“No Spare GPUs Anywhere”: SpaceX’s Earnings Call Signals an AI Compute Crunch

Hello everyone! Today i want to share some trading ideas with you!

1

Elon basically said on the $SpaceX(SPCX)$ earnings call that it tried to rent outside AI compute to free up its own capacity for training but found no spare $NVIDIA(NVDA)$ GB300s, GB200s, older H100s, $Alphabet(GOOGL)$ TPUs or $Amazon.com(AMZN)$ Trainium.

That's pretty insane coming from one of world's largest AI infrastructure builders because the shortage extends across nearly every major chip architecture.

And if SpaceX can't find spare capacity despite its scale and willingness to pay then it's getting much harder to argue we're already in a broad AI compute glut.

2

What's more, $Bloom Energy Corp(BE)$ could be one of biggest beneficiaries of Elon Musk's comments on $SpaceX(SPCX)$ earnings call that power, cooling and electrical infrastructure are "somewhat of a limiting factor" for scaling AI compute with critical turbine parts facing 12–18 month delays.

Thats huge for Bloom since its fuel cells generate power directly at data center without turbine blades or years-long grid connections which could bring billions of dollars of idle GPUs online faster.

3

$SpaceX(SPCX)$ Starlink has tripled its subscriber base in less than two years and Elon Musk expects bandwidth demand to grow 1,000x over the next 5 years.

That helps explain why SpaceX is pushing to increase Starship launches and deploy more powerful satellites.

4

$SpaceX(SPCX)$ says new AI compute deployments are generating "a payback of less than a year" with pricing now above $50 per watt

That's a pretty strong signal for $NEBIUS(NBIS)$, $CoreWeave, Inc.(CRWV)$ and $IREN Ltd(IREN)$ as well because customers are clearly willing to pay a premium for AI compute delivered quickly.

5

Elon Musk says $SpaceX(SPCX)$ & $Tesla Motors(TSLA)$ Terafab project is targeting its first working test chip by May 2027 & volume production in Q3 2028 which is a timeline he calls "insanely fast."

Musk also thanked $ASML Holding NV(ASML)$ for their support as Terafab looks to bring more AI chip production in-house.