Back to threaded view
Full conversation · 15 posts

Elon Musk is either the dumbest High Tech CEO, or the most Blatant and Rapacious Grifter...

TyroneIrish 26438 posts Joined:
Posted

For the Engineers and Scientists in the audience, especially...take a look at this video and recall your College Heat Transfer Class(es)...the areas required for a Gigawatt AI datacenter are Ginormous!...and Expensive (LOTS AND LOTS of Rockets, Materials, etc.). That last part strikes me as the main reason for this Boondoggle. Closed System Cooling in Underground Datacenters is much easier and less costly...plus they're much more convenient to upgrade when new components become available, or repair/replace existing ones.

We need Government agencies that are not business partners with the Contractors and have serious Inspectors General looking out for our taxpayer dollars.

  • Level 1

    To:...Baron, T3K, Jakers, NCin2009, et al MAGAs...read my 8:02am post with the APS Summary of Musk's

    Datacenters in Space proposal...it deals with Basic Physics...like it or not, it is undeniable fact...Elon's plan is not only Ill-Conceived...but an Outrageously Wasteful use of money and resources. As American citizens...in your own 'Self Interest'... you Should be Downvoting...in public...Elon, and not the APS. Whether you will, or not, is another question.

  • Level 1

    LOL. Dumb post. Musk’s 1 GW space cluster is Not a single monolithic ground style building in space

    I.e. it is not like those datacenters we saw on earth.  

    It’s  a constellation of 100k to 1 mil  small satellites (micro-servers), distributed in the space. Each individual satellite consumes electricity probably similar to an EV, not much(But together they consume huge energy). We already have achieved far more than that in space. Just take a look at U.S., Russia…built solar powered international Space Station. ISS consumes much more electricity than an EV.

    I think I said it before. The left is like a theorist, know nothing about real world. But they like to call their idea science and call others anti science. The right is more like an engineer, understanding the real world and practical. Your dumb post is a perfect example to prove my point. 

    • Level 2

      You couldn't have watched the video and posted that...watch it and then get back to us. I'll

      simplify it for you...go to the 00:30:00 mark, where the physicist talks about the radiative cooling requirement for just One AI rack from NVIDIA. Btw, I never said the program entailed a singular behemoth...just that it was on the order of 1GWe, and Very, Very Large...as well as costly...and extremely difficult to maintain, or upgrade.

      Now, do your homework...

      • Level 3

        Stupid. We have about 17,000 satellites in space now. They have no any power or cooling issue.

        Musk’s Space datacenter will be like that, just in huge numbers. Maybe 100k satellites, maybe 500k satellites.,, or maybe 1 million satellites.

        You and your YouTube misunderstood Musk’s space datacenter (actually not just Musk’s, the idea also shared by Amazon, Google and Chinese). You use ground based datacenter to calculate future space datacenter. Very ignorant about what Musk actually talked about and meant. But I know you don’t admit your ignorance and will try to spin away from your stupid post like you did many times. Because you don’t have courage and integrity. Everybody knows it.

        • Level 4

          Not stupid at all, Eli...here is a summary of the American Physical Society's opinion on Musk's plan

          for AI Datacenters in Space...Again, you need to watch...and listen to...the video presented by a PhD in Astrophysics...

          Now...to the APS summary of this issue...btw...a serious question...what education do you have in Physics?...have you ever studied Heat Transfer?

          ________________

          AI Overview

          The American Physical Society (APS) and affiliated physicists have highlighted that Elon Musk's proposals for orbital AI data centers face a major physics and thermal dissipation bottleneck regarding how heat is removed in a vacuum. [1, 2]

          Key Physics Challenges of Space Cooling

          • Absence of Convection: On Earth, data centers rely heavily on air, water, conduction, and convection to move heat into the atmosphere. In the vacuum of space, air and liquid-to-air convection do not exist.[1, 2]
          • Reliance on Thermal Radiation: Heat in space can only escape via thermal radiation(electromagnetic waves), which requires massive surface areas. [1, 2]
          • Massive Radiator Footprints: Calculations referenced in scientific and astrophysical evaluations show that expelling the colossal heat loads generated by gigawatt-scale or multi-megawatt AI server racks requires enormous radiator panels (millions of square feet or acres of surface area), adding severe structural and launch-weight penalties. [1, 2, 3]
          • The "Space is Cold" Misconception: While the ambient vacuum is nominally cold, thin orbital gases and direct solar radiation create intense thermal management hurdles. Traditional systems like the International Space Station's ammonia loops manage only a tiny fraction of the thermal output envisioned by commercial AI clusters and have historically suffered from mechanical and material wear (such as loop or coating degradation). [1, 2]
          • Proposed Workarounds vs. Reality: Proponents and Musk have pointed to custom silicon operating at higher temperatures (e.g., higher Kelvin/Celsius thresholds) or specialized two-phase and high-temperature systems to shrink radiator footprints. However, independent physics breakdowns—including discussions featured at gatherings like the American Physical Society Summit—note that while orbital computing is not strictly forbidden by physical laws, the sheer scale of radiator mass and launch economics makes it vastly more complex than terrestrial alternatives
            • Level 6

              Here are Cost Estimates for Space-Based or Terrestrial AI Datacenters...

              Don't forget the AI Infrastructure will change over time...and lifetimes are shorter in Space than on Earth, so System Costs will increase even more up there...

              _________________

              AI Overview

              Building an AI data center in space is estimated to cost 3 to 100 times more than building an equivalent facility on Earth, depending on the scale, launch prices, and assumptions used. [1, 2]

              Cost Comparisons by Estimate

              • Boston Consulting Group (BCG) Estimate: Orbital data centers carry a 2.5x to 3x cost premiumtoday, which may narrow to about 1.5x over the next decade as technology improves. [1]
              • IEEE Spectrum / Industry Baselines: A 1-gigawatt orbital site is estimated at over $50 billioncompared to roughly $16 billion on the ground (around 3x higher). [1]
              • Google Energy Estimates: Google energy executives state that space data centers can cost up to 100 times more than terrestrial ones under current economic conditions, citing extreme launch and operational hurdles. [1]
              • SemiAnalysis Building Block Analysis: A small AI server "building block" costs about $100,925 per month in space compared to $27,724 per month on Earth, largely driven by launch expenses and a shorter asset lifespan in orbit (5 years vs. 15 years on Earth). [1]

              Why Space Data Centers Cost More

              • Launch Expenses: Lifting heavy AI hardware, specialized shielding, and massive copper/semiconductor arrays into orbit remains expensive, though proponents argue reusable rockets (like Starship) could eventually lower per-kilogram costs. [1, 2]
              • Shorter Hardware Lifespan: Equipment in space faces harsh radiation, thermal swings, and difficult maintenance, leading to a much faster replacement cycle than ground-based servers. [1, 2]
              • Cooling and Infrastructure: Rejection of massive thermal loads from high-density AI chips requires huge deployable radiators in a vacuum rather than traditional water or air cooling on Earth. [1]

              The Counterargument

              Some industry voices (including groups like SpaceX) argue that space-based facilities could eventually become competitive or cheaper if terrestrial permitting delays and power bottlenecks become severe enough, or if launch costs drop below $200 per kilogram. You can review further breakdowns via the Boston Consulting Group Report and Tech-Insider Analysis.

            • Level 6

              LOL...I should charge you for Tutoring ;-)...

              The Physics of Heat Dissipation/Transfer in Space doesn't change, whether Elon Musk or Google is trying to build a Datacenter there...once again, watch the video and read the APS Summary on this topic. Bottom Line:...It's a Waste of Money and Resources to build AI Datacenters in Space, as opposed to doing so on Earth. Only someone who makes and launches Rockets would even Dream of putting them in Space...(Get the Hint???) ;-)...Another way of thinking about it..."Could you do it?...yes...Should you do it?...NOOOOOO! ;-)

              Btw, here's a portion of the Wikipedia page for Datacenters in Space...same story...

              https://en.wikipedia.org/wiki/Space-based_data_center

              _______________

              Disadvantages

              The deployment of space-based data centers raises several technical, economic, and environmental concerns.

              ·       Existing launch costs are substantial and remain the main cost of space infrastructure deployment

              ·       Cooling is limited to heat dissipation through radiation only. This is inefficient compared to convection used in terrestrial data centers

              ·       Space infrastructure must be designed to survive launch and to work under environment conditions of radiation, wide range of temperatures, in vacuum and in microgravity

              ·       In-space assembly is in an early development stage to enable deployment of mega-structures

              ·       Megastructures are particularly exposed to orbital debris

              ·       Solar arrays efficiency decrease 0.5% to 0.8% per year due to exposure of ultraviolet rays, space weathering and orbital thermal cycles[20]

              ·       Hardware is designed for limited lifespan[21]

              ·       Maintenance and repair in space, known as on-orbit servicing (OOS), is still in an early stage of practical implementation[22][23]

              ·       Disposable data centre: technology obsolescence of AI data centre being a concern and difficult maintenance in space imply the single-use purpose of those space data centres[citation needed]

              ·       To extend lifetime, space infrastructure will require either refueling or orbit raise by the servicer, which would increase its operational costs

              ·       Sun-synchronous orbit is a limited resource and proper management and sharing of it is required[24]

              ·       The environmental impact on Earth has its own challenges

              ·       The environmental impact of launches need to be addressed[25][26]

              ·       Deployment consumes Earth resources that cannot be recovered or recycled; computers require many resources, some of which are strategic and recycling e-waste is already a challenge on Earth and extremely unlikely in space[27]

              ·       Space debris (orbit pollution) is another sustainability challenge for space

              ·       Orbits are, like any resource, a limited physical and electromagnetic resource and available for all mankind; the accumulation of satellites on a particular orbit reduces the use of space for other purposes[28]

              ·       A consequence of the increase of satellites in orbit is a higher risk of the runaway of space debris (see Kessler syndrome) which means some orbits could become unusable[29][30][31]

              ·       Latency and bandwidth are constrained in space, and consume limited electromagnetic resources

              ·       Satellite flares could inhibit ground-based and space-based observational astronomy[32]

  • Level 1

    Guys...this is serious...Elon really wants to do this $2T Scam using YOUR MONEY...pay attention...

    DJT is Dumb enough to give him the Go Ahead...for the Right Price (to Him), and you know this is true.

  • Level 1

    LOL...Baron...the video runs for 40 min...you can't possibly have watched it in order to appreciate

    how bogus Elon Musk's plan is...;-)...but at least you are Consistent...always shooting off you opinions without stopping to think...

    I know you're just trying to needle and get a rise...but that's not happening...you're just acting the Fool as you dissipate whatever credibility you think you have.

End of conversation