Hey, it’s Jeff Crusey, your friendly neighborhood VC here at PA.
I tend to look for hard physics before the market admits (or realizes) it's a market. When I wrote the first check to Castelion and led Xona’s seed, it was because I had developed very specific convictions on hypersonics and navigation, respectively. Half a decade later, the world (and markets) caught on.
Today I’m doing something new: open-sourcing the process of building this conviction in a new domain.
I’ve been thinking a lot about what’s going on in the ocean…
IN THIS WEEK’S EDITION:
🌊 The subsurface leash
🦭 The Per Aspera seal, revealed
🐐 HALO, Buffett, Midjourney Medical, and more
🏆 2 Factory 002’s + more community wins
🥹 Slick Six: It’s not goodbye, it’s see you later…
Forwarded this? Subscribe to Per Aspera
This March, the Defense Innovation Unit and the Navy picked Anduril’s Dive-XL for a program called CAMP. It imposes hard requirements for systems that are already hard enough to build, in a place that’s already hostile enough as it is: cross 1,000+ nautical miles without surfacing, operate below 200 m, hold a course where GPS can’t reach, and set payloads down on the sea floor. Anduril is now spinning up a line at Quonset Point, Rhode Island to turn out dozens of these things a year.
There’s one requirement that the Navy and DIU conspicuously left out, and we can’t fault them, as it’s an issue that nobody has been able to solve. Once a system like Dive-XL…. dives, you can’t really reach it.
This is the harder half of a problem we’ve spent the better part of two centuries working from the other end.
Then and Now
On August 5, 1858, Queen Victoria sent a 98-word tweet message to President Buchanan across the first transatlantic telegraph cable, which had been completed days earlier.
The first transatlantic telegraph took ~16.5 hours to cross, and the cable that carried it burned out within a few weeks. But the core idea lived on: a signal-carrying core that we armored, buried, and landed at shore stations.
We’ve perfected the pipe (it’s a great pipe sir), evolving from copper and Morse, to copper + coaxial for analog telephony, to glass fiber carrying light (not unlike the shift currently underway in datacenters). Today, a single modern fiber pair can push ~200 TB/s across an ocean.
Beyond any reasonable doubt, we’ve mastered the fixed link: the thing bolted to the bottom, or tethered to a ship. But in our 165 years as a seabed cable-laying species, we’ve been fundamentally unable to solve high‑bandwidth, stealthy wireless at depth.
The moving link has eluded us.
Not for lack of trying, but due to the tricky physics of it all. Anyone who sets out to solve underwater comms is navigating a tough trade space, and headed headlong into a three-way trap:
Saltwater swallows radio. Only the very lowest frequencies make it to a submarine at depth, in one direction, at a few hundreds bits per second, from megawatt shore stations the size of small towns.
Sound carries, but it crawls along (~1,500 m/s) and tops out in the kilobits (e.g., 1992 dial-up speeds on a good day) if you want real range. Therein lies the challenge: optimize for range, and you’ll sacrifice bandwidth.
Light is fast — with blue-green lasers able to push gigabits — but the beam dies past ~200 m and demands a marksman’s aim between two moving platforms in a messy medium. (I’d bet somewhere, right now, an ex Starlinker is working on this — if that’s you, get in touch)
If you’re the United States Navy, you’re also facing an operational trap: talking at all is a tell, because the two ways to get real bandwidth will blow your cover:
Surface and your mast, periscope, or buoy is in view of RF and satellites.
Transmit sound and you are… transmitting sound… by definition, announcing yourself to the ASW ears that are built to hear exactly that.
Fire and forget
So, stealth and connectivity are in direct tension. The depth and silence a dual-use platform so desperately seeks is precisely what’s cutting it off from Command and Control (C2). We work around these constraints by taking C2 out of the loop and turning to autonomy. But the comms challenge means that today's undersea autonomy is largely "fire-and-forget": you program the mission, launch, and hope, because you cannot retask in real time.
And don’t get me wrong — we need more AUVs (autonomous underwater vehicles) — and by goodness, is industry rising to the call for more autonomous, attritable, agile, FBC surface and subsurface vessels. Catch me in a pub, and I could probably rattle off at least two dozen startups (plus the bigco programs) building these systems, just from the top of my head. There’s enough work to go around for these craft builders: ISR, mine countermeasures, ASW, seabed surveying, payload/strike, etc.
But there’s still that one stubborn problem: you can build the robot, but you can't talk to it in real time once it's in the deep.
A Blue-Ocean Opportunity
I do not submerge you with nautical knowledge simply because it’s the latest obtuse obsession of mine, or a nice-to-have science project, but because the ocean is one of the least appreciated, most critical strategic domains of our time.
Conservatively, 95%+ of intercontinental data goes through subsea infrastructure, and a large share of cross-border oil and gas moves through seabed pipelines and offshore infrastructure. All of that is installed and maintained by a shockingly small fleet — on the order of sixty‑odd large, specialized cable‑laying and repair ships, many of them closer to retirement than launch. In a world where the incremental hull, craft, and vessel leaving a shipyard — whether ours or an adversary’s — wants to be autonomous, which way do you think the scales are tipping: toward attackers or defenders?
Indeed, down in the deep, as you’ve probably seen, cables are being cut and pipelines are blowing up. There’s plenty more ‘plausibly deniable’ activity down there that doesn’t make the front page news. (Look no further than my editor Ryan’s dispatch from Taiwan). Russia has an entire naval unit — the secretive, deep-sea GUGI, with assets purpose-built to tap or cut undersea infrastructure.
My call to action (or request for startups, if you will) is for the right team to attack this and build the missing piece. Right now every undersea vehicle (besides exquisite subs) is autonomous by necessity, not by choice. Persistent comms would give us more control, help the West defend its seabed infrastructure, and give the team who cracks it a huge mission and big ($$$) market.
Get in touch if you’ve got a company, product, a plan, or even concepts of a plan in this dept! To going where the signal doesn't yet reach. 🥂
Joy Shin here — Creative Director at Per Aspera. Popping in, because alongside last week's new website launch, we quietly shipped something new: our seal logo. (Created by Branden Maurin, new to the PA team but long-time PA subscriber. You’ll meet!)
We ran Per Aspera for a year without a real logo — ~4-5 designers, hundreds of iterations, thousands of inspos (we even played w/ Virgil Abloh's Sinking Man logo for Off-White). But the right combo never landed. It had to honor the "hard pursuits" of everyone here (you!), feel Americana without relying too much on 50s-60s nostalgia, and respect both our pursuit of the future + foundations in the past.
Where We Landed
As Ryan mentioned in his one-year letter, Per Aspera doesn’t deal in absolutes — and that's reflected in our seal. So, why a seal? We see this work as generational and structural, so we wanted something that departs from trends and eras: both the flat modernity of tech brands and the nostalgia 'trap' of the current Americana movement.
It was designed with contradictions, open to debate — much like the heady conversations inside our ecosystem.
At base, two Sisyphus figures bear the whole thing on their shoulders: condemned by Zeus to roll his boulder for eternity, Sisyphus is often read as a warning about hubris. We keep that caution and embrace the uphill pursuit, but reject the one fixed part — eternal damnation (because, yes, we will win 🇺🇸).
Above them, the shield holds a clock and the Angel of Progress beneath a crowning bald eagle. The Angel (adapted from Columbia in John Gast's 1872 American Progress) carries progress and the question of progress at all costs. The clock represents the tension of time itself — urgency and permanence, no time to waste while building things meant to last.
Ringing it all, fifty stars for fifty states, because we don't exist to serve an elite pocket or a single niche. We're simply a hard-pursuits tribe — drawn from every arena of the country. That's our elite crowd.
In the next few weeks, we will publish a deep dive, but wanted to give you all a sneak peek into our inner thinking. And if you see something else in it — tell us!
001A / SMART MONEY RE-UPS ON ATOMS… Goldman Sachs Alternatives, led by Co-CIOs Philippe Camu and Tavis Cannell, raised $3B+ at the first close of West Street Infrastructure Partners V — hitting 75% of its $4B target in <6 months, with 80% coming from existing LPs re-upping. PE, of late, is looking like a tale of two cities:
Frozen. Traditional buyout is stuck in a rut, with PE dealmaking running ~40% below prior-year levels. Exits are scarce, distributions are slow, and LPs are balking at new commitments as they wait for cash back. Software‑heavy funds in particular are feeling the squeeze as AI erodes the neat SaaS moats they originally underwrote and LPs scrutinize those theses more harshly.
HOT: The HALO trade — Heavy Assets, Low Obsolescence — looks as good as ever. Infrastructure as an asset class offers what institutions want right now: structural (vs. cyclical) demand drivers ✅ … durable & non-discretionary ✅ … (relatively) defensive, contracted cash flows ✅. The institutions already long the physical economy are doubling down — more power to ‘em.
001B / LEST WE FORGET… A tip of the cap to the 🐐, Warren Buffett, who handed the keys to the Berkshire kingdom to Greg Abel on Jan 1. He steps back with the last laugh, as the ‘boring’ trade now branded HALO was a key strategy for building his Berkshire empire. We said as much last fall (#026): the 🐐 averaged ~20% annual returns for six decades by placing bets on the physical backbone of our world — railroads, utilities, housing — while everyone else chased tech. A parting lesson we took from him (our words):
Build long-duration capital pools that think in generations. Point them at the hard, indispensable systems that everything else depends on. Let operators lead instead of financiers. Reinvest relentlessly in capacity, competence, and compounding.
A reconstructed body volume, swept slice by slice (legs)
002 / RENDERS TO RADIOLOGY (?!?)… Midjourney (yes, the Discord-born, image-generating diffusion model company) has had the technology world in a tizzy for the last 48 hours, in the best possible way, with the launch of a new division, “Midjourney Medical.”
Its first product, the Midjourney Scanner, will be a full-body imaging system that uses sound, water, and dense arrays of ultrasound-on-chip transducers to reconstruct sub-millimeter 3D images of your body in 60 seconds. Lots to still sort out…
does it work as advertised
can it rival MRI in clinical settings (while winning on cost)
how do you compensate for skulls/ribs/lungs
what’s the regulatory pathway from wellness spa scan to a cleared diagnostic
…but consider us believers. A huge kudos to Midjourney founder David Holz and his tiny team for taking this B.F.S. (big ‘freaking’ swing 😉)… one that’s less of a stretch than it may seem. Holz is a prior hardware founder (Leap Motion), who — fun fact — got his start as a researcher at NASA Langley!
Watch this space… We could (and will) go deeper on this soon, as the physics/technology/math behind tomography (the ‘T’ in CT scanner) are quite similar to what Ryan’s company Array Labs is doing in space.
In the meantime, any folks in healthcare / medical imaging reading this? Want to offer your take? And everyone, help: Dan Goldin is still looking for an accurate body measurement tool. Apple Watch and Whoop are not working out for him. Write in with your recommendation!
003 / PERMISSION TO GLOW… This week, Helion Energy — the Everett, WA fusion outfit chaired by Sam Altman and backed by Thrive, SoftBank, and Ford Motor’s Bill Ford — reached a big milestone for the category, becoming the first in the world to secure the key regulatory licenses needed to operate a fusion power plant. This buys down regulatory risk as the startup prepares to break ground on Orion, its first commercial plant, and brings the whole sector one step further from “always three decades away.” 🤞⚡🇺🇸
So many friends of the PA community have been winning lately — and nothing makes us prouder. A few we have to shout out:
001 / KELLEN GIUDA (KG CAPITAL). The Pentagon tapped Kellen Giuda, Founder and CEO of KG Capital, for its new Science, Technology and Innovation Board. (If the name Giuda rings a bell: his wife, Michelle Giuda, is also a force in the PA community. A power couple! 🇺🇸🤝)
002 / AARON SLODOV (ATOMIC INDUSTRIES). "Mr. Build, Baby, Build" himself, Aaron Slodov, pulled off a double header this week:
Debuted Atomic Industries' (AI-driven injection molding) Factory 002
Hosted the third Reindustrialize Summit in the great city of Detroit.
This man does not rest! 🙌🙌🙌
003 / ANDURIL & GA DRONE WINGMEN SELECTED. The U.S. Air Force selected both Anduril (FQ-44) and General Atomics (FQ-42) for Phase 1 of its Collaborative Combat Aircraft (CCA) program, covering engineering, development, and initial production of semi-autonomous drone wingmen to fly alongside manned fighters like the F-35. CCA calls for at least 150 aircraft by decade's end.
Left: Anduril (FQ-44); Right: General Atomics (FQ-42)
As GA boss Neal Blue toasted at our launch event: “entrepreneurship and individual initiative … are the characteristics which will preserve the republic which he holds so dear.” 🥂
004 / DIVERGENT’S FACTORY 002. PA Founding Sponsor Divergent announced America's most advanced industrial metal 3D printer, the Monolith One, and the opening of their own Factory 002 — a 430K sq. ft. facility in Long Beach, CA that will 8x annual production output. Dan has said before: “Watching Divergent’s growth is like watching flowers bloom in the springtime.” 🌸🌻🌷🌹
(Editor’s note: the below is best read with Wiz Khalifa’s “See You Again” playing in the background.)
Goodbye "Slick Six". This week, teams demolished Space Launch Complex-6’s legacy towers at Vandenberg Space Force Base. Farewell, Mobile Service Tower, Fixed Umbilical Tower, and Tail Service Masts!!
Slick Six turned 60 in March and has seen a whole lot over its career. It was built for the Manned Orbiting Laboratory — a crewed spy station the Pentagon canceled in 1969. It was then rebuilt, at a cost of billions, for West Coast Space Shuttle launches. Challenger scuttled that plan in 1986, and not one Shuttle ever flew from the pad.
Jilted at the altar twice, Slick Six finally found honest work lofting the NRO’s largest spy satellites into polar orbit. A late bloomer! Now, under a sole lease to SpaceX, the old pad is on its third act, and a happy one, with SpaceX planning to double Vandenberg’s Falcon cadence to ~100 launches a year.
From a couple false starts to a workhorse of reusable flight: ad astra per aspera.
|