SAN CARLOS, California: Cowboy Space Corporation raised $275 million in Series B funding at a $2 billion post-money valuation on May 11, 2026, and rebranded from its previous identity as Aetherflux. The round was led by Index Ventures, which also led the company’s $50 million Series A last year, and brings total funding to roughly $365 million. Founder Baiju Bhatt, a Robinhood co-founder, is using the capital to build vertically-integrated launch vehicles whose upper stages function as 1-megawatt data centers in low Earth orbit.
The Series B included participation from IVP, Blossom Capital, SAIC, Breakthrough Energy Ventures, Construct Capital, Andreessen Horowitz, NEA, Interlagos, and Bhatt himself. The company confirmed the financing details in its official announcement.
How a Robinhood Co-Founder Ended Up Building Rockets
Bhatt left Robinhood Markets in 2024 to start what was then called Aetherflux, a space-based solar power company. According to public disclosures, his roughly 6% stake in Robinhood remains the bulk of an estimated $6 billion fortune. The original Aetherflux thesis centered on beaming clean power down from orbit. The pivot to Cowboy Space keeps that thread alive, but reframes the company around a different bottleneck: AI compute.
The strategic shift, announced alongside the funding, treats the rocket’s upper stage as the product. Instead of acting as a payload delivery vehicle that gets discarded, the upper stage stays in orbit and runs as a data center node. NVIDIA is supplying Space-1 Vera Rubin Modules for the AI workloads. The company described the architecture as a “vertically-integrated” stack covering launch, hardware, and operations. Cowboy says the first module will provide 1 megawatt of usable compute, with later vehicles scaling to multi-megawatt nodes if early flights validate the thermal and power-handling assumptions.
The pivot is also a hiring story. According to SpaceNews reporting on the rebrand, Cowboy has moved from a roughly 40-person team focused on solar to a launch-and-compute headcount mix and plans to roughly double staff over the next 18 months. The company is keeping its San Carlos headquarters and adding a test facility in Texas.
Index Ventures partner Mike Volpi said in the funding announcement that the firm doubled down because the founding team had moved faster on the hardware roadmap than the original Series A plan called for. Index has now led both of the company’s institutional rounds.
Why are founders betting on orbital data centers now?
Terrestrial data center capacity is constrained by power, cooling, and land permitting, and AI workloads are absorbing capacity faster than utilities can add it. Orbital infrastructure sidesteps those constraints by tapping continuous solar power and natural cold-sink cooling in space, which is the bet Cowboy Space is asking investors to underwrite.
That bet is happening against a backdrop of historic AI infrastructure spending. Anthropic announced a roughly $4 billion deal with SpaceX’s Colossus 1 facility for 300 megawatts of capacity last week, a sign that hyperscale demand has outrun what even the largest terrestrial operators can deliver on traditional timelines. The same week, OpenAI launched a $4 billion deployment company with TPG to push enterprise AI rollout, another signal of capital chasing compute. Cowboy’s $275 million is small in that context, but the company is selling a category, not a contract. Demand signals like Anthropic’s 80x quarterly growth are what make a multi-year orbital bet legible to generalist investors.
The cap table tells its own story. Andreessen Horowitz, NEA, and Breakthrough Energy Ventures sitting alongside Index, IVP, and SAIC is an unusual mix of generalist tech, deep-tech, and strategic capital. That breadth suggests investors are pricing in optionality across launch services, compute leasing, and energy delivery rather than a single product line.
What to Watch Next
Two near-term milestones matter. The first is the space-to-Earth power-beaming demonstration the company says will happen later this year, which carries over from the Aetherflux roadmap and is the technical proof point for the energy side of the business. The second is the first orbital launch of the data center upper stage, which Cowboy says is targeted no earlier than the end of 2028. Slip dates are the norm in deep-tech hardware, and that timeline gives the company roughly two and a half years of runway before the market judges execution.
Regulatory clearance is the other variable. Operating a high-power AI accelerator in low Earth orbit will require FCC spectrum coordination for the data link and FAA launch licensing for the upper-stage configuration, neither of which is routine for a vehicle that doesn’t separate from its payload. Cowboy has not disclosed which launch provider it will use for the demonstration flight, leaving open whether the company plans to ride share on a SpaceX Falcon 9 or develop its own integrated lift capability sooner.
For founders watching, the more interesting question is whether orbit becomes a credible greenfield for AI infrastructure or stays a Bhatt-specific bet propped up by founder capital and brand-name investors. The next leading indicator is who else raises in the same lane within the next twelve months. If a second well-funded orbital compute company emerges by mid-2027, the category is real. If not, Cowboy is alone in the sky.
