I had to read the article/paragraph a few times the other day to make sure I hadn't misread it. I couldn't believe. Data Centers. In. Space. His core reason? Unlimited solar energy. First, and foremost, a core requirement of datacenters is stability and reliability. This is achieved through redundancy. Many drives, many processors, many memory sticks, many wires. This all weighs a FUCK TON. And putting stuff in space..... costs a FUCK TON. I threw the following into ChatGPT: *Based on the average cost per pound to put objects into space (satellites), and the difference in power a single solar panel can generate in space per day (full time sun and higher intensity), along with the average electrical consumption of a data center (relative to how much power a panel can make up there and how many panels would be required), how long would it take to break even putting a data center in space versus keeping it on earth with solar panels* It basically said that a datacenter in space would have to run for more than a decade to JUST break even. That doesn't even account for the difficulties of getting a person up there to fix an issue if something breaks (Memory sticks /etc die all the time. I imagine that it will be more frequent in a harsh environment such as space). Also, a data center would be a large target for a small piece of space debris. With all the land/space available, rather than building a large solar array on earth (make it bigger and he could be a utility provider selling energy off to a surrounding area) , he wants to build a somewhat smaller one and lift all that into space and hope it doesn't break. I work in enterprise data systems. Shit breaks. A lot. His other reasons? Scalability (Space in space), Grid Relief, and Thermal Control: "Unlimited Space" . Yeah, no it's not unlimited. There's a lot, sure. At varying altitudes. But its VERY expensive to get there, stay there, and fix things that break there. Outages can be long in worst case scenarios. It would be easier, and cheaper, to put it on an old (or hell, brand new), oil rig platform in the ocean and utilize solar, wind, and ocean wave powered sources of energy for it. That would take power off residential/commercial grids. And you could literally pump sea water through radiators for better thermal control. It's been a while, but I feel like I recall that thermal control is kinda challenging in the vacuum of space. This is all a stupid fucking idea perpetrated by a person with the emotional and intellectual maturity of a toothpick.

---

What Musk announced is de facto slaughter of most AI competition - definitely OpenAI. Niche and in-demand ones like Antropic will survive. What he aims for is to get AI industry out of the conflict with economy - no competition for power, water or even land. The semiconductors competition remains unsolved, but just the limit on available power generation, and the price spike on grid power, will become competitive disadvantage for OpenAI any other large AI company. There are a lot of ifs stacked in this: BFR should fly reliably, supply chain should accomodate the million-satellite buildup, satellites should prove a useful endurance in space, etc. But if he does this, lots of things will get wrecked.

---

off-grid self-sufficiency in middle of bumfuck nowhere.

---

Yeah the data center in space is dumb, and it'll simply never happen. It would cost a hundred million dollars to get one server rack worth of processing power in space. But cool right? That does.... Absolutely nothing for you. Like you said, we have data centers on earth. There's no advantage to sending your task to one in orbit. None, zilch, nada. Meanwhile you could have that same rack running on earth for a startup cost of less than $25,000. Spacex will have a lot less government work after they deorbit the ISS. It also actually has some competition now with the Chinese rockets and Bezos. Their only real valuable offering is Starlink which is a genuinely compelling offering. However the answer to space where there's already a lot of competition. The idea that you could have cell phone service or internet anywhere on Earth is awesome and will get a fair amount of use. However most customers already have cell phones and internet from terrestrial sources that are faster and cheaper. Your average home user isn't going to switch to Starlink when they have a fiber connection. They aren't going to drop Verizon. This is more for people who live off the grid, countries with poor infrastructure, people on boats, etc. So they will always absolutely have customers there, but we're not talking hundreds of millions of subscribers. But wait? They are also giving you X and Grok. Obviously they both have some value, but Grok has a pitiful amount of traffic compared to either openai or gemini. To put this in perspective, it would be like when Google completely took over and was king of the search in 2010 and someone said hey want to buy Lycos? Now X has real value. However it lost a lot since Musk took over. X is all advertising, and most brands don't want the optics of being associated with a controversial brand. Because of this, they make a lot less money on advertising than they used to. TLDR. $1,500,000,000,000 has two too many zeros. If it went IPO for $15 Billion, it would sound right. Still inflated from a business point of view, but hey, it's Musk. The guy is famous. $1.5 trillion is just fiction. It'll crash hard!

---

Yep. Was off-grid in Ontario for a week and was able to get cell signal by pointing my phone at passing satellites, pretty cool tech to see in person. Worked pretty well too, on iPhone 16 Pro.

---

> Solar power is free on the ground You need to build 6x the panels + a shitton of battery storage on the ground for the same output as in space. Wikipedia says: > ... state of the art for flown spacecraft, which as of 2015 was 150 W/kg (6.7 kg/kW), and improving rapidly. Very lightweight designs could likely achieve 1 kg/kW So assuming 1 kg/kW (at least it's not a Musk estimate), 1 kW of reliable electricity costs: * on Earth: 6 kW of regular panels + enough battery storage, let's say 24 kWh (which probably isn't enough to account for weather but we got to start somewhere) + land to put it * 1 kW of space grade panels + 1 kg launch cost Assuming 1 kW of lightweight space grade panels, manufactured at obscene scale, cost the same as 6 kW of regular panels (I could imagine them getting *cheaper* than earth panels once the scale is there, simply because it's less material), the question would be whether battery storage is cheaper than launch. https://ember-energy.org/latest-insights/how-cheap-is-battery-storage/ estimates $30/kWh of battery storage which already includes some grid connection costs. That means that in our model, *if* launches become cheaper than $720/kg, space power may be cheaper than earth power. Musk dreams of Starship launches costing $100/kg, but we all know what that is worth. Of course, the power would need to be *sufficiently* cheaper to make it worth running the rest of the DC in space.

---

I don’t think it’s worth the hassle. Elon’s already working for Tesla and has a 10 year plan. Space x +xai then layer on Optimus, energy, autonomy, robotaxi then shareholders+ bureaucracy that comes with going public the shareholders BS for both Tesla and now space x and xai haha that’s so not Elon. Who in their right mind would want to deal with that. The answer is nobody absolutely nobody. If you have read elons bio or followed him and his position on progress going public will slow that progress down. Space X is not going public anytime soon but it’s fun to think about. I mean mid terms he saw DC with doge and space isnt the major focus right now. National defense, grid modernization, energy security is.

---

Not quite. You need either a grid connection or storage. Grid connections have limited availability, storage is still fairly expensive.

---