I have questions.

Disclosure: The office owns SpaceX and is planning to sell on Friday. Fingers crossed I have estimated the total number of lemmings currently with brokerage accounts when forecasting the liquidity that will be available.

Last week, SpaceX announced that Alphabet's ($GOOGL) Google will pay $920 million per month to rent 110,000 Nvidia ($NVDA) GPUs, as well as CPUs, memory and other components from October 2026 through June 2029, with computing capacity beginning to ramp in September 2026. This agreement can be canceled for any reason as long as either party give 90 days notice or can be canceled by Google if SpaceX fails to “deliver access to the committed amount of GPUs by September 30, 2026.”

I thought there was a backlog of GPUs and memory and other other component required to build a datacenter? Maybe I missed the memo about this supply chain backup having been cleared. I do get a lot of emails so I probably did miss it. If there was a two quarter backlog, carry the one add banana, that would mean SpaceX can’t get all of that by September.

SpaceX’s IPO is Friday, June 12th. Ninety days from June 12th is September 10th. The terms of this deal really only commit Google to potentially having to pay two weeks of rental costs “at a reduced rate” for the month of September. And SpaceX’s first earnings call will be in September.

But, that isn’t what I have questions about. I fully understand how the IPO escalator to nowhere goes. What I have questions about is this slide:

The phrase “deployable liquid radiators” is not something you normally read. Let alone read three times in a graphic. Twice in the same beaded list.

Is it one radiator… or multiple? Who knows? not whoever threw this together.

First question, why would this need to be deployable? Wouldn’t that just require an unnecessary mechanical system that would be used once and then just sit there? Is there really ever a situation where you would un-deploy the radiators? Are you really designing a system with completely unnecessary single-use components?

Second, isn’t the whole point of putting a data center in space that you don’t need to use a liquid? Space data centers were supposed to use the vacuum of space as a radiator heat sink. Right? RIGHT? I can fully explain why the idea of using the vacuum of space is impractical and stupid but I don’t have to because it says, in triplicate, liquid.

Wate is the best liquid for this sort of thing or a water/glycol mixture. So I am sure they mean water. So the new plan is to not use the vacuum of space but to carry water up to space and then build a data center in space that is cooled with water and have all the same issues as you would on the ground. Only, in space.

Third, where are the pumps? If you are going to use a liquid you will need pumps. This seems like a pretty important detail and another mechanical part. And mechanical parts, you know, break.

Elon Musk’s companies are the public sector equivalent of a Nigerian Prince scam. There are enough mistakes and obvious errors/inconsistencies/illogicities (I made that word up just for fun) that a rational, educated, intelligent person isn’t going to fall for it.

Which is the intent I think.

SpaceX wants to find the people who will buy the IPO and hold the stock forever so that everyone who bought before the IPO can get out. Lemmings. The video game. Not the cute little intelligent animal the director of the Wild Kingdom threw off a mountainside.

The video game. Where you’d blow up a few little guys and order some other little guys to build stairs so people could escape. I used to love that game.

The target buyers of this IPO are the techno equivalent of being a gold/silver bug. People “stack” silver and gold and then die and their kids sell the gold if they can figure out the combination to the safe. There are plenty of people out there that never sold there marijuana dispensary shares or there solar-water tower shares.

We live in a reality where an likely AI-generated marketing slide can fully justify a $75 billion dollar IPO. Kind of nifty actually.

Let’s hope the stock hits $400 on Friday. GoPro hit $100 after its IPO. It hasn’t done so well since but it did hit $100 which was up 4x from the $24 listing price.

I am very curious what is the next new-new thing? So far humanity has not reached self driving, not made a hyper-loop, not generate power from solar roof tops, not replaced long-haul trucks with electric ones, not reached AGI, not made it to Mars, or even orbit yet, there is no fleet of robotaxis and there is no robot in my house folding my laundry.

SpaceX is partnering with a (wait for it) solar company out of California called Ambrosia Energy which I am just going to call Solar City 2.0 (space edition.)

So I guess the new-new idea is just to go back to the original new-new idea that failed already. Only this time the solar roof will be in space.

-AJ

MACRO: PPI is tomorrow!