A growing group of startups now sees orbit as a future destination for computing capacity, betting that space-based infrastructure could eventually help address the world’s escalating demand for artificial intelligence processing. The latest entrant, Orbital, arrives with a proposition that would have sounded speculative only a few years ago but now attracts serious venture capital backing. The company emerged from Andreessen Horowitz’s Speedrun accelerator program in May with a $5 million seed financing round and a mission centered on deploying computing systems beyond Earth. Its formation reflects a broader shift in investor attitudes toward ambitious aerospace projects that require substantial capital and lengthy development timelines. That evolution has become one of the most notable consequences of the commercial space industry’s rapid maturation.
Venture Capital Embraces Long-Horizon Space Infrastructure
Orbital enters the market at a moment when investors appear increasingly willing to support infrastructure concepts that may require years before generating meaningful returns. The company attracted participation from Basis Set, Human Element, Wayfinder, Antler, Anti Fund, Zero Knowledge Ventures, and more ventures alongside. Such backing illustrates how venture capital has broadened its appetite beyond software businesses that can scale rapidly with limited physical assets. Comments from investors backing Orbital suggest growing interest in space infrastructure projects that require longer development timelines than traditional software startups. The emergence of orbital computing startups reflects a growing willingness among some investors to explore space-based infrastructure opportunities tied to the expanding AI economy. Consequently, Orbital’s fundraising demonstrates that founders with experience outside aerospace can attract backing for ambitious space-related ventures when investors see a compelling market opportunity.
That dynamic helped Orbital founder and Chief Executive Officer Euwyn Poon move from urban mobility to orbital infrastructure. Poon previously founded electric scooter company Spin in 2017 before selling the business to Ford the following year and joining the automaker. After departing Ford, he explored multiple startup concepts before settling on the idea of space-based computing infrastructure. His eventual focus aligned with growing industry concerns about the availability, cost, and deployment speed of AI computing capacity. Rather than viewing AI infrastructure as a terrestrial challenge alone, Poon concluded that future demand could justify entirely new approaches to where computing systems operate. The resulting company represents an effort to connect the economics of artificial intelligence with the expanding capabilities of the commercial space sector.
AI Demand Pushes Infrastructure Beyond Earth
The underlying thesis supporting orbital computing remains straightforward. Artificial intelligence models require increasingly large amounts of compute power, while deploying conventional data centers often involves lengthy permitting processes, power constraints, environmental reviews, and land acquisition challenges. Proponents of orbital infrastructure argue that space offers continuous solar energy access and removes several physical limitations associated with terrestrial development. However, the industry continues to confront a significant economic obstacle. Launching large amounts of hardware into orbit remains expensive enough to challenge the financial viability of many proposed business models.
Orbital’s strategy depends heavily on future reductions in launch costs. The company expects meaningful scaling only after SpaceX’s Starship vehicle begins operating regularly for commercial customers. “We will get to full scale when Starship comes online,” Poon explained. He added that current launch economics do not support large-scale deployment, noting that Falcon 9 pricing “makes this not economically feasible.” That perspective is shared across much of the orbital computing sector, where many business plans rely on dramatically cheaper launch capacity before full deployment becomes practical. Therefore, the timing and performance of next-generation launch systems could play a decisive role in determining whether space-based data centers evolve into a viable industry.
Orbital Begins With Demonstration Missions
While waiting for launch economics to improve, Orbital is concentrating on technology validation and incremental milestones. The company has assembled a team of roughly a dozen employees in Los Angeles, drawing talent with experience from organizations including Amazon’s LEO, SpaceX, and Northrop Grumman. Its near-term objective involves testing critical systems needed to operate advanced computing hardware in the harsh conditions of space. Radiation protection and thermal management remain among the most significant engineering challenges facing orbital computing platforms. Successfully addressing those issues will be essential before companies can deploy larger and more powerful processing systems beyond Earth.
The company’s first major demonstration mission aims to place an Nvidia Blackwell chip aboard a partner satellite. Orbital plans to use the mission to evaluate shielding technologies and cooling architectures designed specifically for orbital environments. The test represents an important step toward the company’s broader ambition of launching dedicated data-processing spacecraft later in the decade. In 2028, Orbital hopes to deploy its first computing satellite powered by Nvidia’s Space-1 Vera Rubin-class GPU technology. That milestone would move the company from technology validation into the initial stages of commercial operations. Furthermore, it would provide a foundation for generating revenue before larger constellation deployments become possible.
Building Revenue Before Full Constellation Deployment
Rather than waiting for a massive orbital network before entering the market, Orbital intends to pursue a phased commercialization strategy. The company plans to perform piece-wise inference workloads once its initial spacecraft become operational. This approach would allow individual satellites to contribute revenue as they enter service rather than requiring a complete constellation before monetization begins. Such a model mirrors strategies adopted by other companies pursuing orbital computing infrastructure. The objective is to establish commercial proof points while broader launch and deployment capabilities continue to mature.
Competitor Starcloud has adopted a similar path, already operating a GPU in orbit and planning additional launches before large-scale deployment becomes feasible. These early missions serve as both technical demonstrations and market validation exercises. They also help companies build operational expertise that could prove valuable once orbital computing systems become substantially larger and more complex. Meanwhile, incremental deployments provide observable technical and operational milestones as companies advance toward larger-scale systems. The gradual buildout strategy reflects the practical realities of developing capital-intensive aerospace infrastructure in an emerging market.
A Vision for Gigawatt-Scale Orbital Compute
Orbital’s long-term ambitions extend well beyond a handful of satellites. The company ultimately seeks to deploy approximately 10,000 spacecraft capable of delivering a distributed gigawatt of computing power. Under that vision, each satellite would provide roughly 100 kilowatts of power dedicated to computational workloads. Such a network would rank among the larger commercial space infrastructure concepts currently being proposed by private-sector companies. Achieving that scale would require major advances in launch economics, satellite manufacturing, operational reliability, and capital availability. Nevertheless, founders and investors increasingly discuss these targets as long-term possibilities rather than distant science fiction.
The proposed architecture would place Orbital among several companies pursuing large-scale orbital computing networks. SpaceX has indicated expectations for AI-oriented satellites capable of generating up to 150 kilowatts of power, while Starcloud envisions spacecraft rated at approximately 200 kilowatts. These differing approaches highlight the experimental nature of the sector, where companies continue evaluating alternative designs and deployment models. Yet all participants share a common belief that future AI demand could justify infrastructure projects of unprecedented scale. As a result, orbital computing has evolved from a fringe concept into a competitive arena attracting founders, investors, and aerospace engineers alike.
Competition Intensifies Across the Orbital Compute Sector
The industry’s growing momentum has encouraged multiple companies to pursue different strategies for reaching orbit. Some startups have concluded that waiting for future launch systems introduces too much uncertainty into their plans. Cowboy Space Company, another venture backed by Andreessen Horowitz, recently decided to develop its own launch vehicles rather than rely entirely on external providers. The decision underscores how access to space remains one of the industry’s most critical bottlenecks. Companies seeking orbital infrastructure often face strategic choices not only about computing architecture but also about transportation capabilities.
Established aerospace players are also entering the conversation. Blue Origin has announced plans involving orbital data center deployments using its New Glenn launch vehicle. The participation of larger companies signals growing confidence that demand for off-world computing infrastructure may eventually support multiple business models. Poon believes the market opportunity is broad enough for numerous participants to succeed simultaneously. “There’s so many lanes for companies in our space to pursue,” he told TechCrunch, before outlining different approaches to AI workloads, spacecraft designs, and infrastructure concepts. Accordingly, the sector appears less focused on a single dominant architecture and more centered on experimentation across a wide range of technical and commercial frameworks.
From Scooters to Satellites
Supporters of Orbital argue that Poon’s entrepreneurial background provides relevant experience despite his nontraditional route into aerospace. Andreessen Horowitz partner Andrew Chen pointed to Poon’s role in scaling Spin across 100 cities and deploying approximately 250,000 scooters as evidence of operational capability. Building aerospace infrastructure differs dramatically from operating a mobility platform, yet both require managing logistics, supply chains, manufacturing relationships, and large-scale deployments. Investors increasingly view those execution skills as transferable across industries. Consequently, founders with records of scaling complex businesses are finding opportunities in sectors once dominated exclusively by aerospace veterans.
Chen also highlighted how dramatically venture capital attitudes have changed toward ambitious infrastructure projects. “This kind of thing would have sounded crazy 10 years ago when we were all building mobile apps,” he said. “Starting it in 2026 just lets you tap into all the energy and excitement that’s happening in the capital markets.” His comments reflect his view that advances in commercial launch capabilities and growing AI infrastructure demand have increased investor interest in larger-scale technology projects. Investors such as Andreessen Horowitz have publicly shown greater interest in long-duration infrastructure opportunities than was common during the mobile-app-focused venture cycle of the previous decade. That shift may prove just as important as technological advances in determining the future of orbital computing.
The Compute Opportunity That Changed a Founder’s Direction
Poon’s path toward orbital infrastructure began with a practical experiment rather than a space mission. After leaving Ford, he purchased an Nvidia A100 GPU and colocated it inside a Santa Clara data center to serve open-weight AI models. The experience provided direct exposure to the economics of compute delivery during a period of rapidly rising demand. Operating hardware firsthand offered insight into where value accumulates within the emerging AI ecosystem. Rather than focusing on applications alone, Poon became increasingly interested in the infrastructure layer that powers them.
That realization eventually led him toward a much larger vision. Instead of simply operating GPUs on Earth, Orbital aims to explore whether future generations of compute infrastructure could function beyond the planet altogether. The company still faces substantial technical, economic, and operational hurdles before that vision becomes reality. Yet the willingness of investors to support the effort demonstrates how profoundly the AI boom and commercial space industry have reshaped entrepreneurial ambition. A decade ago, building a space-based data center company would have remained largely theoretical. Today, entrepreneurs are raising capital, hiring engineers, and preparing hardware with the goal of putting thousands of GPUs into orbit.
