Starcloud, a startup focused on AI inference satellites in orbit, has informed TechCrunch about a $250 million addition to its $170 million Series A funding round from March, bringing the company’s valuation to $2.3 billion.
This new capital will enable the company to establish a larger manufacturing plant and accelerate the development of Starcloud-3, its most significant orbital data center spacecraft, designed for launch aboard SpaceX’s upcoming Starship rocket. CEO Philip Johnston is also securing funds to guarantee satellite launches amid increasing constraints in the rocket transportation market.
Johnston conveyed to TechCrunch, “We anticipate needing to secure a vast amount of launch capacity.” Starcloud has already sought FCC approval to operate 88,000 spacecraft.
Johnston stated, “We aim to secure contracts for launches, such as with Starship, as soon as possible.” He added, “A significant cost now involves guaranteeing launch capacity… the launch market is currently quite restricted as SpaceX’s Falcon 9 program is slated to conclude in 2028.”
The expenses associated with launches were already among the most significant hurdles for companies developing orbital data centers, leading one startup to opt for constructing its own rockets.
SpaceX intends to gradually discontinue its primary launch vehicle and introduce the considerably larger, yet unproven, Starship rocket, complicating logistical planning for satellite operators. This challenge is heightened by the irregular flight schedules of competing rockets such as Blue Origin’s New Glenn and ULA’s Vulcan, and the fact that new vehicles like Rocket Lab’s Neutron have not yet commenced operations.
Currently, Starcloud is prioritizing the launch of two of its next-generation 8 kW compute satellites, known as Starcloud-2, via rideshare flights in 2027. These satellites are slated to execute orbital inference tasks for clients, including U.S. government entities. Starcloud is also exploring the possibility of acquiring a dedicated Falcon 9 launch for additional spacecraft and entering into agreements with other providers to facilitate upcoming missions.
Nevertheless, Starcloud’s strategy fundamentally relies on Starship’s potential to significantly reduce launch expenses, enabling the creation of an orbital inference layer capable of rivaling ground-based data centers. Johnston expresses ongoing confidence in SpaceX’s capacity to prove the rapid and frequent reusability of the world’s most powerful rocket.
This week, Elon Musk, CEO of SpaceX, announced that the company would postpone its attempt to catch a returning Starship rocket for several months, with the first re-flight of the vehicle now anticipated in late 2026 or early 2027.
Johnston remarked, “Clearly, a lack of available SpaceX launch capacity in 2029 would present a significant challenge for us.”
The extended funding round for Starcloud was spearheaded by Manhattan West Ventures, with contributions from Nvidia and Cisco. An individual knowledgeable about the transaction indicated that Nvidia invested $25 million in Starcloud. Additional investors comprised Benchmark, EQT, Soma, NFX, 776, Cedar Capital, Goanna Capital, and Standard Capital.
Johnston highlights Nvidia’s investment as a crucial indicator of Starcloud’s competitive edge in the emerging space computing industry. Starcloud is reportedly the sole company to date operating a Nvidia H100 terrestrial data center GPU in orbit and the first to use it for model training; most other space-optimized GPUs are tailored for edge processing. Starcloud is collaborating with Nvidia, sharing insights as the chipmaker develops its inaugural dedicated space GPU, the Vera Rubin Space-1 chip.
He explained to TechCrunch, “Their decision to invest now stems from the extensive data we acquired from Starcloud One.” He added, “They conducted considerably more technical due diligence on this than any other venture capitalist.”
While the space-optimized chip is still under development, Starcloud aims to launch it into orbit by late 2028. Johnston notes that his engineers are focusing on critical design considerations, including the correlation between the chip’s operating temperature and the dimensions of its heat-dissipating radiators, the strategic placement of radiation shielding, and the necessary reinforcement for the chips to withstand the intense forces of a rocket launch.
The company, which currently has 25 employees and is expanding, is establishing production lines at a 100,000-square-foot facility in Woodinville, Washington, close to where SpaceX and Amazon manufacture satellites for their communication networks.