Cowboy Space Raises $275M at $2B Valuation to Build Rockets Whose Upper Stages Become Orbital Data Centers

Cowboy Space — the San Carlos, California startup founded less than two years ago as Aetherflux by Robinhood co-founder Baiju Bhatt — has raised $275 million at a $2 billion valuation in a Series B funding round announced May 11, 2026. Index Ventures led the round, having also led the company's $50 million Series A last year. New investors IVP, Blossom Capital, and SAIC participated alongside existing backers Breakthrough Energy Ventures, Construct Capital, Andreessen Horowitz, NEA, and Interlagos. Bhatt himself, who co-founded financial services platform Robinhood before starting Cowboy, also participated in the funding round. Total capital raised across the company's history is approximately $365 million. The financing positions Cowboy as one of the space industry's fastest unicorns — privately held companies valued at $1 billion or more — crossing the threshold just over a month after Starcloud, the other prominent orbital-data-center startup, hit unicorn status with its own $170 million Series A.

The Architectural Bet: Rocket Upper Stage as Data Center

Cowboy's structural differentiation is the architectural decision to design the rocket upper stage and the orbital data center payload as a single integrated vehicle from day one rather than as separate systems flown together. Bhatt described the choice in operational terms: each upper stage leverages the full mass and volume of the vehicle to package power generation, cooling, and compute together — including using the stage structure itself as a radiator. The design choice is the first public commitment by any orbital-data-center company to build a rocket purpose-engineered around the data-center mission profile rather than adapting an existing or third-party launch vehicle. Competing orbital-data-center developers, most prominently Starcloud, plan to rely on SpaceX's Starship heavy-lift rocket to deploy their orbital compute nodes. Cowboy's vertically integrated architecture eliminates the redundant mass that comes from designing a satellite to be packaged inside a separate launch vehicle, optimizing the amount of power and compute that can be delivered to orbit per launch event.

The Vehicle: Between Falcon 9 and Starship

The Cowboy launch vehicle would be larger than SpaceX's 70-meter-tall Falcon 9 but smaller than Starship, which stands at more than 120 meters, targeting a payload capacity of 20,000 to 25,000 kilograms to low Earth orbit. The scale is deliberately positioned between the workhorse Falcon 9 class and the ultra-heavy Starship class — large enough to make the integrated upper-stage-as-data-center mission profile economically meaningful (a 1-megawatt orbital data center is not a small payload), small enough to maintain a credible production-cadence and operational-cadence target relative to the development burden of a Starship-class super-heavy lift vehicle. Design work on the vehicle is underway across multiple cities, with satellite engineering in Seattle and rocket engine development in Los Angeles. Bhatt told SpaceNews that the company is designing the launch vehicle, upper stage, and orbital compute platform as a single integrated system optimized specifically for AI infrastructure in orbit, and that the level of integration the company is pursuing requires full vertical control rather than relying on external providers for core architecture.

The Mission Sequence: Apex Demo, Galactic Brain, 1MW Data Center

Cowboy has disclosed a three-step mission sequence. The first mission, planned for later in 2026, is a small satellite built in partnership with four-year-old satellite manufacturer Apex Space to demonstrate wireless power beaming from low Earth orbit to the ground using infrared lasers. Bhatt described the demonstration as a subscale technology demonstrator sized to prove the physics and validate the approach, and he framed it as a building block — with the long-term business focus on using that power in orbit for AI compute rather than primarily beaming it to Earth. The second step, planned for early 2027, is the deployment of the company's first Galactic Brain data center node using NVIDIA Space-1 Vera Rubin modules designed for AI computing in LEO. Bhatt described the Galactic Brain mission as the first step toward demonstrating orbital AI compute powered by solar energy. The third step, targeted for end of 2028, is the first launch of Cowboy's proprietary rocket carrying a 1-megawatt orbital data center — the integrated vehicle architecture's first operational demonstration.

The Team: SpaceX and Blue Origin Heritage

The Cowboy team includes two named senior hires from the established U.S. launch industry. Tyler Grinnell, the former SpaceX director of launch operations, brings operational experience from the launch vehicle company that has industrialized commercial reusable launch in a way no other company has matched. Warren Lamont, who led rocket engine and booster stage development at Blue Origin, brings propulsion and large-stage development experience from a company that has been the principal architectural alternative to SpaceX in the U.S. heavy-lift launch industry. The two hires together represent a deliberate assembly of the launch-vehicle development capability needed to deliver a Falcon-9-to-Starship-class vehicle on a credible production schedule. Bhatt himself contributes the founder profile that has been one of Cowboy's distinctive characteristics — Robinhood co-founder, with a track record of reimagining massive markets from first principles, as Index Ventures Partner Jan Hammer described it, and a lifelong passion for physics that frames space as the ultimate market opportunity for that ambition.

The Constellation Vision: Many Thousands, Potentially Tens of Thousands

Cowboy's plans call for power and data center constellations numbering in the many thousands, potentially reaching tens of thousands of satellites. The number is structurally important context: the company is not positioning Galactic Brain or the 1-megawatt 2028 data-center launch as endpoints, but as early demonstrations of architecture that scales to constellation-scale orbital AI infrastructure deployment. Bhatt told SpaceNews that the size of the power-beaming and computing constellations will ultimately be driven by the economics Cowboy seeks to validate through the upcoming technology demonstrations. The startup has not yet filed plans for the constellations with the U.S. Federal Communications Commission beyond experimental license requests. Bhatt's framing — focus on demonstrating capability before scaling to full commercial services, with initial commercial services to follow once performance, reliability, and economics are validated at scale — is the conventional sequence for a constellation-scale infrastructure deployment but is also a meaningful disclosure that constellation deployment is years away from initial commercial revenue.

The Investor Syndicate: Index, IVP, Crossover Signals

Index Ventures' decision to lead both the $50 million Series A last year and the $275 million Series B at this scale is a notable concentration of conviction by a single firm. Index Partner Jan Hammer framed the investment thesis around demand for AI compute and energy beginning to outstrip what terrestrial infrastructure can support, with Cowboy positioned to address that constraint through the orbital infrastructure architecture. The new investor additions are also structurally meaningful. IVP is one of the most established U.S. growth-stage venture firms, and its participation in a private round at a $2 billion valuation is a credible signal of pre-IPO institutional readiness. Blossom Capital adds European venture conviction with significant late-stage activity. SAIC — the publicly traded U.S. defense and government services company — represents a strategic-investor signal pointing toward defense and government applications of orbital data centers and space-based power infrastructure. Existing backers Breakthrough Energy Ventures, Construct Capital, Andreessen Horowitz, NEA, and Interlagos round out a syndicate that combines climate-and-energy conviction (Breakthrough), industrial-tech specialization (Construct), generalist venture conviction (a16z, NEA), and frontier-tech venture (Interlagos) — a deliberately diverse syndicate appropriate to a company that spans launch vehicles, space-based power, and in-orbit compute.

The Reusability Question and the Launch Market Reality

An open architectural question is how Cowboy will approach launch vehicle reusability. Bhatt described the company as designing for reusability where it makes economic sense, but with the primary focus on delivering maximum compute and power to orbit efficiently — the architecture is optimized for mass efficiency and performance, not just reuse for its own sake. The framing is structurally different from SpaceX's dominant approach to commercial launch — where reusability has been the central thesis driving cost reduction and operational cadence — and reflects the architectural reality that a launch vehicle whose upper stage becomes an operational on-orbit asset cannot return to Earth in the conventional sense, since the upper stage is, by design, a satellite. Cowboy did not disclose other details about how it plans to enter the notoriously challenging launch market, or build the cadence needed to sustain meaningful LEO operations. The execution risks are substantial — building a Falcon-9-to-Starship-class vehicle from a clean sheet over a roughly 30-month development cycle to the 2028 launch target is a meaningfully aggressive schedule by any historical launch-development benchmark.

Outlook: Demonstrations, Capital, and the Orbital AI Infrastructure Thesis

Cowboy's $275 million Series B at a $2 billion valuation is one of the largest space-industry growth-round commitments of 2026 to date and is structurally significant for several reasons. It is a public validation that institutional growth-stage capital sees orbital data centers as a credible AI-infrastructure category rather than an experimental concept. It supports a company-specific architectural bet — rocket upper stage as integrated data center — that is structurally distinct from competing orbital-data-center developers and that, if validated, gives Cowboy a defensible architectural moat against well-capitalized incumbents including SpaceX. And it demonstrates that founder profile, technical talent depth, and a clearly articulated mission sequence can mobilize substantial growth capital even before any operational space hardware has flown. The execution risks remain proportional to the ambition: a clean-sheet Falcon-9-to-Starship-class launch vehicle plus an integrated orbital data center plus power-beaming infrastructure plus thousands-to-tens-of-thousands-of-satellites constellation deployment is collectively one of the most ambitious vertical integration programs ever announced in the commercial space industry. The 2026 power-beaming demonstration with Apex, the 2027 Galactic Brain deployment with NVIDIA Space-1, and the 2028 1-megawatt data center launch will be the operational milestones against which the architectural thesis is validated or recalibrated.

Frequently Asked Questions