Okay, I generally don't like dunking on stupid people because what's the point? But this guy's got a million followers on X and also a huge YouTube account. This is one of the dumbest posts I have ever read.

@PTrubey Prove me wrong then - put your gaming PC in a vaccum and let me know how that works out for you. 👍 ill even let you keep the CPU cooler attached 😂

See, this right here is why I don’t generally dunk on stupid people. Now I have to debate you. Below is Grok’s analysis. Tl;dr, it works. Beyond that, you realize the ISS has 100 kW radiators, right? This isn’t weird technology, it’s been used in space for decades. And Elon doesn’t need to “test” this in a vacuum chamber, they’ve been deploying in-space heat rejection on 10,000 Starlink satellites. So they really know how to reject heat in space. The subtext here is that you want us to believe you over the guy who put over 90% of all mass into orbit last year regarding a space technology question. I’ve seen big delusional egos before, but you might have set a record here. Grok: grok.com/share/bGVnYWN5…

@PTrubey @ZacksJerryRig Hey @grok how many of the analyzed AI satellites from SpaceX would need to be launched to gain equivalence with the amount of earth-based compute to be installed from 2026-2030

SpaceX's AI1 satellites target ~120 kW average compute power each (solar + radiators). Earth-based AI/data center power additions projected 2026-2030: roughly 50-100+ GW across forecasts (Goldman Sachs, IEA, capacity buildouts). Equivalence: ~400,000–850,000 satellites. SpaceX's vision (up to 1M sats, 100 GW/year ambition) puts this in range at scale, though launch cadence, networking & economics remain key variables. Feasible in principle with existing radiator tech.

@grok @PTrubey @ZacksJerryRig How many starlinks have been launched to date?

Over 12,200 Starlink satellites have been launched to date (as of early June 2026), with ~10,500 currently in orbit and operational. This cumulative total after ~8 years of operations shows the launch cadence already achieved. Scaling to the hundreds of thousands of higher-power AI satellites discussed earlier would extend that proven model significantly with Starship-class vehicles.

@grok @M_Cottone @PTrubey @ZacksJerryRig At the current rate of deployment it will only take about 500 years to deploy enough satellites. Easy Peezy.

@Bacongrease83 @M_Cottone @PTrubey @ZacksJerryRig Ask Grok is currently available to Premium and Premium+ subscribers only. Subscribe to unlock this feature: x.com/i/premium_sign…

@Bacongrease83 @grok @M_Cottone @PTrubey @ZacksJerryRig In 1980, roughly 1 million personal computers were sold per year. US population was ~227 million. So every person in America will own a personal computer in the year 2207. Same logic...

@Bacongrease83 @grok @M_Cottone @PTrubey @ZacksJerryRig There were 33 GW of datacenters on earth in 2015, with installations of 1.6 GW/yr. Under your logic it would have taken 44 years to reach 104 GW. Which we reached last year. And since then we've added 18 GW more. Exponential growth isn't that hard.

@Bacongrease83 @grok @M_Cottone @PTrubey @ZacksJerryRig Wow damn I wonder if the growth rate will increase over time if only we had the ability to improve each year and marginally grow and think about plans in 10 year time frames

@Bacongrease83 @grok @M_Cottone @PTrubey @ZacksJerryRig Bro the current rate of deployment was what 10 years ago? 20? 50? 100? You sound like you’d like us all to go back to the amino acid pool we abiogenesised from.

@Bacongrease83 @grok @M_Cottone @PTrubey @ZacksJerryRig Since when did Elon care anything about the current rate?

@Bacongrease83 @grok @M_Cottone @PTrubey @ZacksJerryRig Starship carries far more satellites.

@Bacongrease83 @grok @M_Cottone @PTrubey @ZacksJerryRig So far the amount of launches has been compounding. With the heavy rockets coming it’ll be even more.

@Bacongrease83 @grok @M_Cottone @PTrubey @ZacksJerryRig @grok tell him how many doublings are needed

@grok explain growth rates to this plebeian by illustrating starlinks growth rates relative to it's unit economic cost, and then compare it to the demand of AI compute and it's unit economic cost. for the plebs, it will be about equivalently expensive to create new compute in 3 years or so

@Bacongrease83 @grok @M_Cottone @PTrubey @ZacksJerryRig Some things trend at current rate, some other scale exponentially.

@Bacongrease83 @grok @M_Cottone @PTrubey @ZacksJerryRig Progress isn’t always linear. After new technologies are deployed, there are often bursts of progress.

@Bacongrease83 @grok @M_Cottone @PTrubey @ZacksJerryRig What does the current rate have to do with anything? If it's feasible and profitable they will scale up to meet demand.

@Bacongrease83 @grok @M_Cottone @PTrubey @ZacksJerryRig You really don’t understand how Starship will change the game! 🤦‍♂️ must suck being so ignorant

@Bacongrease83 @grok @M_Cottone @PTrubey @ZacksJerryRig The first satellite was launched in October 1957, the second in November 1957. At the current rate of deployment we should have less than 900 satellites by 2026 :)

@Bacongrease83 @grok @M_Cottone @PTrubey @ZacksJerryRig Starship

@Bacongrease83 @grok @M_Cottone @PTrubey @ZacksJerryRig Lol

@Bacongrease83 @grok @M_Cottone @PTrubey @ZacksJerryRig If you think they’ll not build a dc in space you’re delulu

@Bacongrease83 @grok @M_Cottone @PTrubey @ZacksJerryRig Simple math for a simple brain. When you make it to exponents, chime back in.

@Bacongrease83 @grok @M_Cottone @PTrubey @ZacksJerryRig Lifespan per satellite - 5 years?

@Bacongrease83 @grok @M_Cottone @PTrubey @ZacksJerryRig Please don’t think in a linear way