That’s another reason I’m increasingly interested in the picks-and-shovels behind commercial drones — autonomy chips, batteries, sensing and edge AI.
No positions but watching the following.
$AMBA $AMPX
Amazon reportedly wants to go from around 85 same-day hubs today to 1,000+ by 2031.
Most people will look at this as Amazon vs Walmart.
I think there’s another angle worth watching.
Amazon has already said Prime Air drones will launch from its Same-Day Delivery sites.
And Prime Air is targeting nearly 500 U.S. cities/towns by the end of this year.
So these aren’t necessarily just warehouses.
Amazon may be quietly building the physical network needed to eventually scale drone delivery nationwide.
Novo just partnered with Anthropic to use Claude Science in actual drug discovery.
The obvious takeaway is “AI + biotech.”
I think the more interesting one is this:
AI models themselves may become commodities. Proprietary biology becomes the moat.
Big Pharma doesn’t need to build the best AI model. They can rent Claude and connect it to their own data, scientists and labs.
That makes me more interested in companies owning proprietary datasets, wet labs and validated drug-development infrastructure.
Compute could increasingly shift toward inference actually deploying AI at scale.
That keeps me interested in the less obvious AI beneficiaries: cooling, power, transformers, switchgear and rack infrastructure.
$MOD $TSSI are a few I am watching
Anthropic is calling for slowing frontier AI development… while signing a deal for a 2.16 GW AI data center.
I think there’s an important distinction being missed here:
Slower frontier training doesn’t necessarily mean slower AI infrastructure spending.