Back to Headlines
Tech
Jun 09, 2026
Analyzed by GPT OSS 120B

E‑scooter Founder Raises $5M to Launch Space‑Based AI Data Centers

AI Summary
Former e‑scooter CEO Euwyn Poon’s new venture Orbital closed a $5 million seed round to build AI compute data centers in orbit. The startup plans a 2028 demo flight and a long‑term goal of a 10,000‑satellite constellation powered by SpaceX’s Starship.

Orbital Secures $5M Seed to Pioneer Space‑Based AI Compute

Orbital, a startup spun out of a16z's Speedrun accelerator in May, announced a $5 million seed round led by a roster of venture firms including Basis Set, Human Element, Wayfinder, Antler, and others. Founder and CEO Euwyn Poon, who previously built and sold e‑scooter company Spin to Ford, will use the capital to develop a prototype satellite that runs AI inference workloads in space.

  • Lead investors: Basis Set, Human Element, Wayfinder, Antler, Anti Fund, Ascent, Rubik, Zero Knowledge Ventures, LYVC, Feld Ventures, New Legacy, FNDR, UpHonest, Asterisk.
  • Team size: ~12 engineers in Los Angeles with backgrounds at Amazon LEO, SpaceX, Northrop Grumman.
  • Immediate milestone: Demo flight of an Nvidia Blackwell chip on a partner satellite to validate radiation shielding and thermal management.

Funding Breakdown and Projected Capital Needs

The seed round provides the runway for hardware development, satellite integration, and a 2028 demonstration launch. While the $5 million covers early‑stage R&D, Poon and investors acknowledge that scaling to a full constellation could require $5 billion over the next decade, a timeline that venture capitalists are increasingly comfortable with.

Implications for AI Compute Landscape and Space Launch Economics

Orbital’s strategy hinges on the economics of SpaceX’s upcoming Starship. Current launch costs with the Falcon 9 make space‑based data centers “not economically feasible,” but a fully reusable Starship could lower per‑kilogram costs enough to justify a distributed gigawatt of compute power. The company aims for 10,000 satellites, each delivering 100 kW of power, comparable to Elon Musk’s target of 150 kW per AI satellite and rival Starcloud’s planned 200 kW units.

Roadmap to 2028 Demo Flight and Long‑Term Constellation Goals

Key upcoming milestones:

  • 2026‑2027: Complete radiation‑shielding and thermal‑management prototypes; secure a launch slot on a partner satellite.
  • 2028: Conduct the first flight carrying an Nvidia Blackwell chip, demonstrating in‑orbit inference capability.
  • Post‑2028: Incrementally launch satellites to generate revenue per unit, mirroring Starcloud’s “piece‑wise inference” model, while awaiting Starship’s commercial availability for full‑scale deployment.

Even if Starship’s timeline slips, competitors like Cowboy Space Company and Blue Origin are exploring alternative launch solutions, underscoring a broader industry push toward space‑based AI infrastructure.

Long‑Term Outlook: A Decade‑Scale Bet on Space‑AI Fusion

Analysts see Orbital’s ambition as a “10‑year, $5 billion” play that aligns with the growing demand for AI compute and the desire to bypass terrestrial constraints such as cooling and regulatory reviews. If successful, the venture could reshape the AI hardware supply chain, create a new market for satellite‑hosted GPUs, and validate the economic case for large‑scale, space‑borne data centers.