A couple of big swings in AI this week, and they all kind of rhyme. Everybody is spending money like the bill never comes due. Here are three that caught my eye.
SpaceX Wants to Put AI Data Centers in Orbit
On June 8, SpaceX showed off the AI1 satellite, its first crack at an orbital data center. The thing is 20 meters tall with a 70 meter wingspan, basically racks of AI chips strapped to giant solar panels and liquid radiators. The company has already asked the FCC for permission to launch up to a million of them. This week CNBC ran a piece asking the obvious question: does any of this actually pencil out?
Why it matters: nobody wants a data center humming in their backyard, eating up power and water. Space sidesteps the neighbors and gives you all the sun you can soak up. If it works, the cost of compute shifts in a way that touches everything else.
My take: I love a big swing as much as the next guy, but launching a million of anything into orbit is a nightmare. You cannot send a tech out with a ladder when a rack goes down 300 miles up. The idea is right and the timeline is fantasy. Ask me again in ten years.
Qualcomm Is Shopping for an AI Chip Company
Qualcomm is in early talks to buy Tenstorrent for somewhere between 8 and 10 billion dollars. That deal would hand Qualcomm a real seat at the AI hardware table, the one Nvidia has been hogging for years.
Why it matters: right now if you want to train or run a big model, you mostly pay the Nvidia tax. Every serious buyer is hunting for a second option. More chip makers in the fight means better prices and shorter lines for the rest of us.
My take: this is the smart kind of spending. Qualcomm already knows how to ship silicon at scale, and Tenstorrent gives them a running start on the AI part. I would rather see money go here than into a satellite that needs a rocket to reboot.
The Funding Numbers Quit Making Sense
Q1 AI funding came in around 255 billion dollars, which beats the entire year of 2025 all by itself. Anthropic closed a 65 billion dollar round that put it near a trillion on paper, and SpaceX is talking about going public at 1.75 trillion or more.
Why it matters: that much money chasing one idea sets the weather for everybody. It decides which startups live, what gets built, and how hard the fall is if the music ever stops.
My take: I am old enough to have watched a couple of these cycles. Some of this money is buying real things that work. A lot of it is buying a seat at the table because nobody wants to be the one who sat out. Both can be true at the same time. Keep your own books honest and do not bet the ranch on a number with that many zeros.
That is the week. Big dreams, big checks, and not a lot of brakes. Stay sharp out there.