Here’s one way to gauge the impact of SpaceX’s upcoming IPO: The company has so thoroughly transformed the venture capital world’s view on long-term, capital-heavy space projects that even a founder with no background in aerospace can raise money to build a data center in orbit. Orbital, a startup that emerged in May from a16z’s Speedrun accelerator program with a $5 million seed round, is the latest to promise AI inference in space—once Starship begins regular flights. Other backers include Basis Set, Human Element, Wayfinder, Antler, Anti Fund, Ascent, Rubik, Zero Knowledge Ventures, LYVC, Feld Ventures, New Legacy, FNDR, UpHonest, and Asterisk.

Founder and CEO Euwyn Poon previously launched the e-scooter company Spin in 2017 and sold it to Ford a year later, joining the automaker. You’ve probably heard the pitch before: Demand for AI compute is insatiable, and deploying it on Earth is slow. Why not move to space, where there’s endless sunlight and minimal environmental review? The main hurdle is the punishing cost of launching things into orbit, which currently makes the business case unworkable. Like many competitors, Orbital is betting that SpaceX will perfect its Starship rocket and offer it to commercial customers. “We will get to full scale when Starship comes online,” Poon explained. The price of the Falcon 9, the current industry standard, “makes this not economically feasible.”

For now, Poon and his team—about a dozen people in Los Angeles with experience at Amazon LEO, SpaceX, and Northrop Grumman—are working toward a demo flight. That test will involve flying an Nvidia Blackwell chip on a partner’s satellite to validate Orbital’s radiation shielding and thermal management technology. By 2028, the company hopes to launch its first data-processing spacecraft, equipped with Nvidia’s Space-1 Vera Rubin-class GPUs. At that point, Orbital aims to begin piece-wise inference work, generating revenue with each satellite deployed. This mirrors the path of rival Starcloud, which already has a GPU in orbit and plans to launch more to generate income until Starship enables a full constellation.

Orbital’s ultimate goal is to deploy 10,000 satellites, delivering a distributed gigawatt of computing power, with each satellite providing 100 kW. For context, Elon Musk has said SpaceX expects its AI satellites to produce up to 150 kW, while Starcloud plans to field larger 200 kW-rated spacecraft to run its chips. Some companies are too impatient to wait for Starship. Cowboy Space Company, another a16z-backed space data center startup, recently decided to build its own rockets. Jeff Bezos’ Blue Origin has also announced plans to launch data centers into space using its New Glenn rocket.

Poon is confident that the sheer breadth of AI demand will allow many players to succeed. Chen noted that Poon’s experience scaling Spin—which deployed 250,000 scooters across 100 cities—demonstrates he can handle the complex task of building an aerospace company. Over the long term, a project like this could take a decade and cost $5 billion or more, but Chen said venture firms are now more comfortable with such timelines. “This kind of thing would have sounded crazy 10 years ago when we were all building mobile apps,” he said. “Starting it in 2026 just lets you tap into all the energy and excitement that’s happening in the capital markets.”

Poon arrived in the space data center business by an unusual route. After leaving Ford, he bought an Nvidia A100 on a whim, co-located it in a Santa Clara data center, and used it to serve open-weight models. That hands-on experience convinced him of the value in delivering compute during the AI era. Now he just needs to put a couple thousand GPUs in space.