Key Takeaways
- Cowboy Space Corp., formerly Aetherflux, raised a $275 million Series B round led by Index Ventures to fund its orbital AI‑compute infrastructure.
- The company plans a constellation called Stampede that will deploy solar‑powered AI data centers in space, each housed in a rocket upper stage that becomes a 1‑megawatt compute hub on orbit.
- Cowboy intends to own its full manufacturing chain and operate dedicated launch sites to accelerate launch cadence and meet surging AI demand that outpaces terrestrial power‑grid expansion.
- Technical expertise comes from alumni of SpaceX, Astranis, Kuiper, and NVIDIA, which is supplying an AI‑focused chip platform for the orbiting data centers.
- Near‑term milestones include a power‑beaming demonstration launch later this year, an optical‑communications cluster in 2027, and the debut of Cowboy’s own launch vehicle in 2028.
Company Origins and Rebranding
Cowboy Space Corp. traces its roots to 2024, when Robinhood co‑founder Baiju Bhatt founded Aetherflux with an initial focus on space‑based solar power. After securing a $50 million Series A round roughly a year later, the venture underwent a strategic pivot and rebranded to Cowboy Space Corp., signalling a broader ambition to serve the AI era rather than solely harvest solar energy. The name change reflects the company’s new identity as a builder of orbital infrastructure designed to deliver high‑performance computing directly from space.
Funding Round Overview
In May 2026 Cowboy announced a $275 million Series B financing round, led by the venture‑capital firm Index Ventures. The capital is earmarked for the development, testing, and deployment of the company’s flagship “Stampede” constellation, as well as for the creation of a proprietary launch vehicle whose upper stage doubles as an orbiting data center. This sizable infusion underscores investor confidence in Cowboy’s unconventional approach to solving the compute‑power bottleneck that AI workloads are creating on Earth.
Strategic Vision: Orbital Infrastructure for AI
Cowboy’s leadership argues that terrestrial energy grids cannot keep pace with the explosive growth of artificial intelligence. In a May 11 statement posted on X, the company noted that average lead times for new data‑center connections in major U.S. markets stretch five to seven years—or longer—while AI demand continues to outstrip existing capacity. By moving compute resources to orbit, Cowboy aims to bypass ground‑based constraints, offering a scalable, low‑latency alternative that can be deployed far more rapidly than traditional data‑center construction.
Stampede Constellation Concept
The centerpiece of Cowboy’s plan is the Stampede satellite constellation, a fleet of spacecraft each equipped with solar panels and AI‑optimized hardware. Although the exact number of satellites has not been disclosed, the architecture envisions a mesh of orbiting nodes capable of performing high‑performance computing tasks and transmitting results to Earth via optical (laser) links. The constellation is designed to function as a distributed super‑computer, providing on‑demand compute power for AI training, inference, and data‑intensive workloads across the globe.
Integrated Rocket‑Data Center Design
A hallmark of Cowboy’s technology is the tight integration of launch vehicle and payload. The company is developing a homegrown rocket whose upper stage, after delivering the satellite to orbit, remains attached and is repurposed as a solar‑powered AI data center. Once in space, this stage operates as a 1‑megawatt hub, delivering continuous compute capability while benefiting from the constant sunlight available in orbit. This “one‑piece” approach reduces mass, simplifies deployment, and lowers the cost per unit of orbital compute relative to separate satellite and launch‑vehicle architectures.
Manufacturing and Launch Strategy
To achieve rapid scaling, Cowboy intends to own its entire manufacturing chain, from component fabrication to final assembly of both rockets and satellite‑data‑center stages. Additionally, the company plans to establish dedicated launch sites that will support a high cadence of flights, minimizing reliance on external launch providers and reducing turnaround time between missions. By controlling both production and launch infrastructure, Cowboy aims to achieve a launch frequency that can keep step with the accelerating demand for AI compute.
Technical Partnerships and Talent
The engineering team behind Cowboy draws on deep experience from leading aerospace and technology firms. Alumni of SpaceX, Astranis, Kuiper, and NVIDIA have joined the effort, bringing expertise in launch systems, satellite communications, and advanced semiconductor design. Notably, NVIDIA is supplying an AI‑focused chip platform that will power the orbiting data centers, ensuring that the hardware is optimized for the matrix‑heavy workloads characteristic of modern AI applications.
Near‑Term Milestones
Cowboy’s roadmap begins with a demonstration launch later in 2026, aimed at proving the feasibility of beaming power from space to Earth—a critical step for validating the energy‑generation component of its orbital data centers. A follow‑on mission scheduled for 2027 will deploy a cluster of satellites to test high‑speed optical (laser) communications, laying the groundwork for low‑latency data links between orbit and ground stations. Finally, the company expects to fly its inaugural launch vehicle in 2028, marking the transition from payload‑only missions to a fully integrated launch‑and‑compute service.
Philosophical Shift: First‑Principles Approach
Baiju Bhatt has described Cowboy’s solution as a “first‑principles departure from the traditional [satellite] constellation model.” Rather than simply adding more satellites to an existing architecture, the company reexamines the fundamental constraints of power, cooling, and data transmission that limit terrestrial data centers. By leveraging the unique advantages of the space environment—uninterrupted solar flux, vacuum cooling, and line‑of‑sight optical links—Cowboy seeks to construct a compute platform that fundamentally outperforms ground‑based alternatives for AI workloads.
Implications and Market Outlook
If successful, Cowboy Space’s orbital AI data centers could reshape the landscape of high‑performance computing, offering a viable alternative to the prolonged lead times and energy constraints that currently hinder expansion of terrestrial facilities. The ability to deliver on‑demand, scalable compute from orbit may attract hyperscalers, research institutions, and AI‑focused startups seeking to accelerate model training and inference without waiting for new grid connections or power‑plant construction. Moreover, the company’s vertically integrated model—spanning rocket design, satellite manufacturing, and launch operations—could set a new precedent for how space‑based services are developed, potentially lowering costs and increasing access to orbital infrastructure across a broader range of industries.