Blue Origin has sought Federal Communications Commission authorisation to launch and operate Project Sunrise, a constellation of satellites that would provide in-space computing services. The company is proposing an orbital data centre system with up to 51,600 satellites, representing the latest corporate bet that artificial intelligence's insatiable appetite for computing power can be solved by looking upward rather than building more data centres on Earth.
The proposal deserves scrutiny beyond its technical specifications. Like other companies that have announced plans for orbital data centres, Blue Origin argues that a satellite constellation is the best approach to meet the growing computing requirements, and thus power demand, of artificial intelligence applications. The pitch is superficially attractive: satellites powered by uninterrupted solar energy, avoiding the land costs, grid strain, and permitting nightmares that plague terrestrial data centre development. Yet the plan raises deeper questions about fiscal prudence, regulatory effectiveness, and whether private ambition can outpace genuine technical feasibility.
Blue Origin is not alone in this pursuit. In late January, SpaceX filed an application with the FCC for a constellation of up to one million orbital data centre satellites. The FCC also recently accepted for filing a proposal from Starcloud, a startup focused on developing orbital data centres, for a constellation of up to 88,000 satellites. The ambition is undeniable; the coordination is almost non-existent.
Several legitimate concerns deserve serious attention. Sun-synchronous orbit is the single most congested highway in low Earth orbit, and objects in this orbit are the most likely to collide with other satellites or debris. The more objects and debris there are in low earth orbit, the higher the probability that an accidental collision leads to Kessler syndrome, a chain reaction where debris from one collision hits another satellite, creating more debris that hits another, and so on. Around 14,000 active satellites currently orbit Earth, alongside several thousand pieces of space junk. By some estimates, the number of satellites could grow to more than 60,000 or even above 500,000 by 2040.
Environmental risks compound the engineering challenges. When satellites and other spacecraft debris burn up, they release aluminium oxide and other metals and elements, which can eventually contaminate the atmosphere, land and water. There are three main risks: changing the Earth's radiative balance, harming the ozone layer and releasing toxic elements, such as mercury and cadmium, which can rain down and pollute freshwater ecosystems.
Beyond physics and chemistry sits a stubborn economics problem. Launching all of that mass into orbit would be prohibitively expensive. To become "economically viable" would require costs to fall below $200 per kilogram, a "sevenfold reduction from current levels." That threshold isn't expected until the mid-2030s. Blue Origin's New Glenn rocket has flown twice; deployment of thousands of satellites at scale remains speculative.
There is also the matter of regulatory capacity. Low Earth orbit is largely a Wild West due to regulation lagging behind its rapid commercialisation. The FCC has received the filing and sought public comment, but the agency has no coherent framework for managing orbital congestion or environmental risk across multiple mega-constellations. International coordination through the International Telecommunications Union remains underdeveloped.
Fair-minded assessment demands acknowledging the genuine appeal of the concept. The feasibility of deploying infrastructure in orbit has improved markedly due to falling launch costs. The progress of SpaceX, particularly through reusable rocket technology, has reshaped the cost curve for sending payloads into space. With the development of heavy-lift vehicles such as Starship, the idea of placing large-scale computing infrastructure in orbit is no longer purely theoretical. What was once prohibitively expensive seems now to be edging into the realm of commercial viability.
Yet ambition and engineering reality remain misaligned. Large-scale orbital data centres remain science fiction unless some moonshot-level hurdles are overcome. While the known cost of scaling terrestrial data centres remains high, the unknown costs of sending data centres to space en masse is even higher. They are not a real solution for the investment, innovation, interconnection, permitting, and other needs of the artificial intelligence industry today.
Australia has a stake in how this unfolds. If orbital mega-constellations proceed with inadequate international governance, the consequences for satellite communications, weather forecasting, and space sustainability will be felt globally. If they proceed with proper environmental assessment and collision-avoidance protocols, they represent a genuine innovation frontier. The risk is that regulatory capture and corporate momentum carry the day before serious questions are answered.
Blue Origin's filing is not inherently reckless. But it deserves harder scrutiny from the FCC than it may receive. The company should be required to demonstrate not just technical feasibility but genuine environmental assessment, orbital traffic management integration, and realistic timelines rooted in actual launch capacity. Saying something can be done does not make it prudent to do so at the scale being proposed.