Newsletter 32

Happy New Year’s Eve, y’all. We know (or we hope) that most of you are AFK — away from keyboard — so today’s issue is quite light. Thanks again for joining the Per Aspera community this year and for being part of something bigger: what we like to call ‘The Renaissance in Hard Pursuits.’ We see you all — including our friends around the world — showing up to push the work forward.

With 2026 around the corner, the 250th year of the American experiment, there’s only one thing we’re certain of: we will keep going. Everything we care about is shaped by whether we stay in motion.

Wishing you and yours a year of prosperity, exploration, security, and dynamism. See ‘ya in 2026!

— Ryan, Dan, Joy, Jeff

Forwarded this email? Subscribe to Per Aspera here.

Rather than write another “year in review” postmortem or “year ahead” prediction post, we’re sending along five of the most popular Antimemos (long-form essays) that we published this year. Whether you’ve joined us recently or you’re looking for a longer read this holiday season, we’ve got you covered. Dive into one or more of the essays below, and don’t be afraid to drop us a line letting us know what you thought.

At a time when everyone and their mothers are talking about space-based datacenters – and SpaceX is planning an IPO primarily to capitalize its forthcoming orbital compute layer – let’s rewind the tapes back to May 19. It was only Per Aspera’s second week publishing, and for our first Antimemo, we dropped a 16,000-word beast titled Realities of Space-Based Compute. Yes, it covered space-based datacenters. And it was dense, packed with intricate modeling and tons of analysis on unit economics, physics, and more, but that didn’t scare you away. This was our most popular post of 2025 – and to be honest, we weren’t expecting it to take off like it did – but guess it pays to be prescient (and lucky, timing-wise!) A spoiler alert: we present a more nuanced, measured assessment on the concept than you see online: neither full, <5-year bull nor ‘it will never work.’

Probably the second most talked-about tech trend of this year (ex AI) seemed to be “forward deployed” this or that, so it’s fitting that our field manual on building an FDE (forward deployed engineering) team was our second most popular essay. AI companies are hiring for it; Satya and Benioff have praised it; and half of LinkedIn slapped the title on roles that were really just sales engineering in a trench coat. So we called in someone with actual receipts: Mark Scianna, a former Palantir FDE – one of the very first – who spent months dodging rocket fire in Kandahar while debugging databases, and who now runs Forward Deployed VC. Together, we wrote the definitive playbook on what FDE actually is (and isn't), the unforgiving economics most companies can't stomach, and a four-part acid test to tell the real thing from "Sparkling Sales Engineering." If you're considering the model – or just sick of the buzzword – this one's for you.

Top Gun: Maverick opened with Tom Cruise punching past Mach 10 in an experimental jet – an iconic, memorable scene that Hollywood chose because hypersonic flight still feels like the edge of American ambition. The frustrating truth, as we lay out in Why Aren’t We Flying Faster, is that we achieved that edge 60 years ago. The X-15 hit Mach 6.7 in 1967. The X-43A set a scramjet record in 2004. The X-51 sustained hypersonic cruise in 2013. Each time, we proved the physics worked – then shelved the program. Meanwhile, China and Russia are steadily advancing their hypersonic programs (and fielding, and parading, real-world systems). Dan, who greenlit the X-43A as NASA boss, brought in Mark Lewis, former Pentagon chief scientist for research, to diagnose what Ryan has called "strategic amnesia" – and chart a path forward. We’ll give you a hint of our verdict: we don't have a physics problem, we have a follow-through problem.

Here's a fact that should bother you: in frontier AI clusters, moving data can now burn nearly as much power as computing it. The will of our epoch is a hunger for intelligence – at scale, on demand, everywhere. We're building cathedrals of compute to feed it, but the substrate underneath is cracking. Resistance, heat, and the physics of copper are throttling these cathedrals. In Architecture Movement: Photonics, we lay out why we believe the escape route is light. Photons are massless, chargeless, and immune to the resistance taxes strangling copper. Electrons gave us the industrial internet; photons are poised to power the cognitive one. Mark our words: 2026 will be a massive year for photonics.

For two centuries, American land obeyed a law as reliable as physics: the closer you were to a city, the more your land was worth. Distance was a tax. Every wave of infrastructure – rail, highways, broadband – reinforced this rule, reaching dense urban cores first. Now, this law is under quiet assault from 340 miles overhead, led by Starlink. LEO broadband inverts the traditional economics of connectivity, since shared-capacity satellite networks perform best where fewer people live. Working with our friends at Mach33, we built a model to understand what happens to land values when connectivity reaches the places that fiber and capital forgot. Does it redraw the price map of American land? We think it might. One thing’s for sure: the frontier feels alive again…

We want to close out 2025 by thanking our sponsors who saw something in our collective mission and decided to bet on us this year. Your support means everything to us.

Thanks to you, we've been able to make Per Aspera free now and for the foreseeable future for anyone who wants to participate in this great American Renaissance. It's been a busy year, and a most fulfilling start to what we're building together.