Jun. 17 at 3:54 PM
My core AI-infrastructure portfolio — focused on the physical bottlenecks that actually determine how fast AI can scale.
$VRT — thermal management + onsite power systems for data centers; critical as rack density and heat loads accelerate
$BE — fuel cell systems tied to grid stabilization; plays directly into peak electricity demand constraints from AI workloads
$AAOI — optical transceivers enabling high-speed data movement across servers and distributed compute clusters; key photonics link in the stack
$IREN — integrated data center + energy operator; strong positioning around secured power capacity, the real bottleneck in large-scale AI buildouts
This isn’t just “AI exposure” - it’s a bet on the physical infrastructure layer (power, cooling, and photonics) that determines how far the entire AI cycle can extend.
Sharing framework and levels - no paywall, just structured positioning thinking
