
Happy Monday, y’all. Today, we’re looking back:
It’s been a few months since we started publishing Per Aspera. 100 days in, 14 editions, and 100,000+ words later, what were the biggest stories of the summer? Keep reading for the team’s rundown…
Personally, I just celebrated two years at my job (head of BD at Array Labs). Below, you’ll find a reflection on 24 things I’ve learned in 24 months, covering things that nobody tells you when you’re starting out in ✨deep tech✨. Hopefully it’s helpful for some of you on your own journeys — and instructive for those of you trying to find your way into this world.
Looking forward, one quick housekeeping note: we won’t be sending next Monday in observation of Labor Day. S/O The American Worker!!!
P.S. Were you forwarded this email? Subscribe here.



Hot American Renaissance Summer
We launched Per Aspera ~100 days ago with a bet: that a renaissance was stirring in the hardest pursuits of science, industry, and technology. In that short time, we’ve witnessed enough to prove the thesis….while discovering just how far we still have to go. This is the state of the renaissance as summer 2025 draws to a close.
Star of the Season, or Summer Slump?
Artificial intelligence was the main character this summer. AI startups captured a majority (53%) of venture funding in H1 2025 — in the U.S., that share was even higher, at 64%. For us, the real story is measured not in headlines about the AI application/wrapper layer, but in the billions upon billions poured into the physical bedrock of computation. Between now and 2029, Morgan Stanley expects global datacenter spending to approach $3T – roughly the size of France’s economy. Nearly half of that capex is expected to flow from the balance sheets of hyperscalers/Big Tech (Amazon, Microsoft, Google, Meta). The remaining $1.5T will need to be filled by external investors and developers, setting off the largest construction and financing boom in the history of digital infrastructure.
AI has had some wind knocked out of its sails in recent weeks, with mounting mentions of the B-word, as frontier lab leaders reset expectations and corporate pilots overwhelmingly fall flat.
An Energetic Stagehand
If AI was the star of summer, energy was the stagehand struggling under its weight. Datacenters are now the largest incremental source of electricity demand in America. Northern Virginia developers alone have filed power requests equivalent to the entire load of NYC. Lead times for high-voltage transformers have stretched to 4–5 years. Utilities invested $178B in 2024 capex, the 13th annual record in a row, yet after a decade of flat demand, it still isn’t enough. Uncle Sam, hyperscalers, and opportunistic state governments are looking toward the cavalry on the horizon – nuclear – but it doesn’t arrive fast enough to unclog today’s grid queues.

Inputs, Chips, & Minerals
Summer reminded us that trillions poured into AI and energy mean little without control over inputs. Without wafers, magnets, uranium, and processing capacity, the renaissance is fundamentally rate-limited.
On chips, the U.S. just confirmed that it’s taken a 10% stake in Intel, underscoring both the fragility and centrality of America’s last integrated chipmaker. $200–300B of fab investment has been announced across Ohio, Arizona, Texas, and New York, but the bleeding edge remains abroad.
Minerals tell a parallel story. China’s springtime rare earth squeeze forced Ford to idle its Chicago line, spiked prices, and rattled defense primes. The U.S. remains 100% import-dependent on 12 critical minerals and 50+% reliant on 31 others. This summer brought progress, but there’s much more work to be done.
Human Capital
The renaissance happens on factory floors, job sites, and control rooms, and right now, those places are running out of hands. Across semiconductors, batteries, aircraft, robots, minerals, ships, and grids, the bottleneck is often talent. Projects hit the same wall again and again: not enough welders, machinists, electricians, or operators.
America’s generational bets are not just competing with foreign rivals — they’re competing with each other for the same scarce pool of skilled hands.
The upshot
Summer 2025 showed both strength and fragility: capital is flowing and factories are breaking ground, but shortages of power, inputs, and people threaten to choke the momentum. Week by week, it seems that physics, policy, and the powers that be are conspiring to push this forward. What will determine whether this moment endures, though, is breadth and depth:
Breadth: training and mobilizing enough welders, machinists, electricians, and engineers (or, deploying a workable technological substitute)
Depth: preserving, rediscovering, and scaling the tribal knowledge and process mastery that turn theory into reality.
Without breadth, the renaissance starves for hands. Without depth, we revert to dependency. Above all, the hard path means cultivating stronger men and women willing to do the work. As JFK said: do not pray for easy lives, pray to be stronger. That is what this renaissance demands.
We want your perspective: What will the biggest story be for the rest of 2025?
Vote Below

Closeshoring™ is driving America’s technology-enabled industrial reinvention, converging localization, automation, and digitalization to secure our position as the global industrial leader. Led by Alan Simon (CEO), Jeb Bush (Chairman), Brad Skinner (Sr. Partner), Jeb Bush Jr. (Managing Partner), and Santi Barrosso (Partner), Closeshoring™ helps growth-stage and large teams strengthen supply chains, revitalize communities, and accelerate national capabilities in defense and space — laying the foundation for America’s strategic future.


HOW I LEARNED TO LOVE THE PAIN, CHEWING GLASS, AND HARD PROBLEMS
Two years ago I switched teams. After half a decade covering the frontier as a writer, I joined a space startup in Silicon Valley as part of the founding team. The inside story of building a hardware company is far messier and more complex than one could ever appreciate from the outside looking in. You come to appreciate firsthand the shape and texture of the company’s journey between the peaks and valleys. You live through the messiness and learn because it’s your life.
As I cross the two-year mark, I’ve pulled together a list of 24 things I’ve learned in 24 months at Array Labs. Most reflections come as postmortems or victory laps. This one is contemporaneous — a look at how we’ve managed to keep calm and carry on. Hopefully this is useful for someone else on their journey. At the very least, it should be a good reminder you’re not the only one chewing glass.
(P.S. if this sounds energizing, then apply for a job at Array!!!)


Last week, we published Starlink, Rural America, and A New American Frontier, which explored how Starlink could shift the value of land in America by erasing “the distance tax” that has defined rural life for centuries.
Afterward, we asked you: do you own a Starlink? Here’s how you answered.

ONE NOTABLE RESPONSE
“I bought Starlinks for my 75 years old mom who lives in a remote area in Costa Rica. The best investment for human connection!”



TSMC’s Fragile Armor. Taiwan’s silicon shield is showing some hairline cracks, per MIT Tech Review. TSMC has long acted as the island’s “silicon shield,” deterring Chinese aggression due to its existential role in the global chip supply (and, consequently, the entire technology value chain). Facing U.S. pressure, TSMC is scaling production far and wide. This is raising hackles among nationalists in Taiwan, who fret the island is losing leverage. This has fueled intense debate domestically, prompted the country to pass a law that prohibits TSMC from exporting its leading-edge technology, and fueling Chinese wolf-warriors/disinformation campaigns, which say the company’s overseas expansion will “hollow out” Taiwan and erode its strategic value. ICYMI, a few weeks ago, we dug into the global high-wire act that is chip supply.

Back to End of Thirst Traps. A Sci Am feature takes a closer look at the viability of subsea desalination. Researchers have deployed submerged pods off the coast of Santa Monica, CA and Mongstad, Norway, promising cleaner, cheaper fresh water by letting ocean pressure do the heavy lifting of reverse osmosis — and conveniently sidestepping the zoning/regulatory battles you get onshore. For Southwestern states and water authorities, as we laid out in End of Thirst Traps, clever pilots aren’t enough to close the massive structural deficit they’re facing. What’s needed now is a real willingness — politically and financially — to scale what really works. Our modest proposal is building a true solar/storage-powered desal megaproject.

X-37B. The Pentagon has launched the X-37B for its eighth mission, this time testing high-bandwidth laser satellite communications and quantum inertial sensors in orbit. The ability to return the Boeing-built orbital spaceplane for reuse accelerates learning cycles, offers a rapid testbed for new technologies, and helps harden USG space architectures against countermeasures. Also, who doesn’t love spaceplanes?!?
What’d we miss? Have something others participating in the Renaissance should know? Hit reply and drop us a line at [email protected].


