THE POST
Larry Page has made many accurate predictions and this is one of them https://t.co/HE2ro6Z2ey
QUOTED
Hosun Chung@hosun_chung
Larry Page reportedly has a number for how much computing will end up in space: five nines. As in 99.999%. @patrick_oshag heard it secondhand: "Someone asked Larry Page how much of compute will be orbital, and he said five-nines." @martin_casado's reason it might happen: "It's just basically regulatory." He thinks there are 500 proposed bills to stop data center building in the US, Texas included. From @InvestLikeBest.
https://x.com/hosun_chung/status/2119198237047590971The quoted post carries a video clip of about 43 seconds from the Invest Like the Best podcast. It is not copied here and I have not watched it.
ENGAGEMENT
- likes19,257
- reposts2,035
- replies1,429
- quotes161
- bookmarks2,797
- views2,637,324
AS RETURNED BY THE X API · FETCHED 2026-11-09 06:18 UTC
WHAT IT SAYS
Musk quotes a clip in which podcast host Patrick O'Shaughnessy recounts, secondhand, that Larry Page was asked how much compute will be orbital and answered five nines, 99.999%. Musk says Page has made many accurate predictions and this is one of them. The quoted post adds investor Martin Casado's view that the reason would be regulatory, citing about 500 proposed bills against US data-centre building.
CONTEXT
OfficeChai, which I opened, reports that Page has not confirmed the figure publicly and that O'Shaughnessy said he had heard the story rather than heard it from Page. It places the post in a week of US data-centre backlash after the midterm elections, notes Google's Project Suncatcher put a prototype TPU satellite in orbit in October and that SpaceX has filed with the FCC for an orbital data-centre system, and records sceptics, including Sam Altman, who call orbital data centres unrealistic in the foreseeable future.
On Latitude Media's Catalyst podcast, Energy Impact Partners' Jake Elder argues orbital solar can reach about 95% capacity factor and 5 to 10 times the energy per panel of ground panels, but that energy is only 5 to 15% of an AI data centre's cost, chips cost the same in orbit, and heat rejection, debris and maintenance are hard. His ten-year guess was that 5 to 10% of compute could be in space, with most still in grid-connected hyperscale sites.
WHAT IT MEANS FOR K
Moving compute to orbit would mean harvesting sunlight that never reaches Earth, which is the step past a planet-bound K. Today the claim is a secondhand quote about a distant future and adds no watts; the engineering and the launch cadence it needs are still unproven.
CLAIMS
- Larry Page expects 99.999% of compute to end up in orbit (secondhand).
- That prediction is accurate (Musk).
- Regulation on the ground is the reason it could happen (Casado, in the quoted post).
EVIDENCE
- The Page figure is hearsay recounted on a podcast; OfficeChai notes Page has not confirmed it.
- Google has a prototype TPU satellite in orbit under Project Suncatcher, per OfficeChai.
- An energy investor's analysis puts orbital compute at 5 to 10% of the total in a decade and flags cost, heat and maintenance limits.
REASONING
- The direction is consistent with real prototypes and filings.
- The specific number has no evidence behind it and the timeline is unstated.
- So the post is speculative for K.
SEE ON THIS SITE
TAGS
- COMPUTE
- SPACE
- ORBITAL-DATACENTERS
- DATACENTERS