Satlyt, a burgeoning company co-founded by former Google and SpaceX product manager Rama Afullo, has successfully raised an $8 million seed round. The startup aims to revolutionize space computing by developing software that enables artificial intelligence to run directly on satellites, allowing multiple spacecraft to collaboratively handle complex data processing tasks in orbit.

Afullo’s vision for in-space computing emerged from his experiences at Google’s cloud division and a brief tenure at SpaceX’s Starlink. Despite initial internal rejections at both tech giants, he departed SpaceX to establish Satlyt, headquartered in Sunnyvale, California, and Nairobi. Unlike companies such as SpaceX, Google, or other startups like Starcloud and Cowboy Space Company, Satlyt is not focused on building its own satellite hardware. Instead, it positions itself as a software provider, akin to VMware or Snowflake, offering a platform for running sophisticated AI models without requiring users to manage the underlying space infrastructure. Afullo likens their approach to building an “Android” for the space industry, fostering an open ecosystem across various satellite platforms.

The timing for Satlyt’s innovation appears opportune, especially as SpaceX has increasingly invested in orbital data centers. Satlyt’s software is slated for an imminent launch aboard a SpaceX rocket, alongside the inaugural prototype for Google’s Project Suncatcher, an initiative focused on space data centers. The company has already demonstrated its capabilities through two prior missions, including deploying Google DeepMind’s Gemma AI model on a Momentus spacecraft, which reduced the size of onboard software error transmissions by over 60%. Such efficiencies, Afullo notes, could save hundreds of thousands of dollars per satellite annually by processing data and resolving anomalies in space rather than relying on costly and slow downlinks to Earth.

Looking ahead, Satlyt’s software will be integrated into a spacecraft built by Indian startup TakeMe2Space, serving three distinct customers: NASA, which will test cloud computing protocols in space; Stellerian, a space surveillance firm testing image processing workloads; and TakeMe2Space itself, demonstrating multi-tenant software hosting. The company’s next major goal is to establish a shared computing system spanning two different satellites by next year, paving the way for a true third-party compute cloud in orbit. This ambitious plan, backed by Non Sibi Ventures and former NASA astronaut Bernard Harris, anticipates that by the end of the decade, virtually all new satellites will incorporate GPUs and advanced processors, making Satlyt’s software an essential component for turning satellites into revenue-generating managed services.

Key Takeaways

• Satlyt, founded by a former Google and SpaceX manager, has secured $8 million in seed funding to develop AI software for satellites.
• The company focuses on providing software for in-orbit computing, enabling AI models to run directly on satellites to improve efficiency and reduce reliance on ground stations, rather than building its own spacecraft.
• Satlyt's technology has demonstrated significant cost savings and is set for an upcoming launch with SpaceX, involving customers like NASA and Stellerian, with plans to create a multi-satellite compute cloud.

Editor’s Analysis & Impact

Satlyt’s successful funding round and innovative approach signal a significant shift in the space industry towards decentralized, in-orbit data processing. By focusing on software rather than hardware, Satlyt positions itself as a crucial enabler for the burgeoning space computing ecosystem, potentially democratizing access to advanced analytics and AI capabilities for satellite operators. This move could drastically reduce operational costs and latency associated with traditional ground-based data processing, fostering new business models for satellite services. The future outlook suggests a rapid integration of GPUs into satellites, making Satlyt’s platform increasingly vital. This development could accelerate the commercialization of space, enabling more sophisticated applications from Earth observation to space surveillance, and establishing a true ‘cloud in orbit’ as a third-party utility.

Frequently Asked Questions

Q: What is Satlyt's core business?
A: Satlyt's core business is developing software that allows artificial intelligence models and complex computing tasks to run directly on satellites in orbit. This approach aims to reduce the need for data to be sent back to Earth for processing, saving time and money.

Q: How does Satlyt differ from companies like SpaceX or Google in space computing?
A: Unlike SpaceX or Google, which are investing in building their own orbital data centers or spacecraft, Satlyt focuses exclusively on creating the software platform. It aims to be an 'Android' for space computing, providing a horizontally integrated, open ecosystem that can operate across various satellite hardware platforms, rather than building its own.

Q: What are the main benefits of running AI on satellites?
A: Running AI on satellites offers several key benefits, including significant cost savings by reducing expensive data downlinks to Earth, faster decision-making for anomaly resolution, and more efficient processing of sensor readings directly in space. This enhances operational efficiency and enables new, more responsive applications for spacecraft.