Catapulting satellites into space. Making all the data center heat useful. Why memory technology needs disrupting. All that and more in this week’s Prototype. To get it in your inbox, sign up here.
As I mentioned earlier this month, the space economy is facing a bottleneck when it comes to rocket launches: Demand is currently outstripping supply. It’s also possible that there will be fewer launches in 2026 than there were in 2025. To be sure, there are plenty of rocket companies racing to close the gap, but “It’s not rocket science” is a cliché for a reason: Actual rocket science is often filled with delays and setbacks.
It’s why Winnie Lai wants to eliminate rockets from at least part of the equation. She’s the founder of Auriga Space, which is developing a linear electromagnetic accelerator to catapult rockets to high altitudes, where their engines kick in to bring them to orbit. It essentially replaces the typical first stage of a rocket.
This isn’t a new idea–physicist Gerard O’Neill proposed it in the 1970s, and subsequent generations of scientists have worked on the problem. Today, the U.S. Navy’s aircraft carrier USS Gerald R. Ford uses an electromagnetic system to enable fighter jets to take off.
What’s kept such systems from space launch, Lai told me, is that the underlying technology wasn’t ready yet–but advances in electronics and semiconductors have changed the equation. Today, her company has already built prototype systems that can fire small metal slugs up to Mach 2.4 (that’s over 1,800 miles per hour).
So far, Auriga has raised more than $12 million from investors and Department of Defense grants. Earlier this month, the company announced that it’s going to market with its Prometheus accelerator–not for bringing payloads into space, but for testing materials at hypersonic speeds to start building the revenue the company will need for further development. Axiom Materials has signed on as a pilot customer.
Auriga is also developing its technology for defense purposes, such as anti-drone weapons. And in the long run, Auriga aims to build a multi-kilometer accelerator that can deliver small satellites into orbit.
“Our ultimate goal is to make space launch more efficient,” she said. “By increasing efficiency, we believe we can bring down the cost and enable more frequent launches.”
Quick housekeeping note: I will be on vacation next week, but will be back with a new edition on July 24th.
Discovery of the Week: This Tiny Tube Could Turn A Data Center’s Heat Into Electricity
One of the challenges of the AI era is managing the heat generated by the servers that run it, which must be cooled with air or water to avoid overheating. That uses around 20 to 40% of the power going into them. One tantalizing possibility in managing this heat are thermoelectric products that convert heat directly into electricity. Such devices are already used for some applications, but efforts to bring them to industrial scale have tended to fall short due to economics or mechanical inefficiency.
New research from Korean university Postech could change that. Scientists there have developed a method that uses silicon nanotubes for thermoelectric devices, which could make it possible to scale up the technology and enable data centers to recycle their waste heat into electricity.
The key, according to their research, is silicon nanotubes: As it turns out, these tiny, hollow tubes work much better than solid wires for thermoelectric devices. The hollow center traps heat-carrying vibrations, while still allowing electricity to flow. Nanotubes reduced heat flow by up to 70% compared to silicon wires, solving one of thermoelectrics’ biggest problems: blocking heat while still allowing electricity to flow. That’s been especially difficult with silicon, which has historically let heat pass through too easily, limiting the efficiency of thermoelectric devices.
Next up, it needs to be shown the technology can be scaled up to observe its performance in a larger commercial generator. If successful, silicon nanotubes could be used in thermoelectric devices attached to AI servers, EV batteries or even industrial machinery to turn their waste heat into useful energy.
– by Sydney Goitia-Doran
The Hot Take: Memory Tech Needs More Disruption
Each week, I ask investors for their take on tech trends within their industries. Today the answers come from Michael Stewart, managing partner at M12, Microsoft’s venture fund. His investments are focused on generative AI, hardware and gaming.
What tech is being overhyped right now?
Humanoid robotics. Why? Because as much as the use case makes total sense to me, robots are generally not universal. They’re almost always specialized, right? So it could happen with a huge behavioral mindset change in society to accept humanoids everywhere as some people think. But I believe the initial experiment with humanoids will be phase one: humanoids. Phase two? A bunch of super specialized non-humanoid looking things, but we’ll see.
What should more people be talking about today?
Innovation in memory technologies in semiconductors. Everybody can comment on the memory cycle, the prices, but disruptive innovation in memory has been absent for as long as I’ve been investing. And I’m one of the only people I know that actually has invested in memory startups. So that is to me the biggest area in this conversation that we could be deploying more in.
What are we all going to be talking about in five years?
I think we’re still maybe underestimating how attached and important personal AI will be in our lives. It will integrate at a level that’s very difficult to compare to something today. In fact, we’ll be maybe wistfully, not in a bad way, but wistfully looking back and thinking like, “Remember those days when the AI had to be prompted and told how to help out?” And it will get to a point where it’s quite essential for your everyday life in a way where you’ll be safeguarding it and being very careful about it.
What’s Entertaining Me This Week
I recently rewatched one of my all-time favorite movies, A Man For All Seasons. Based on Robert Bolt’s play of the same name, it stars Paul Scofield as Thomas More, who was persecuted by King Henry VIII for refusing to take an oath acknowledging him as the head of the Church in England. It’s a cliché to say “they don’t make movies like this anymore”–they do, but there are simply too few of them. This is a whip-smart film that explores issues of corruption and moral integrity, showing how much easier it is to go along than to stand your ground. It never dumbs itself down to the audience and features some of the most crackling dialogue of all time: “Why Richard, it profits a man nothing to give his soul for the whole world… but for Wales?”
