src=”https://techcrunch.com/wp-content/uploads/2026/06/satellite.png” />
The Emergence of Orbital: A New Frontier in Space-Based AI Compute
The venture capital landscape is undergoing a significant recalibration, demonstrably shifting its perception of capital-intensive, long-term space ventures. This evolution is underscored by the emergence of Orbital, a new company aiming to establish data centers in space, attracting substantial seed funding despite its founder’s lack of prior aerospace experience. This development signals a broader acceptance within the industry for ambitious, frontier-pushing technology initiatives.
Founding and Funding
Orbital, which officially launched in May from Andreessen Horowitz’s (a16z) startup accelerator program, Speedrun, has secured $5 million in seed funding. The round was led by a16z, with participation from Basis Set, Human Element, Wayfinder, Antler, Anti Fund, Ascent, Rubik, Zero Knowledge Ventures, LYVC, Feld Ventures, New Legacy, FNDR, UpHonest, and Asterisk. The company’s founder and CEO, Euwyn Poon, brings a track record of successfully scaling ventures, having previously founded the e-scooter company Spin, which he sold to Ford Motor Company.
The Space Data Center Proposition
The core premise behind Orbital, and indeed many nascent space data center companies, addresses the burgeoning demand for artificial intelligence compute power. The pitch posits that traditional terrestrial data centers face limitations in deployment speed and regulatory hurdles. Space, with its perceived abundance of solar energy and fewer environmental constraints, presents an alternative. However, the primary challenge remains the prohibitive cost associated with launching payloads into orbit, a significant barrier to economic viability.
Reliance on Next-Generation Launch Capabilities
Orbital, like many of its peers, is strategically banking on the successful development and widespread commercial availability of SpaceX’s Starship rocket. Poon explicitly stated that the company’s full-scale operations are contingent upon Starship coming online, noting that current launch costs via the Falcon 9 make the economics “not economically feasible.” This dependency highlights the critical role of advancements in launch technology for the entire space data center ecosystem.
Technological Roadmap and Milestones
Currently, the Orbital team, comprising approximately a dozen professionals with experience from entities like Amazon LEO, SpaceX, and Northrop Grumman, is focused on a near-term demonstration flight. This initial mission aims to test their proprietary radiation shielding and thermal management technologies by launching an Nvidia Blackwell chip aboard a partner’s satellite. Looking further ahead, the company has set a target for 2028 to deploy its first dedicated data-processing spacecraft, equipped with Nvidia’s Space-1 Vera Rubin-class GPUs. Post-deployment, Orbital plans to offer incremental inference services, generating revenue on a per-satellite basis.
Scalability and Competitive Landscape
Orbital’s long-term vision involves deploying a constellation of 10,000 satellites, capable of delivering a distributed gigawatt of computing power, with each satellite contributing 100 kW. This ambitious scale rivals projections from other industry players, such as Starcloud, another data center startup with a GPU already in orbit, and Elon Musk’s estimated 150 kW per satellite for SpaceX’s AI satellites. The competitive environment is also marked by companies like Cowboy Space Company, which is developing its own launch vehicles, and Blue Origin’s plans to deploy data centers using its New Glenn rocket, indicating a multifaceted approach to overcoming launch-related challenges.
Market Opportunity and Founder’s Vision
Poon expresses confidence in the vastness of the AI market, asserting that it can accommodate numerous companies pursuing different niches within the space data center sector, including varied AI workloads, hardware designs, and architectural concepts. Andrew Chen of a16z views Poon’s experience in scaling Spin as evidence of his capability to manage the complexities of an aerospace venture. While acknowledging the decade-long timeline and substantial capital investment potentially required, Chen notes that venture firms are increasingly amenable to such long-term horizons, a stark contrast to the mobile app-centric focus of a decade prior.
Poon’s own journey into this domain was catalyzed by a hands-on experiment with an Nvidia A100 GPU, which underscored the fundamental value of compute delivery in the AI era. His current objective is to successfully deploy thousands of GPUs into orbit.
Business Style Takeaway: The convergence of massive AI compute demand and the evolving economics of space launch is creating a new investment thesis for infrastructure-in-orbit. Businesses and investors should monitor the interplay between launch vehicle maturation and the development of specialized space-based computing platforms, as this sector represents a potentially transformative, albeit high-risk, frontier for technological innovation.
Learn more at : techcrunch.com