Happy Friday folks. On this day in 1975, an Apollo and a Soyuz docked nose-to-nose and the two crews shook hands 150 miles up. The rendezvous took three years of joint engineering to pull off; getting this issue wrapped up for you all took a busy week and one candle, burned at both ends, right up to COB Friday.
And while we’re here, Monday marks 57 years since Apollo 11 touched down at Tranquility Base. Here's to the hundreds of you reading this who are working on our return trip as we speak — to ensure we're the first to land boots (back) on the Moon, and this time around, to stay.
IN THIS WEEK’S EDITION:
🔩 Should you have to build the whole stack?
🫡 Dan salutes Chris Scolese
📊 You have spoken re: SPACs
🛥 Maritime first, a new FPV prime, & FDEs
🇨🇭 The reality machine shuts down
Forwarded this? Subscribe to Per Aspera
MORE MAGNETS, MOTORS, CASTINGS, & CONNECTORS
Or: Why Hard-Tech Founders Shouldn’t Have to Chase the Whole Stack
There are a couple of things that we keep saying:
Whether it’s a robot, chip, substation, or spacecraft, we should be able to map a critical system’s supply chain several tiers deep, down to the irreplaceable parts — and never discover that the only way to build any of these parts at scale runs through a jurisdiction that doesn’t wish you well (or, through a single, supply-constrained source)
Capital, mindshare, policy, and talent tend to chase the tip of the spear — the most exciting things — like: frontier models, GW-class datacenters, $10B+ megaprojects, finished systems, and the like.
Put the two together and we as a community worship at the altar of maximal vertical integration, thanks to the strength of the outlier success stories: SpaceX! Tesla!
It is fine and good to want the whole stack — the integrator captures the narrative, hearts and minds, margins, and multiples. But should you have to be full-stack in order to succeed?
Today, in an informal essay, we get into a trend that’s got us very excited: the merchant base is reconstituting. Inside: the hero’s journey of a vertically integrated maximalist, the trials and tribulations of the long tail, and a new wave of startups who have recently emerged (Westmag, Heron Power, Stone Power, Arcturus, Molten Industries, etc) to build actuators, motors, turbines, piece-parts, and upstream precursors that go into everyone else’s beautiful machines.
Hello, Dan here popping in to celebrate Chris Scolese, who officially retired from the NRO last week 🥳. Chris goes down in history as the 19th Director of the NRO (the 2nd longest serving of the agency) — and my friend of the past 30+ years.
Chris Scolese early in his career (left); Scolese in action (right).
Chris — who began his career in the U.S. Navy in 1978 as a Naval Officer — was personally selected by the legendary Admiral Hyman Rickover, father of the Nuclear Navy, to serve at Naval Reactors (NAVSEA). Under Admiral Rickover (who many of you in the PA community know had uncompromising standards), Chris built instrumentation, sensor systems, and multi-processor architectures for nuclear reactor control.
I met Chris when I arrived at NASA in 1992. At the time, the Earth Observation System was projected to cost what I remember to be ~$18B. It was seen as unaffordable and headed for cancellation.
Chris rescoped the entire program and delivered the full working constellation for $8B. I’ve said this publicly at the Space Symposium before: Chris saved $10B (returning funds back to the American taxpayers). A compatriot in the Faster, Better, Cheaper army.
All 3 EOS flagships — Terra (1999), Aqua (2002), and Aura (2004) — are still in orbit today, operating well beyond their design lifetimes, along with Landsat and many other missions. They deliver critical data to farmers, firefighters, and first responders here and around the world.
Terra (1999), Aqua (2002), and Aura (2004)
Chris is quiet, focused, and brilliant. He's always been in the background. He went on to serve as Chief Engineer, Associate Administrator, Acting Administrator, Director of Goddard, and 7 years leading the NRO. Nearly five decades of service.
Chris — thank you for your precise, calm, and impactful service to America.
— Your colleague friend, Dan Goldin
Last week, we talked about how SPACs are making a comeback for companies in defense, energy, space, and the like. And, given this second act, we wanted to know what you all make of it. So we asked: 1) Are SPACs redeemable, 2) or terribly radioactive or 3) simply not sure?
You answered, and… the results were more balanced than we thought they would be (surely we’re all feeling blind-pool-warrant-sweetener PTSD)!
That said, a number of PA readers — people who want hard tech to get capitalized, who cheer when a robotics or fusion company finds a path to the public markets — still won't touch the instrument. Some notes from you all:
“Companies pursuing a SPAC are out of options — they don't have the track record to go public via IPO and they don't have private investor interest. Betting on one means that private investors got it wrong. Private investors are far from perfect, but the ratio of diamonds to coal is about the same as good companies pursuing a SPAC.”
“I know several fusion companies running to SPAC and they are all the ones that the industry knows have run out of road on their concept. Hugely damaging.”
001A / FIRSTS COME IN PAIRS… Three Saronic Corsairs — 24-foot autonomous boats built in Austin — struck the submarine maintenance yard at Bandar Abbas on July 12, marking the first combat strike ever carried out by American sea drones.
Three Corsair USVs strike Bandar Abbas Naval Base, Iran, July 12, 2026.
Source: U.S. Central Command (@CENTCOM)
This is the very same platform that just five weeks earlier pulled two downed Apache aviators from the water off Oman after two hours adrift — the first at-sea personnel recovery by an unmanned vessel.
Splash Industries' Typhoon USV, via Ivan Avanesov
001B // AND ONE MORE FOR THE ROAD… The Pacific logged its own first the same week as the Gulf (^), when Splash Industries’ ten-foot Typhoon autonomously delivered 3D-printer parts to the USS Essex and recovered itself into the moving ship's well deck. Splash CEO (and friend of PA) Ivan Avanesov calls it a world first — and no small feat: Typhoon had to match speed with 41,000 tons of a moving warship, and then drive itself into the open well deck. Watch the video here.
002 / GUNDO STARTUP’S NOW A PRIME… El Segundo darling Neros, founded in ‘23 by former world‑champion drone racer Soren Monroe-Anderson, has landed a $500M Army contract to supply its Archer attritable FPV attack drones.
Soren Monroe-Anderson (left), Olaf Hichwa (right), Co-Founders of Neros.
Source: NY Times
The company has rapidly grown from a small El Segundo shop into a factory operation with dozens of engineers and operators building U.S.-made FPVs at scale, positioning a relatively lean team to deliver on a ceiling usually reserved for legacy primes. Oh how the times have changed! 👇
003 / FDE… On July 15, Anthropic x Blackstone x Hellman & Friedman launched Ode with Anthropic — a $1.5B forward‑deployed engineering + services business that embeds hit teams of engineers inside companies to build custom systems on top of Claude.
Ode with Anthropic's founding partners and investor consortium.
Ode’s key targets are mid‑market and PE portcos — community banks, regional health systems, and mid‑size manufacturers — that lack deep engineering benches. They tend to be domestically anchored, with local customers, supply chains, and workforces, so they generally couple to their local/regional economies. In theory, this means the upside should also accrue to the local community base in addition to the effort’s sponsors. (What can we say: we’re optimists and techno-optimists.)
Also note: Goldman’s horizontal play as investors (alongside Spanish bank BBVA) in the OpenAI Deployment Company (DeployCo), a $4B+ venture that similarly embeds FDEs to bring GPT into prod at large enterprises.
Speaking of… ICYMI last year, we teamed up with Mark Scianna, one of Palantir’s first-ever FDEs, to write: How to Build Your 1st FDE Team.
THE REALITY MACHINE IS DOWN FOR MAINTENANCE. At 6 a.m. on June 27, CERN's Large Hadron Collider — the 27-kilometer ring buried under the French-Swiss border, the most powerful particle accelerator ever built — dumped its final beams and went dark for four years. Long Shutdown 3 has been on the calendar for years: a ~$1.5B rebuild that will swap out 1.2 km of the ring for the High-Luminosity upgrade — including the first Nb₃Sn superconducting magnets ever operated in an accelerator.
The internet has, naturally, taken this news calmly. TikTokers diagnosed the Mandela Effect reversing and timelines quietly merging due to CERN shutting down, while X treated us to its fair share of bangers, such as users logging their first eight-hour sleep of the year, vanished back pain, afternoons that always “feel like 16pm,” and long-lost mutuals reappearing on the feed. The internet theorists are generally split: one camp believes that catastrophe follows a switch-off (citing Long Shutdown 2, 2018-2022, which contained the entire pandemic), while the other credits the beam dump for the best week of their lives. Then there’s a third camp which is whatever this is: “They cut CERN during a full moon, right before Mercury retrograde — buckle up.”
We regret to inform the internet theorists that A) the machine also switches off every winter and reality has survived each Christmas thus far, and B) the machine’s documented brushes with the uncanny are decidedly mundane:
In 2016 a stone marten chewed through a 66 kV connection and took the collider offline for a week
Then, that November, a second marten repeated the trick and is now taxidermied in a Rotterdam museum
And, in 2009, a passing bird dropped a baguette onto the capacitors and tripped the cryogenics.
For the non-timeline-obsessed — the physics-pilled among the PA faithful — we’ll offer you some parting fun facts instead of more memes:
Switching this thing off is a massive logistical program: the LHC carries 36,000 tons of cold mass at 1.9 kelvin — colder than deep space — chilled by ~120 tons of liquid helium (Asian chipmakers right now: must be nice… )
Warming it back up takes 4-5 months, on a schedule CERN’s engineers say is set by “cryogens inventory logistics.”
Beams will be back in 2028, with full-power physics, at roughly 10x the old collision rate, coming in 2030.
The martens, one can only assume, are regrouping.
This is a gif of the SpaceX recovery team — who rode by boat roughly 2-2.5 days, departing from the port of Exmouth (off the coast of Western Australia), to reach their station 500 nautical miles out in the Indian Ocean. 🇺🇸 This is what it looks like working at the most valuable company in the history of the world — and it looks like doing hard things.
And with that, go forth, do hard things, and prosper - we’ll see y’all back in the inbox next week.
|