Closing Bell - Manifest Space: Semiconductors in Space: Inside Besxar’s Plan to Redefine Chip Manufacturing 10/30/25

Episode Date: October 30, 2025

Morgan Brennan sits down with Ashley Pilipiszyn, CEO and founder of Besxar, a startup that just emerged from stealth with a bold mission — manufacturing semiconductors in space. Pilipiszyn explains how Besxar’s autonomous “fab ships” leverage the ultra-high vacuum of space to create ultra-pure wafers and substrates — the foundation for every advanced chip powering AI, data centers, and defense systems. They discuss the company’s upcoming launch campaign with SpaceX and why space-based manufacturing could dramatically cut costs while securing domestic chip supply chains. Pilipiszyn also shares how her time at OpenAI and SLAC National Lab inspired this fusion of AI, materials science, and orbital production — and why she believes in-space manufacturing is American manufacturing. Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

Semiconductor manufacturing is complicated. It costs tens of billions of dollars to build a fabrication plant and can take years to build. But what happens if you strip all of the earthly constraints out of the production process? One of the core issues and really the genesis for why I even started the company was realizing that as chips continue to get smaller and smaller, the cost of the price point and then the cost to manufacture them only increases. And interesting enough, there's actually a Moore's second law. Everyone talks about Moore's law. There's actually a law that states this called Moore's Second Law, also known as Rock's Law,

which states that every four years, the cost to construct a new leading edge fab doubles. And we're seeing this happen in real time. TSM just announced the other day that their new 1.4 nanometer fab is going to cost nearly $50 billion, up from the $20 billion that was just constructed in Arizona. So we're seeing this play out, and that doesn't scale. So for our approach, we're trying to strip away what actually goes into manufacturing a wafer. And when you really boil it down, you're talking about raw precursor materials, you're talking about the wafer foundations themselves,

and realizing that what goes into a lot of fabs in terms of capital, capital expenditure is HVAC systems, multi-billion dollar vacuum pumps, like I said, humans and bunny suits, a lot of solvents, a lot of water. And for our approach, we're actually removing all of that. Bexar Space Industries is emerging from stealth, founded two years ago by Ashley Pil Lipschen, a former early open AI employee. Bexar is building fab ships to manufacture the ultra-high purity wafers and substrates used in GPU. and do so from space. The company is a member of

NVIDIA's Inception program, has a contract with the U.S. Navy, and is backed by investors including B-Ventures, Union Labs, and SpaceX, which will launch Bexer's missions. We've got 12 launches to go ahead, so we're going to be launching 24 fab ships, two per launch,

and what the important part of this is being able to rapidly iterate. So really, this idea of load and go, being able to put in the wafers, test them, see how they perform. You can think of this similar to the ultimate egg drop challenge. So we want to ensure not only can we get wafers to space, do our manufacturing, but also on re-entry and return that we're able to successfully bring back wafers

without any type of cracking or damage like that. And she says the fab ships will represent the first ever reusable payload program for SpaceX. On this episode, Semiconductors from space with Bexar's Ashley Pellipion. I'm Morgan Brennan, and this is Manifest Space. Joining me now, Ashley Pallipchian, the CEO and founder of Bexar, which just emerged out of stealth. It's great to have you on. Welcome.

Thanks for having me. So let's talk a little bit about that, because you did just bring this startup that you've been working on out of stealth. What does Bexar do? So we leverage the ultra-high vacuum of the vacuum of space to manufacture ultra-high purity wafers and substrates. These are the foundational building blocks for every semiconductor, from cars to houses to every GPU. And one of the issues is that there are only a handful of companies that can manufacture them, and they continue to be more and more expensive to be able to manufacture.

And so we're really focused on bringing next generation materials forward for the future of semiconductors. How quickly can you start sending these, they're called fab ships. How quickly can you start sending these fab ships to space and start making semiconductor materials? So we're actually starting our first flight campaign with our Clipper class. And our Clipper class, much like the name, are for short. sorties and very agile. We're going to fill them with a variety of different terrestrial semiconductor wakers to be able to do some initial testing. We then want to use that data to understand how they perform during launch, time and space, and reentry in order for us to scale

up our production capabilities. So our production capabilities, we're looking to scale in the next two years, and that's where things are going to get really interesting. And we'll be able to provide the materials that we're looking to work with our partners across AI data centers and defense. Do you already have partners or customers lined up? We do. We actually have a contract with the Navy. We're really excited to be working with them on some of their directed energy needs. And then on the commercial side, we're a proud member of the NVIDIA Inception program. And so looking at how do we scale the materials that are going to be required for Next Generation. generation GPUs.

And you also just announced that you have a bunch of launches lined up with SpaceX to do all this. Yes, we're very excited. We've got 12 launches to go ahead, so we're going to be launching 24 fab ships two per launch. And what the important part of this is being able to rapidly iterate. So really this idea of load and go, being able to put in the wafers, test them, see how they perform. you can think of this similar to the ultimate egg drop challenge. So we want to ensure not only can we get wafers to space do our manufacturing, but also on re-entry and return that we're able to successfully bring back wafers without any type of cracking or damage like that.

And so how big are these fab ships going to be and how much manufacturing are you able to do? For our clipper class variety, we're focused on being able to take. take, you know, again, these terrestrial level wafers and being able to actually scale this up for longer duration missions. And so you can imagine starting with a microwave sized for our clipper class and scaling that to kind of like a mini fridge for our next iteration. And so the idea is, again, being able to scale up that production for these different substrates and materials based on different needs. And all of this is going to be autonomous. in terms of being able to do you're saying.

No humans and bunny suits where we're going. Yeah. What is that? I mean, I realize I'm getting ahead of myself for probably a few years from even like fully appreciating or having this conversation. Well, what is this going to do to the cost to be able to manufacture some of these advanced materials?

So it is going to significantly drop the cost. As I referred to earlier, one of the core issues and really the genesis for why I even started the company was realizing that as chips continue to get smaller and smaller, the cost of the price point and then the cost to manufacture them only increases. And interesting enough, there's actually a Moore's second law. Everyone talks about Moore's law. There's actually a law that states this called Moore's Second Law, also known as Rock's Law,

which states that every four years, the cost to construct a new leading edge fab doubles. And we're seeing this happen in real time. TSM just announced the other day that their new 1.4 nanometer FAB is going to cost nearly $50 billion up from the $20 billion that was just constructed in Arizona. So we're seeing this play out and that doesn't scale. So for our approach, we're trying to strip away what actually goes into manufacturing a wafer. And when you really boil it down, you're talking about raw precursor materials, you're talking about the wafer foundations themselves and realizing that what goes into a lot of fabs in terms of capital expenditure is HVAC systems, multi-billion dollar vacuum pumps. Like I said, humans and bunny suits, a lot of solvents,

a lot of water. And for our approach, we're actually removing all of that. Wow. How'd you come up with this idea? A lot of thinking and really a combination of my prior experiences. So I got my start in the world of energy focused on microgrid. So while I was at Stanford, I led a project at Slack National Lab, funded by the DOE called the Grid Resilience and Intelligence Platform. And so we spent a lot of time thinking about how do we apply AI to anticipate, absorb, and recover from grid events. And so really spent a lot of time thinking about the energy side of the equation. And then joined Open AI and quickly saw, oh my gosh, how much compute is going to be required. And again, we're seeing this play out with now all the major hyperscalers going after massive amounts of compute.

And again, a lot of people are looking at the energy side of the problem. But there's also another piece of the equation, which is the material science side. And that really boils down to who is actually going to manufacture all of the chips that need to go into all of these data centers to provide that compute. And that's really where I started to quickly realize, wow, okay, that's a pretty short list. And for the future of materials, there are very few people, if anyone, manufacturing compound semiconductors, so non-silicon-based wafer.

So these are gallium nitrate, diamond, aluminum nitrate, etc. And the benefit of them is that they have higher thermal conductivity. What this means is that they can operate and perform in much higher temperatures. So there's, you know, the meme about GPUs melting. There's a reason for that. And that's because of the materials that are used. because what's happening inside data centers with all of this AI, all of this compute, they're running at 24-7 throttling.

Most data centers were not anticipating that, you know, if you asked them a few years ago. And so what used to be a very nice bell curve, now you're just having that 24-7 throttling. And so seeing this convergence between the demand for AI and compute, the materials at the wafer level that were being manufactured and who was going to, you know, provide that. And then finally, seeing the rapid amount of launches available and the cadence provided by SpaceX and many other companies to be able to access and go to and from space on a regular basis. And so factoring those three things in really gave me the inspiration to start, Baskar, and build off of a lot of the work done by NASA in the early 2000s and late 90s with Wake Shield facility as well. It's super fascinating. I mean, at a time where there's obviously a lot of focus on bringing more semiconductor manufacturing back to the U.S., when I hear about your timeline, I have to think, and knowing how complicated some of these fabs are, these multi, multi-billion dollar fabs that are being designed and developed and constructed here in the U.S., I have to think that you're going to be giving some of them a run for their money and pretty quickly, potentially, if all goes according to plan here.

Yeah, I mean, so really at the end of the day, China is making leaps and bounds in this area already. So actually it was just announced last year that on China's space station, they already have been looking into hand manufacturing compound semiconductor waferes primarily GAN, which they control 98% of the world supply of gallium for both defense and commercial markets. So this is already happening. So we are throwing our hat in the rink. And we are, you know, trying to really help, you know, provide a domestic supply of semiconductors. And more importantly, really make the case that in-space manufacturing is American manufacturing. Now, you're two years in stealth since you started this company. What does that look like?

Are you venture funded? I guess how to think about how to think about that? Yeah, we are venture funded. Our partners include B partners, union. labs and then a bunch of different strategic angels ranging from SpaceX, Open AI, Anderol, Slack National Lab, NASA, and they've all been able to help us in terms of recruiting, in terms of partnerships, in terms of being able to move forward with accessing different types of materials, et cetera. And so the amount of support that we've received has been incredible.

How long is it going to take, once you're up and running from a production standpoint, How long is it going to take to actually be able to manufacture wafers in space? Well, it's a great question, and it depends on which process. So for compound semiconductors, there are really three core modalities or techniques that are used to manufacture these wafers. They also depend on the materials, as I mentioned, whether it's gallium nitrate, diamond, etc. All of that being said, much like ink cartridges, we can pick and choose which ones that we utilize and swap in and out. That's part of the modularity that we're going after. So we can run cycles anywhere from less than an hour runs to a matter

of days to a couple of weeks. So again, with modularity in mind, we can kind of throttle up and down as needed depending on the different material needs and obviously sizes of those wafers as well. So what are the next milestones we should be watching for for Vexar? Yeah. First and foremost, our first launch coming up, so we're very excited. We're looking to get that off the ground before the end of the year. And then we've got all remaining, you know, the 11 after that. So we're very excited for that to becoming valid, you know, getting flight heritage. The other really exciting thing about what we're doing is we are pioneering the first ever reusable payload program on a SpaceX rocket. So what this means is we actually are going to be

reusing our fab ships in between our launches. So again, coming back to that idea of, you know, ink cartridges for a printer. We're going to be swapping out our materials and things like that, but the hardware itself will be reflown. And so we're going to be learning along the way and understanding the integrity of our hardware, because the idea is to take the data of that hardware and be able to use that to scale up to our next version for longer duration missions and higher production throughput. So that's really going to be the key milestones is getting through this flight campaign and validating our Clipper class. I assume you're making these fab ships in-house.

And I guess just as fascinating to me is the autonomous piece of this and what goes into that. Yeah. So the way we like to think about it, and I saw this a lot for my time at OpenAI and prior at Slack as well, is you have to start with the data. And there's a lot of folks talking about using AI for chip design, which is great. But for chip manufacturing, that's a whole different ballgame. And so one thing that we're thinking about is how do we capture data at every single step of the manufacturing process, from the precursor materials to then actually deposition and the nanofabrication and then doing defect analysis or defect detection and failure analysis and then using all that data, feeding it all the way back upstream to optimize our processes. And so that's a lot of how we think about this truly autonomous manufacturing processes that are able to continue to.

update their data, much like Tesla is going around and collecting all this data as their cars drive around, that's actually how we think about our fab ships, is continuing to improve those processes with each production run. That's super fascinating. I look forward to seeing how all of this plays out. Appreciate the time. Ashley Pillipshin of Bexar. Thanks so much for joining me today. Thanks so much for having me. That does it for this episode of Manifest Space. Make sure you never miss a launch by following us wherever you get your podcasts and by watching our coverage on closing bell overtime. I'm Morgan Brennan.