“Data centers in space” may sound like science fiction, but as the SpaceX IPO augurs a space boom, tech giants and startups are betting orbital artificial intelligence is the next frontier. How to Build a Data Center in Space: https://bloom.bg/4fSeMXn Elon Musk and Jeff Bezos plan to make AI extraterrestrial. Building data centers in space is still a relatively nascent idea, but it’s gaining traction and financial backing, as energy and space constraints limit the expansion of massive computing facilities here on Earth. Earlier this year, in the latest leg of their decades-long space race, Elon Musk’s SpaceX and Jeff Bezos’ Blue Origin both announced plans to build and launch so-called orbital data centers. Learn more: https://bloom.bg/4fSeMXn 00:00-01:36 The AI data center squeeze 01:37-03:43 Starcloud’s big ambition 03:44-04:17 Why data centers in space make sense 04:18-8:15 The engineering challenges and possible solutions 08:16-09:28 Reducing the cost of launch 09:29-09:57 Global race to get there 09:58-11:47 China’s different approach 11:48-13:21 Who will control the next internet? -------- Like this video? Subscribe: http://www.youtube.com/Bloomberg?sub_confirmation=1 Get unlimited access to Bloomberg.com for just $1.99 your first month: https://www.bloomberg.com/subscriptions?in_source=YoutubeOriginals Bloomberg Originals offers bold takes for curious minds on today’s biggest topics. Hosted by experts covering stories you haven’t seen and viewpoints you haven’t heard, you’ll discover cinematic, data-led shows that investigate the intersection of business and culture. Exploring every angle of climate change, technology, finance, sports and beyond, Bloomberg Originals is business as you’ve never seen it. Subscribe for business news, but not as you've known it: exclusive interviews, fascinating profiles, data-driven analysis, and the latest in tech innovation from around the world. Visit our partner channel Bloomberg News for global news and insight in an instant.
ADVERTISEMENT
Looks like an ad video for spacex
Rough overview. The main challenges of an orbital data centre are: 1. Cooling: in space, you cannot use air or water; heat must be dissipated using enormous radiators. 2. Energy costs: solar panels and gigantic electrical systems are required to generate hundreds of MW. 3. Maintenance: replacing a faulty server in orbit is vastly more difficult and expensive than on Earth. 4. Hardware obsolescence: GPUs and servers become obsolete within a few years, but upgrading them requires new launches. 5. Latency: for many cloud applications, the distance adds delay compared to terrestrial data centres. 6. Radiation: this increases errors and the degradation of electronics. 7. Network reliability: dependence on satellite links and ground stations. 8. Space debris: risk of collisions with satellites or debris. 9. Security and geopolitics: infrastructure vulnerable to attacks or interference. 10. Business case: the biggest problem; it is currently unclear which cloud service would be economically better in orbit than on Earth. In practice, the physical challenges are difficult but (theoretically) solvable; the real crux is understanding why a customer should pay more to use a data centre in space rather than one on Earth.
Quality reporting. Saying 4km2 and showing 4km by 4km x 2. 🙄
4 square kilometers is 2km by 2km, not 4 by 4
This is a pretty lazy video by Bloomberg standards. It's just a video full of bullet points by the companies involved, and no real investigation of if this stuff is actually realistic. How many shuttle launches would it take to build a 4km wide array? And how much pollution does all those shuttle launches produce? This video just glosses over everything.
Talking about this like its realistic is insane
"We will need" Do we ?
if the lifespan of a GPU is about 3 years, how do you change the graphics cards in space? is the life of a satellite abour 3 years then its space junk. How is that humans need 25 watts of brain power and a nuclear plant can't match the same output? aren't the gains in efficiency rather than brute force?
It would be orders of magnitude cheaper and faster to build them underwater. And that really ain't happening to any significant degree.
4km x 4km is more like 16 km2
All these companies rely on SpaceX to do anything
Please drop the children's TV narrator.
The most important point is: It's not economically feasible.
Video comes out after space x ipo
Laser links in space is talked about like it's something for the future. It's already in use by SpaceX in their Starlink constellation.
so space trash is no longer a problem or we're abandoning astronomy/space science?
Feels like they tried to cram as many buzzwords into the title as possible
what s up with bloomberg just posting ads? This after that Anthropic ad. Its getting weird
How to Build a Data Center in Space: /4fSeMXn
While there's advantage of having data centers in space, I doubt it's going to happen in the next 2-3 decades, for it to be commercially successful. There are a lot of challenges to overcome. How are they solving heat dissipation in large scale? How about maintenance? Datacenters need constant maintenance, things break. These are marketing materials where they promote this is the future, only to allure investors