Welcome to Issue 10, where orbital AI meets its first cost check
The orbital-AI buildout moved from announcement to launched hardware this week, even as the first hard-nosed market check questioned how much of it the economics will support.
On the Space4AI side, the ambition and the arithmetic arrived together. Shanghai Xingshu launched the first satellites of a 1,000-unit space-computing network at the WAIC stage in Shanghai. Days earlier, Bain concluded that launch and operating costs will cap orbital data centers at a modest share of AI compute near-term — a counterweight to the enthusiasm the constellation announcements keep generating.
On the AI4Space side, applied AI showed up as programs and as hardware in the loop. NASA joined the White House Genesis Mission, turning agentic AI loose on 150-plus petabytes of mission data across a whole-of-government effort, while LatConnect 60 unveiled an intelligence-fusion layer for SWIR imagery aimed at Indo-Pacific defence. From there the story climbs off the ground: MDA Space UK was picked to supply the LEIA LiDAR for autonomous hazard avoidance on ESA's Argonaut lunar lander.
Across both sides, the theme is the same — the space-AI thesis is being tested against real deployment and real cost, not slideware.
Specifics below.
Did you know?
We have recently launched the Bits & Orbits Insights
Check out the growing set of independent dashboards, trackers, and maps on the
AI × Space frontier:
Space4AI
Shanghai Xingshu launches first Tiansuan space-based AI computing constellation at WAIC
On July 18, Shanghai Xingshu Tiansuan Space Technology Co. said it had launched the first constellation of a space-computing project that aims to deploy 1,000 satellites in orbit. The company called the launch a step toward the commercial operation of what it describes as China's first space-based computing network. The announcement was made at the World Artificial Intelligence Conference in Shanghai, which ran July 17–20 and which President Xi Jinping attended, casting Beijing as the champion of a new global AI order.
The network, named Xingshu or "Star Hub," is being jointly developed by Fudan University and Shanghai Xingshu Tiansuan, and will eventually comprise around 1,000 satellites, according to Chi Nan, an optical communications professor at Fudan. The design processes satellite data in orbit using AI before sending results to Earth. Chi said at WAIC that it would offer on-demand, time-based computing services, letting governments, researchers and companies in sectors such as energy, shipping and finance rent computing power by the hour.
Xingshu is also intended to combine satellite observations with AI to provide weather forecasts, disaster warnings, port and shipping monitoring, power grid inspections and remote-sensing services for emergency responders, meteorological agencies and natural resource authorities. It follows a series of Chinese space-computing efforts, including the 300-satellite Tiansuan constellation led by the Beijing University of Posts and Telecommunications and state-backed Zhejiang Lab's 1,000-satellite Three-Body Computing Constellation (previously covered in Issue 7). Tiansuan has launched eight satellites so far, while Three-Body deployed its first dozen satellites last year.
Bain report finds orbital data centers will hold only modest share of AI compute near-term
On July 17, Bain & Company published an analysis of the near-term market for orbital data centers, concluding that high launch and operational costs will limit ODCs to a modest share of total AI compute capacity despite heavy investor interest.
The report identifies orbital edge-AI as the most commercially realistic near-term use case, where satellites run onboard inference to cut the volume of raw data that must be downlinked to Earth.
AI4Space
NASA joins White House Genesis Mission, applying agentic AI to 150+ petabytes of mission data
On July 22, NASA announced it is joining the Genesis Mission, the White House initiative to apply artificial intelligence to complex scientific and engineering problems. President Trump created the effort through the Executive Order "Launching the Genesis Mission" on November 24, 2025, and it is led by the White House Office of Science and Technology Policy. The initiative has since expanded into a whole-of-government program spanning more than 15 federal agencies.
NASA framed its contribution around two priorities: strengthening U.S. capabilities in space, and mining more than 70 years of agency science and engineering. The agency plans to explore how AI can unlock discoveries in over 150 petabytes of data collected across decades of missions by its telescopes, satellites, orbiters, landers, and aeronautics programs. NASA says the scale and complexity of those archives make it impractical to examine every observation with traditional methods, and that AI tools could connect data across different missions, instruments, and simulations to surface patterns and improve predictions. A key mechanism is pairing NASA's mission expertise with the Department of Energy's computing and AI capabilities to shorten the path from concept to operational readiness.
"The Genesis Mission is an opportunity to turn that foundation into faster science, stronger engineering, and better mission outcomes," said NASA Administrator Jared Isaacman, adding that the agency can "advance AI tools that accelerate exploration, strengthen American leadership in space, and open new paths to understanding our planet and the universe." NASA said the goal includes developing advanced systems faster than traditional engineering allows to operate in an increasingly crowded and dynamic space environment.
LatConnect 60 unveils proprietary AI intelligence-fusion capability for SWIRSAT amid Indo-Pacific defence interest
On 24 July, Australian Earth-observation and AI company LatConnect 60 (LC60) unveiled a proprietary intelligence-fusion capability that applies recent AI advances to short-wave infrared (SWIR) satellite imagery.
The platform fuses multi-sensor EO data into actionable intelligence for ISR applications. LC60 said it is now open for demonstrations with Indo-Pacific defence and intelligence stakeholders. The unveiling comes ahead of the first launches of the company's SWIRSAT constellation, planned for late 2026.
MDA Space UK selected by OHB to supply LEIA LiDAR for ESA Argonaut lunar lander autonomous hazard avoidance
On July 21, MDA Space UK was selected by OHB to supply its LEIA 3D-scanning LiDAR sensor for the European Space Agency's Argonaut lunar lander.
LEIA feeds the Argonaut spacecraft's Hazard Detection and Avoidance (HD&A) system. During descent, the sensor generates real-time terrain maps that let the lander autonomously identify craters, boulders, and rough terrain, then select a safe landing site without ground intervention.
Till next time,
Meta-beat Column of this week
Read also about the AI Pipeline that sits at the core, producing this Newsletter, including its ups and downs of this week: