How an E-Scooter Entrepreneur Raised $5 Million to Launch Space Data Centers

A crucial metric to observe for SpaceX’s potential IPO this week is how significantly the company has shifted the venture capital viewpoint on long-term, capital-intensive space projects. It now enables a proficient founder, regardless of prior space experience, to obtain funding for a space data center business.

Orbital, a recent entrant that emerged this May from a16z’s startup accelerator Speedrun with a $5 million seed investment, is the latest firm aiming to conduct inference in space — as soon as Starship becomes fully operational. The diverse group of investors includes Basis Set, Human Element, Wayfinder, Antler, Anti Fund, Ascent, Rubik, Zero Knowledge Ventures, LYVC, Feld Ventures, New Legacy, FNDR, UpHonest, and Asterisk.

Euwyn Poon, founder and CEO, previously created the e-scooter company Spin in 2017, which he sold to Ford the following year to join the automotive giant. When he was ready to start a new project, a16z’s Speedrun was eager to back him. Partner Andrew Chen shared with TechCrunch that Poon considered various ideas before focusing on the concept of space data centers.

You’re probably familiar with the pitch: the demand for AI computation is relentless, and deploying it on Earth can be slow. Why not explore space for endless sunlight and fewer environmental constraints? The main obstacle is the challenging economics related to launching materials into orbit, which presently complicates the business case.

Orbital is, like many rivals, depending on SpaceX successfully launching its Starship rocket for commercial use. “We will scale fully when Starship is operational,” Poon clarified. The current pricing of the Falcon 9 makes it “not economically viable.”

Currently, Poon and his team — about a dozen individuals based in Los Angeles, with backgrounds at Amazon LEO, SpaceX, and Northrop Grumman — are preparing for a demonstration flight aimed at launching an Nvidia Blackwell chip on a partner’s satellite to assess Orbital’s radiation shielding and thermal management solutions. The company plans to launch its first data-processing spacecraft equipped with Nvidia’s Space-1 Vera Rubin-class GPUs by 2028.

At that point, the objective is to initiate piecemeal inference work, allowing revenue generation with each satellite deployed. This approach resembles that of competitor data center startup Starcloud, which has already launched a GPU into orbit and intends to deploy more for income generation until Starship enables the launch of their complete constellation.

Orbital aims to deploy 10,000 satellites into orbit that together provide a distributed gigawatt of computing power, with each satellite contributing 100 kW. In comparison, Elon Musk stated that SpaceX’s AI satellites are expected to yield up to 150 kW, while Starcloud plans to operate larger 200 kW-rated spacecraft for chip operations.

Some companies are too eager to wait for Starship’s rollout. Cowboy Space Company, another space data center startup backed by a16z, has recently decided to build its own rockets. Jeff Bezos’ Blue Origin has also announced plans to launch data centers into space with its New Glenn rocket.

Poon is optimistic that the immense demand for AI will enable multiple companies to succeed. “There are various pathways for businesses in our industry to explore,” he told TechCrunch, outlining a range of options that include different AI workloads, designs, and visions for a space data center.

Chen pointed out that Poon’s background in scaling a company that deployed 250,000 scooters across 100 cities showcases his capability to navigate the complexities of building an aerospace venture. Although such a project could take a decade and cost $5 billion or more, Chen noted that venture firms are growing more comfortable with these timelines.

“This type of initiative would have seemed absurd a decade ago when we were all focused on mobile apps,” he remarked. “Launching it in 2026 allows you to tap into the current enthusiasm and energy in the capital markets.”

Poon’s foray into the space data center arena was somewhat unorthodox. After leaving Ford, he whimsically acquired an Nvidia A100, colocated it in a Santa Clara data center, and provided open-weight models. This hands-on experience reinforced his belief in the importance of delivering computational power in the AI era.

Now, he simply needs to position a few thousand GPUs in space.