Previously on FBC… Last week, we (re)introduced you to “Faster Better Cheaper” and laid down the seven laws. The philosophy has practitioners in many high, hard places.
But don't take it from us. Listen to Tom Mueller (!), who was SpaceX Employee #1 and is now founder/CEO of Impulse Space:
Today, in Part 2, we’re asking the hard questions:
How do we turn FBC loose on an America hungry to make things again? How do we scale? What does rebuilding supply chains mean? How do we train, adopt, cluster, and move fast enough for it to actually matter? And, most importantly, how do we measure it?
Forward this to someone who wants to go Faster, Better, Cheaper. And if you’re that someone, subscribe to Per Aspera here.
Part 2 of our Faster Better Cheaper series starts with a confession.
We’re addicted to Shiny‑New‑Thing Syndrome: lionizing the billion‑dollar round, the giga‑mega-super-fab’s ribbon‑cutting, and press releases that hit long before ground is cleared or power is procured. (Present company included. It’s exhilarating to live at the pointiest edge of the spear!)
The problem is that the tip of the spear, and the edge of real-world progress, rests on a base. And this base extends across machining, fabrication, materials science, and a long tail of suppliers, specialty job shops, integrators, and sub‑assemblers who, somewhere along the way, help turn raw materials into finished systems.
Our fixation at the tip has produced a giant sucking sound at the base.
Ross Perot, bless him, was onto something…
That’s the sound — and we’ve all heard it — of sovereign supply chains, processes, parts, and piece-parts being deprived of oxygen and pushed offshore to willing producers.
The premises of FBC 2.0:
We’re in what we call a Capacity Trap: We can’t build at volume because we lack supply chains; we don’t fund supply chains because we don’t have volume; the supply chain never arrives because we never funded capacity.
This leads to the Vertical Integration Reflex: CEOs and founders who see this most clearly respond in a rational way: go fully vertical, or die trying. At the firm level, this is a strategic choice that can often make sense. But at a systems level, it should not be required for a hardware company to survive.
We have a K-Shaped Industrial Comeback: The top arm (fabs, primes, neo-primes, megaprojects) is up and to the right, with the bottom arm inflecting downward (the supplier ecosystem and long tail.) This doesn’t have to be either-or. It can (and should) be both-and!
A proposal for breaking free
Per Aspera has a modest proposal — to build an American Industrial Complex using FBC principles. Our proposal rests on five pillars: 1) full-stack factories, 2) FBC cluster, corridors, and cities, 3) and you’ll get the rest in the full article. 😉
We close with a question we don't fully have the answer to yet: how would we know if any of this is working? We don't have an index of simple metrics and biomarkers of America’s industrial health, but as a community, it's on us to start building one. We’ll work toward this in the coming months, and commit to convening the right folks under one roof in DC this fall to develop the “Scoreboard” we so desperately need.
Read today’s story and don’t be a stranger! Drop us a line if you’ve got thoughts, notes, or ideas for how to help.
If you missed Part 1…. here you go!
Last week, we asked you all to write in with a question for the community. Here’s a good one from David Merrill, Founder and Executive Chairman of Elroy Air.
What are the scarce elements of today's and tomorrow's industrial supply chains that need more/better products/technologies, to fill gaps and accelerate progress?
A most topical question indeed!
001 / OPEN TO INFERENCE… On a single Tuesday (Feb 24), we got three separate bets against Nvidia's AI chip monopoly:
Meta signed a deal with AMD (≤ $100B)
SambaNova announced their $350M+ Series E, unveiling their SN50 inference chip and Intel as both investor + co-design partner
MatX, founded by the designers of Google's TPUs, raised a $500M Series B
These are all bets on inference (also not first of their kind — OpenAI/Cerebras, NVIDIA/Groq). With agentic AI and machines talking to machines at massive scale, inference is eating training. NVIDIA dominates training but inference is wide open and where the AI architecture war is moving. At the same time, China sees its own opening: DeepSeek has supposedly cut both Nvidia and AMD out of the pre-release optimization window for its upcoming V4 model, giving Huawei a weeks-long head start to optimize its Ascend chips for inference.
On one hand, the inference war is great if we produce a robust, competitive, multi-vendor AI ecosystem. (🇺🇸 companies doing what we do best — competing!). On the other hand, some might argue that now is not the time to fragment into a dozen incompatible bets while China consolidates around one stack. Thoughts?
002 / BURIED IN BRIEFS… Canada has advanced five major mining & processing projects this quarter* that feed into allied nuclear fuel and battery supply chains. Comparatively speaking, this is making us feel a little…flat-footed? The U.S. is the second slowest in the world in developing a new mine, with an average discovery-to-product timeline of ~29 years. Canada and the U.S. have comparable regulations, so what gives? It's the long tail of litigation: American mining projects face more legal challenges than Canadian and Australian ones combined.
*About those projects: Denison got the green light to build its Phoenix uranium mine at Wheeler River in Saskatchewan, Foran’s McIlvenna Bay copper‑zinc site is 88% built with wet commissioning underway, Kinross’s $5B Great Bear gold was put on Ontario’s fast track, NMG bought a Bécancour brownfield to supply Panasonic’s 13,000‑tpa graphite anode needs, and Rio Tinto took control of Nemaska Lithium with $300M+ earmarked for Québec’s LiOH plant and Whabouchi mine this year.
And while we’re here: For those in the industry who will be at PDAC next week, so will Ryan! Reach out to him to link up.
003 / ADD TO CART… The Pentagon is kicking off “The Gauntlet" at Fort Benning, Georgia, this month, a first-phase field trial where soldiers fly and rate cheap one-way attack drones from 25 vendors — part of a four-stage, ~$1.1B effort to field tens of thousands of units this year, ramp to hundreds in 2027, and procure 300,000+ by 2028. Competitors included a mix of U.S. drone shops, big and small, and at least two Ukrainian manufacturers tempered by years of war with Russia. This week, the Pentagon also switched on the Counter-UAS Marketplace, an Amazon-style storefront where commanders can browse and buy from a tested inventory of 1,600+ counter-drone tools, from sensors to jammers (echoing Ukraine’s own Brave1 Market).
In Issue #034 we walked through the supply-chain holes exposed by the DJI ban — a market where Chinese firms dominate everything from airframes to motors. Even on the Pentagon's own Blue UAS "trusted" list, recent reports say that a majority of platforms still rely on Chinese-sourced motors and other so-called “duumb” components. The Gauntlet is part of a much bigger push, and a huge (300,000+) order book aimed at conjuring the American supply chain into existence.
Before You Sign Off...
On Saturday, Feb. 28, six planets will line up in the evening sky — Mercury, Venus, Neptune, Saturn, Uranus, and Jupiter — appearing in the first hour after sunset. Four should be visible with the naked eye if the weather cooperates, while Uranus and Neptune will want binoculars or a small scope. If you've had a long week or recent events have left you feeling a little cooked, you could do worse than stepping outside for 15 minutes and looking up. Touch some grass, find a clear patch of sky, and take it all in. We're all still riding on the same rock together, under the same stars, trying to get the next orbit right.
Per Aspera Ad Astra!
|