BESXAR Space Industries is a space industrialization company testing how new material components of semiconductors react in space and how that will influence the construction of future manufacturing. BESXAR leverages the environment of space to rapidly and cost-effectively prototype and produce higher-performance, more energy-efficient semiconductor materials capable of meeting future computing demands at scale. It's facilitating the creation of the manufacturing infrastructure in space with proprietary data and understanding of how new elements and chips behave in space. Manufacturing in a space vacuum eliminates gravity and atmospheric gases, allowing near-perfect crystalline materials to form. The proprietary reusable orbital foundries, called "Fabships," harness the ultra-high vacuum and microgravity of space to manufacture next-generation, ultra-pure semiconductor materials. Space-based manufacturing has moved from lab curiosity to live program with BESXAR having completed its first Fabship test, transforming space into a superior manufacturing process and the next frontier for semiconductor production. The goal is to create a domestic supply of materials and to influence the manufacturing of high-performance chips to meet future demand.
BESXAR Space Industries is raising a $10M Series Seed on a SAFE with an $85 post-money valuation cap. The round follows a $4.4M Pre-Seed raised in August 2025 on a SAFE at a $20M cap, which funded design of the Clipper-class Fabship and the first SpaceX flight campaign. Existing investors and partners include Bee Partners, Union Labs, NEA, and SpaceX. BESXAR is also a member of NVIDIA's Inception program. This round funds a specific transition: from proving semiconductor hardware can survive spaceflight to proving unlocking the next manufacturing step towards autonomously manufacturing in orbit.
The company aims to achieve the following three flight milestones before raising its Series A at year end. In addition to its completed first flight, it’s planning:
Finally, The third milestone is the parallel buildout of the Ground Loop, a matched terrestrial version of the orbital process that runs the same materials through the same recipe on the ground. The Ground Loop is what makes every flight result interpretable: without a controlled Earth baseline, no orbital wafer can be shown to be better than one made conventionally. Together, these milestones convert the platform from a survivability test bed into a working process.
Early demand is taking shape across the customers, such as power-electronics manufacturers and grid technology companies and government, who feel the materials constraint most directly. The first flight on July 5, 2026, carried a variety of substrate samples to test different options including silicon and gallium nitride (GaN) as new alternatives to silicon. The next flight will carry General Atomics, a defense contractor's payload. ABB, one of the largest manufacturers of power distribution and grid automation equipment, has stated that the current wide-bandgap of device reliability and lifetime remain incompatible with the lifetime requirements of grid components, and that improving them would accelerate adoption of equipment the United States grid needs. While the company is pre-revenue, these validation points and established relationships create a customer pipeline for BESXAR.
With 11 remaining contracted Falcon 9 flights, and a launch relationship no competitor in the category has been able to secure, BESXAR enters this capital round with a de-risked first flight and data for accelerated iteration that is unusual for in-space manufacturing. Unlike competitors that focus on dedicated spacecraft and a year per orbital data points costing eight figures, BESXAR is capital efficient, generating flight rideshare margins, reduced costs with reusable fabs and a data iteration learning cadence in weeks. Combined with a new demand to substitute Silicon with GaN to support the scale of data center power delivery, a defense supply chain and elimination of foreign supply chokepoints, the company is positioned to become the first credible domestic source of orbital-grade compound semiconductor materials.
INVESTMENT TIMELINE
Virtual Company Presentation
Thursday, August 6th, 2026
1:00 pm ET // 10:00 am PT
RSVP to Google Calendar invite
Call details available here
Final Investment Commitments Due
Thursday, August 20th, 2026
We will take commitments on a rolling basis. To secure your allocation, please submit final commitments here.
Funding & Documents Due
Thursday, August 27th, 2026
At the end of the commitment period, you will receive details regarding closing documentation and wiring instructions via Carta.
The Company's confidential financing documents and diligence materials are available for review in Carta. Please request access to data room materials at the top of the page. All documents are confidential and not for further distribution. Plum Alley Ventures Company reserves the right to not proceed with the investment opportunity if the $500,000 syndication minimum is not met.
To learn more, watch the SAFE Summit 2026 talk below from Founder & CEO Ashley Pilipiszyn on developing an orbital semiconductor manufacturing platform.
* Building Domestic Dependency for the Right Materials: Gallium nitride (GaN) is the acknowledged bottleneck in two markets accelerating simultaneously: AI data center power delivery and domestic defense electronics. In both, material quality limits performance, not device design. The demand curve doesn't need to be created, only served. BESXAR is building the domestic source that serves it.
* Solving for a Real Supply Chain Chokepoint: China produces ~99% of the world's raw gallium supply. The US produces none of it, yet nearly all of that gallium ends up in the GaN wafers America's chip and defense industries depend on. Meanwhile, Taiwan Semiconductor Manufacturing Company (TSMC), one of the few companies that makes GaN wafers, is exiting the business entirely within the next year, and the largest GaN wafer size anyone needs isn't available to buy at any price, from anyone. The supply chain America relies on for GaN barely exists, and BESXAR is helping address that.
* A Growing Data Advantage: BESXAR has 12 SpaceX Falcon 9 flights under contract, two Fabships per booster, with hardware recovered in under ten minutes. One flight is done, 11 remain. Competitors spend eight figures and roughly a year to get a single data point from orbit. Over the same 24 months, BESXAR will generate close to 10x as many. That matters because BESXAR's real asset isn't hardware, it's a growing playbook of exactly which orbital conditions produce the best material. BESXAR’s playbook only gets better with more flights.
* Validation with a Defense Customer: General Atomics is flying a hosted payload with BESXAR. Defense primes don't put hardware on unproven vehicles casually, which makes this an integration and trust milestone for the company's platform.
* Rare Founder-Team Fit: The deep technical solution that the company provides is supported by a founding operator from OpenAI, a Chief Revenue Officer who built Anduril's largest division beyond $650M in annual contract value and previously served as Budget and Policy Director for the Senate Armed Services Committee, a Head of Engineering who qualified Merlin engine reuse at SpaceX from zero to 20+ flights per engine, and a board director who took multi-junction materials from the lab to mass production while securing 60+ patents.
* Strong IP Associated with Orbital Manufacturing: Hardware can be copied. A multi- year dataset connecting orbital conditions to wafer quality can't, and every flight widens that gap. Reusing the same Fabship is also what turns this from an expensive one-off program into a real manufacturing business: after year one, the cost to fly each mission drops from $29M to $7M
BESXAR was founded in Washington, D.C. in July 2023 by Ashley Pilipiszyn, who approached the problem from the demand side rather than aerospace. As an early OpenAI employee, she saw AI compute scaling collide with power delivery and thermal constraints rather than with algorithms, leading her to conclude that the next bottleneck was moving upstream into semiconductor materials.
Solving the Supply Problem of Next-Generation Chips in Orbit
The company's thesis rests on one testable idea: semiconductor performance is set during material formation, before a chip is fabricated. In compound semiconductors like gallium nitride (GaN), silicon carbide (SiC), aluminum nitride (AlN), and diamond, performance is limited by defect density, impurity control, strain, and interface quality, not circuit design. Downstream processing can find defects but can't remove them, and larger, higher-powered devices give each wafer more chances to carry a flaw. Material quality is the ceiling.
BESXAR isn't building spacecraft to sell; it's building a materials business. The end product is a wafer, the raw slice of material that chipmakers turn into finished chips, sold to power-electronics companies, chip foundries, hyperscale data centers, and defense contractors. The near-term goal is a wafer size no one currently sells commercially, and closing that gap matters most for defense: radar, electronic warfare, and directed-energy systems all require this material to be made domestically, under security requirements that today's overseas-heavy supply chain doesn't meet.
In just over two years, BESXAR has signed a 12-flight launch agreement with SpaceX, joined NVIDIA Inception, completed its first flight carrying institutional customer samples, and finalized the Flight 3 mission baseline.
Unlike other markets, manufacturing in space and space industrialization is being formed with rapid growth. As demand across AI, electric and compute increases the market will balloon beyond the current value associated with substrates, materials and manufacturing.
Power, not compute, is becoming AI infrastructure's binding constraint. Increased voltage for chips to 800V is not supported by silicon which is a material that melts at that level. New materials such as GaN are required to support the demand and also address the supply dependency on international markets across China and Russia.
Three structural forces are converging behind BESXAR providing multiple paths to value creation rather than relying on a single catalyst.
1: AI power delivery has become the binding constraint on compute. Electricity does not reach a graphics processing unit (GPU) in the form the GPU needs. It’s converted several times on the way from the grid to the chip, and every conversion loses a fraction of the power as heat, which then has to be removed, which consumes more power still. That makes conversion efficiency a direct constraint on how much compute a given facility can run. US data center usage is expected to grow from 4.4% of the country's power in 2023 to 7-12% by 2028, per the Department of Energy. New material substrates are required for efficient utilization of the increased demand.
2: GaN is winning share on physics which makes the shift to it durable. The industry roadmap is moving from 48-volt server architectures toward 800-volt direct current (VDC) systems, a transition NVIDIA has publicly staked its next generation of infrastructure on, and silicon can't make that jump cleanly: pushing it to higher voltage means building it thicker, and a thicker device wastes more energy as heat every time it switches, which becomes a real failure risk at 800V power densities. GaN breaks that tradeoff. Its bandgap, the energy required to make the material conduct, is roughly three times higher than silicon's, and it withstands a correspondingly higher electric field before breaking down, so a GaN device can hold the same voltage while staying thin, wasting less energy as heat, and running cooler under the same load. That's why GaN scales with the industry's voltage roadmap in a way silicon doesn't, and why it's being adopted as the standard rather than treated as a niche alternative.
3: The supply base is geopolitically fragile and concentrated. The highest grade of GaN wafer ships from three suppliers, all Japanese: Sumitomo Electric, Mitsubishi Chemical, and SCIOCS. Currently, eight-inch freestanding GaN is not commercially available. Upstream of that, China controls ~99% of low-purity gallium production and is currently not accessible with Beijing's suspension of its US export ban which is anticipated to expire in November 2026 with the continued prohibition of the material for military use. BESXAR is solving for a real supply chain chokepoint and increasing US domestic dependency.
BESXAR's initial wedge is disciplined. Rather than pursuing the entire power chain, it is currently focused on GaN for power and radio frequency devices, specifically the board-level stage of next-generation AI data centers. GaN already holds a clear technical advantage towards material reliability. It will later expand into indium phosphide and diamond.
Technology for BESXAR includes the growth process of substrates and the fabs called Fabship that facilitate the growth. It is not the spacecraft that carries it to orbit. There is a real physical reason orbital growth could produce cleaner material than terrestrial growth. The core claim is that space supplies two conditions terrestrial fabs can’t easily replicate at scale, an open-boundary ultra-high vacuum and microgravity where contaminant exposure at the growth surface is significantly reduced by a factor of 100 to 1,000.
The argument for why Earth cannot simply build a better fab is four-part:
A clean environment in orbit is free. The differentiated engineering artifact is the Company’s Fabship, the autonomous process module built to convert that free environmental advantage into a manufacturing one addressing all the four concerns related to terrestrial fabrication. Flight 3 is the first real technology test of the modular Fabship as a manufacturing device.
From there the roadmap expands steadily up the value chain: cleaner initial growth layers and interfaces, thicker GaN templates, lower-defect intermediate structures, freestanding GaN, reusable seed crystals grown in orbit, and ultimately native 6-inch and 8-inch GaN substrates. While the product that does not exist today, over time, the same manufacturing platform can extend into adjacent compound semiconductors including AlN, SiC, diamond, and indium phosphide (InP), significantly expanding the addressable market without changing the underlying manufacturing thesis.
BESXAR product is selling wafers, not spacecraft. The company positions as an upstream substrate supplier into the power-electronics stack, occupying the same place in the value chain as a steel supplier rather than a car manufacturer, and selling to power-electronics makers, chip foundries, hyperscalers, and defense primes. The eventual flagship is the 8-inch freestanding GaN wafer. It is a form factor that does not exist commercially as the market tops out at 4-inch and 6-inch GaN wafers from Japan. However, the defense angle: radar, electronic warfare, and directed-energy programs need domestically produced high-power GaN that complies with the International Traffic in Arms Regulations (ITAR), and today's supply chain has its concentration point offshore.
The size of the Fabship determines the production capacity of the wafers. Larger fab ships hold more substrates plus manufacturing equipment, which increases output per mission. BESXAR’s reusable Fabships include:
BESXAR’s taking a sequential approach to product development and scale where each step retires a category of risk before the next is attempted. The first flight, completed July 2026, demonstrated that the canister survives and the instruments record usable data. The first attempt at making substrate is scheduled for flight two in December 2026. What exists today is recovered flight hardware, data from the mission and development of baseline and eleven contracted flights on which to iterate the process.
BESXAR is anticipated to hold a leading domestic supply position in a form factor nobody sells, at a moment when the two largest sources of domestic demand from AI and defense have no alternative.
BESXAR's buyers aren't purchasing space flights. They're purchasing a wafer engineered to solve a problem the industry hasn't yet had the material to solve.
A power electronics company evaluating a BESXAR wafer is not making a bet on orbital manufacturing; it is assessing the substrate that lets its device run hotter, hold higher voltage, and last longer, at a cost it can defend. That framing sets the entire commercial strategy. Today these companies face a tradeoff they cannot engineer around: they can have performance and reliability, or they can have affordability, but the current generation of materials does not give them both. BESXAR anticipates that a cleaner GaN substrate collapses that tradeoff.
BESXAR is meeting the customer where they are in their approach. They are winning the device-level companies that want BESXAR material inside their products, and let their demand create the pressure that makes foundry qualification worth doing.
Customers include:
BESXAR is a merchant materials supplier. It sells qualified GaN wafers by the palette to power-electronics manufacturers, foundries, hyperscalers, and defense primes, and does not sell spacecraft, launch services, or finished devices. A palette is the unit of launch capacity the model prices against, carrying up to 40 Fabships, so palette cost and wafers per palette are the two numbers that set unit economics. Revenue is per-wafer and volume-driven at a ~$15,000 average selling price (ASP) that the model holds flat from 2030 through 2035 rather than assuming premium capture for space-made material. That flat-ASP assumption is a deliberate conservatism and the most important thing to understand about the model: the entire margin story is built on cost compression, not on pricing power.
The cost structure is unusual for a semiconductor company because launch, not fab capital, is the dominant line item. Launch represents ~60% of fully loaded cost per wafer, which makes flight cadence the primary economic lever, since fixed program cost divided across one flight per year produces a fundamentally different unit economic than the same cost divided across six. Five drivers compound in the model: launch cadence, Fabship reuse, launch-price deflation, wafers per flight, and yield learning. Of these, yield learning is the one the company controls most directly, and arguably the most important, since the economically relevant output is not wafers returned but wafers that meet customer specification.
Reuse is the mechanism that converts a capital-intensive program into a manufacturing business. After the first production year, only about 10 of 40 Fabships per palette are new builds, dropping per-flight Fabship capital expenditure from $29M to $7M. This is why refurbishment discipline is a business-model question rather than an engineering detail: every additional flight per unit directly compresses cost per wafer. The model's progression from a modest negative gross margin in 2030 to 27% in 2031, 60% in 2033, and 70% by 2035 is almost entirely a function of cadence rising and per-unit capital falling in parallel.
The most concrete asset is the SpaceX agreement: 12 contracted missions for a first-of-its-kind reusable, booster-integrated payload program. It gives BESXAR a flight cadence that competitors cannot currently purchase, which matters because most in-space manufacturing ventures fly once, learn once, and reset the learning curve on the next mission years later.
Flight history has begun. Flight 1 launched in July 2026 as a partner mission with the University of Texas at Austin and the University of Virginia, gated on canister structural survival and telemetry capture, both of which it met. Flight 2 follows in Q3 2026 as a paying customer flight with General Atomics, carrying improved sensors and boards that both verify Flight 1 data and demonstrate the ability to integrate and fly an external hosted payload under real mission constraints. Flight 3 in December 2026 is the gallium deposition demonstration. That is three missions inside seven months, which is the cadence claim being made operational rather than asserted, and cadence is the entire strategic argument.
The company anticipates becoming a revenue-generating company in year 2 and scaling to over $600M in revenue by year 5.
Customer-side validation is early but real, and the distinction between kinds of validation matters here. General Atomics is a paying external payload customer rather than a research partner, which is the more meaningful signal. Separately, Pietro Cairoli, Head of United States Corporate Research and Development at ABB, has gone on record describing a critical and growing need for high-efficiency, long-life power semiconductor devices, and noting that the reliability and lifetime of the current generation remain incompatible with the required lifetime of many grid components. That is a named executive at a major power-distribution manufacturer articulating the exact problem BESXAR is solving, which is stronger than a generic advisory relationship.
Commercial infrastructure is being built ahead of product. BESXAR has been present at Data Center World in ABB's Innovation Alley, has a public announcement cadence tied to flight milestones, and has structured a formal evaluation package with defined characterization methods for prospective customers.
The competitive set splits three ways:
In-Space Manufacturing Peers
The most direct competitor isn’t the best-funded one. Varda has raised 2x more capital than BESXAR and flies a real reentry cadence, but its commercial focus is pharmaceutical crystallization and hypersonic reentry services, not compound semiconductor substrates. Space Forge is the better-known semiconductor analog with genuine flight heritage, but its architecture requires building and recovering its own spacecraft, which is structurally why its per-data-point cost and cycle time are higher. Astral Materials is the closest match to BESXAR’s actual thesis: a 2024 startup building mini-fridge-sized crystal growth furnaces with an explicit initial focus on GaN, funded through a NASA Small Business Innovation Research award. Astral is materially less capitalized and flies on third-party capsules, but it targets the same material with a comparable payload footprint and holds a Sierra Space agreement covering Dream Chaser flights and potential integration into commercial space station systems. Worth noting that Space Forge holds a parallel Sierra Space agreement and opened a Florida Space Coast office specifically to pursue federal semiconductor funding, contesting BESXAR’s domestic-supply positioning directly rather than tangentially. Plum Alley Ventures Company has created the following competitive analysis table:
BESXAR’s structural claim is iteration rate, and the mechanics support it, with one real counterpoint. Two Clipper canisters fly per Falcon 9 booster as integrated payloads at rideshare marginal cost, against competitors spending eight figures and roughly a year per orbital data point on dedicated free flyers. Over a 24-month window that is a difference of roughly an order of magnitude in data points generated, which matters enormously in a business whose only compounding asset is a recipe library linking orbital process conditions to measured wafer outcomes. BESXAR trades environment quality for iteration speed. That is a defensible choice at this stage, but it means competitors could reach a stronger materials result on fewer flights.
International Competition
The company notes that the closest analogs to BESXAR’s approach are from China and Russia, both of which have subsidized national programs. Pilipiszyn points to Russia’s recent retrieval of a molecular beam epitaxy (MBE) reactor that had spent six months mounted outside the ISS. That detail confirms that at least one serious state program considers orbital epitaxy worth flying real hardware for, which is external validation of the physics premise from an unexpected direction. A state-funded program with unlimited dwell, meaning the ability to sustain a process continuously without time or cost constraints, isn't a commercial competitor and will never qualify a wafer for a US foundry. It could, however, publish a materials result that reframes the category before BESXAR reaches Freighter.
Terrestrial Incumbents
Pilipiszyn describes the direct commercial field as very small: roughly three to four companies globally competing for GaN epitaxial wafer share, with two in China, one in Japan, and one in South Korea. If that holds, merchant GaN epitaxy is among the most concentrated categories in semiconductors, and it sits almost entirely in Asia. That concentration is the strongest argument for BESXAR’s domestic-supply positioning, because it means a US buyer seeking an alternative supplier currently has no domestic option to switch to rather than simply a worse one.
BESXAR Space Industries is raising a $10M Series Seed on a SAFE at an $85M post-money cap, following a $4.4M Pre-Seed closed August 2025 on a SAFE at a $20M cap. Existing investors and partners include Bee Partners, Union Labs, NEA, and SpaceX. The Pre-Seed funded Clipper-class Fabship design and the initial SpaceX flight campaign; this round funds the transition from hardware survivability to demonstrated autonomous orbital process. This precedes an anticipated $50M Series A that the company is targeting by Q1 2027.
Use of proceeds is gated to three sequenced milestones:
Ashley Pilipiszyn, Founder & CEO: OpenAI's first Technical Director to the CTO, contributing to the GPT-3, DALL-E, and Codex launches. Before that she led the DOE-funded Grid Resilience and Intelligence Platform at SLAC, applying AI to monitoring and protecting US critical infrastructure. She understands the AI compute demand curve from the inside and the grid customer from the DOE work.
Diem Salmon, CRO: Former VP of Strategy and Growth for Anduril's largest division, Air Dominance and Strike, which reached over $650M in annual contract value in 2025 in under six years. Built it into three business lines and won two programs of record, launching the YFQ-44A for the Air Force CCA program, Barracuda, and Lattice for Mission Autonomy. Previously Budget and Policy Director for the Senate Armed Services Committee. For a company whose most defensible early market is domestic ITAR-compliant GaN, and whose second flight carries a General Atomics payload, this is the highest-leverage seat on the team.
Neil Mason, Head of Engineering: Founding engineer at Antares Nuclear; ten years at SpaceX in propulsion, where he took Merlin from zero reuse to 20-plus flights per engine without inter-flight refurbishment by designing and running the qualification campaign that proved it, saving millions in capex and opex and enabling higher launch rate. Reuse qualification is exactly what Fabship economics depend on. Direct skill transfer, not adjacent.
Homan Yuen, PhD, Semiconductor Manufacturing & Board Director: Co-founder of Solar Junction, where he took a new class of III-V multi-junction materials from lab to mass production using molecular beam epitaxy. Over 60 patents in materials, semiconductors, devices, and energy. PhD and MS in Materials Science and EE from Stanford, BA in Physics from Berkeley. He is the materials credibility, he sits on the board rather than in an advisory role, and his MBE production experience maps directly onto the PA-MBE roadmap BESXAR is decomposing across flights.
Michael Ellis, Strategy and Mission Operations Advisor: Former Head of Operations for Space at Anduril and Director of National Security Space Launch at SpaceX, overseeing all DoD and NRO missions from integration through launch. Previously grew Raytheon and Iris Technology's space electronics business tenfold.
This is a strong team for the two things this company must do next: operate a reusable flight program at cadence, and sell a novel material into defense and hyperscale procurement. Yuen's board seat closes more of the materials gap than we credited on first read. What is still missing is a named operating leader for epitaxy and metrology, since third-party characterization is the gating deliverable for the entire first phase and the current requirements documents distribute that ownership across individual engineers. That hire is the most consequential one between this round and the Series A, and we are tracking the search.
BESXAR is in the early development phase. As a Seed-stage company, it’s still derisking product. It’s only after flight 3 that the company can achieve substrate development and then start securing revenue.
.png){kind=logo}