Space DC: Napkin Math Debunked
Preface
To preface, I'm a Elon fan, had my earliest big stock win via all-in $TSLA and used to work at Tesla's Palo Alto office. I do think Space DCs will be a part of the solution to alleviate the power bottleneck but I'll show that they are not cheaper than terrestrial DCs.
To demonstrate, I'll only need to add one reason to @WR4NYGov's (Warren Redlich) napkin math to debunk and I won't need any. Warren is one of $TSLA investors community's early core members much like @FransBakker9812 is for $IREN.
I don't doubt Elon's engineering capabilities, so I'm going to take all of Warren's assumptions on radiant cooling and radiactive shielding working, reusable Falcon heavy to drive down launch costs, etc. Provided that all the engineering and cost optimizations works, Warren comes out with $10m to power on a Vera Rubin rack in space vs $12m per rack on the ground.
Key Point
The key point Warren misses is that terrestrial DC can be retrofitted to fit the new generation at 3-4m / rack while Space DCs need a relaunch, however reusable the rocket is. Using Warren's numbers, we will show the 20 year cost of Space DCs and terrestrial DCs under both cases of 5 year GPU lifetime and 7 year GPU lifetime.
For 5 year GPU lifetime, Space DC will have 4 launches over 20 years = 40m. For terrestrial DCs, it will be 12m initial build cost and (3.5m * 3 retrofits) = 22.5m.
Now, longer GPU lifetimes help the case for Space DCs. For 7 year GPU lifetime, Space DC will have 4 launches over 21 years = 30m. For terrestrial DCs, it will be 12m initial build cost and (3.5m * 2) retrofits) = 19m.
Conclusion
You see, I am giving Space DC maximal benefit of the doubt that all the engineering works and working off Warren's assumption that the initial cost including solar, radiant cooling of Space DCs is cheaper than terrestrial DCs to begin with.
Sure, Warren has stated that he hasn't counted the savings in electricity cost yet, but for companies like $IREN with cheap re-usable power in West Texas, South Australia and relatively smaller sites in Spain and British Columbia, electricity costs aren't the difference maker of 17.5m per rack (5 year GPU) or 11m per rack (7 year GPU).
So yes Space DCs are cost competitive with novel terrestrial solutions like $BE Fuel Cells but not with grid connected power. And none of this argument needs to touch on cost of redundancy needed in space since you can't economically make repairs in space, cost of insuring expensive Vera Rubins in space, and not to mention maintenance. Maintenance is actually a big one, Starlink needs little maintenance since it runs the same loop of software while GPU Cloud runs a diverse workload that many need all types of manual intervention. My wife works at Oracle Cloud Infra and it's not all to uncommon to call the Datacenter technicians.
You literally just signed SB 73 before the LA mayoral election. This law prohibits unauthorized access, disruption, modification, or seizure of voter rolls, voter lists, or certified voting technology by law enforcement (including federal agents) without a court order or specific state election law investigation.
You are actively making laws to make it HARDER for there to be election integrity.
Time to take down the California Communist Party
People demanding "proof" of election fraud are not understanding how crime works. I worked at Manhattan DA for over 2 years, one in Homicide. We never had video proof of the crime. We almost never had DNA. These are things that occur on CSI on TV, not in real life. And we still convicted people all the time.
What we had was testimony and circumstantial evidence. Travel times, bank records, cell phone data, gate access codes. Motive, capability, benefit, time and place. Never direct proof. Of course the defendant always denied the crime, but there was enough evidence to show that one had to have occurred nonetheless.
If what we have in the LA Mayoral election is a statistical anomaly that is beyond reasonable explanation with anything besides fraud, that is enough to prove a crime. This has been true since the beginning of Western Civlization.
If you’re an $IREN investor, I highly recommend following Fran and subscribing to his group. His research, analysis, and due diligence are among the most comprehensive I’ve seen. He put in real capital to support his thesis and Do deep dive research.
The value of the insights and community discussions alone makes it well worth it for anyone serious about understanding the $IREN investment thesis.
Highly recommend to follow Fran’s if you $IREN Bull. STAY Ahead of the curve.