Blasting Into Orbit?

研报英文原文证据摘录

Blasting Into Orbit?

f May 20, leaving it 1,314

satellites short of the FCC’s 1,618 interim milestone. If we assume every scheduled launch plus

four additional Falcon 9 missions not on the manifest, Amazon could plausibly only launch ~550

satellites by July 30. Thus, Amazon is building satellites faster than it can launch them. This

continues to highlight AMZN’s need to acquire a launch capability perhaps via M&A.

8) Heavy-lift rockets remains the gating factor. Amazon’s deployment plan was built around

heavy-lift rockets. Vulcan and New Glenn together account for 2,672 of the 3,478 unflown

satellites in the manifest, or about 77% of the remaining deployment plan. Both vehicles are

currently grounded. Even if every Falcon 9 and Ariane 64 contract flew tomorrow, Amazon would

still clear only 1,110 satellites, which is 508 short of the FCC milestone. This creates a real

chance that commercial service timing (slated for September this year) slips if the heavy-lift fleet

does not return quickly.

9) Orbital data centers are interesting, but not the near-term Leo thesis. Orbital data centers

quickly became a major space-sector debate. The bull case is easy to understand. Power is a

key constraint for AI data centers on Earth, and terrestrial facilities face limits around land, water,

grid interconnect, and permitting. In orbit, solar exposure can be very high, and operators avoid

local permitting and grid constraints.

But the physics are hard. Cooling does not disappear in space. Every watt of heat has to be

pushed out into vacuum. Launch cost is also still far too high. Even at $1,500/kg, the economics

are not close to terrestrial data centers, and Quilty thinks another 5x to 10x reduction in launch

cost is needed. Radiation is another issue.