2022)
• Investing in Hadrian (a16z, 2022)
• Anduril Industries and Hadrian Announce Strategic Partnership (Anduril, June 2023)
• Exclusive: Chris Power of Hadrian: $10B or Zero, Revitalizing American Manufacturing (Sourcery, 2025)
• Hadrian Reviews (Glassdoor) (Glassdoor, accessed 2026)
• Hadrian Automation (Wikipedia, accessed August 2026)