Thread by @IDrawCharts

ok I promised a thread about my current investment/"macro" (blurgh 🤮) equity plays

so here it is

fair warning, it's a long one

The general thesis is "Kardashev 0.5"

Energy sources are practically unlimited, we just haven't figured out how to reach/harness most of them

we haven't needed to because we wouldn't have been able to use it

AI and robotics changes that, to the degree that we can physically manufacture chips and servos.

Eventually, the marginal costs of compute, energy and non-creative labor will all approach 0 (I don't believe this necessarily implies AGI/ASI but that's a different conversation)

whether money will even matter at that point is debatable, but it certainly does in the interim

Resource constraints on rare earths, metals, energy etc only really exist until we can send an army of self-propagating robots into space to gather all of these things, where they exist in practically unlimited quantities.

so how close are we to that future?

as far as R&D goes, we're honestly not too far off.

you want a task-specific robot? We can build it in a couple years of R&D

you want an AI that can reason about a specific problem? we're building those fast too

you want to pair those two together? We're doing it, see tesla self-driving, amazon's robotic warehouse system, multiple robotics platforms for performing surgery, etc

so it's an arms race for scale

reaching that scale starts with building an AI capable of receiving a task specification as input and ouputting a design for both the robotics and the controller system

this requires compute and energy

but implementation requires even more of both, plus servos.

To get compute, you need chips and energy.

To get chips, you need raw materials and energy.

To get servos, you need raw materials and energy

To get raw materials, you need energy.

The further down the road we go, the less value there will be in the "completed" product; whether that is a chip, a servo, a robot, whatever.

competitive edges in design will narrow and general-purpose systems will become less efficient than task-specific implementations

All this to say, I'm no longer interested in chip manufacturers.

Compute will be incredibly valuable, but cloud compute will never be as efficient as on-board

robotic design will gravitate toward application-specific hardware for both onboard compute and mechanical components

I'm investing in the areas that are not application-specific.

My first and largest bet is very boring: the S&P500, $SPY

Why? Because any non-creative work will make efficiency gains from AI/robotics.

I don't know which companies will benefit maximally, so I want them all.

That one's a no-brainer, so let's get more specific

I want exposure to nuclear energy generation, particularly companies who already own/operate reactors.

Nuclear is expensive to build, cheap to operate. It's consistent power, which is perfect for compute and robotics.

My second-largest position is Duke Energy $DUK, which operates 11 reactors over 6 plants, providing 10.7GW of capacity (not including other sources).

Current P/E ratio of ~18, with an operating margin of 27% over the last 12 months

I expect energy companies to become (and perform) like growth companies over the next decade, and investors to bet on future production over current output. I also expect to see new equity issuances in order to finance massive infrastructure and powerplant deployment costs.

$DUK sits well below the S&P average P/E of 28 with healthy operating margins and pays a dividend of 3.68%.

Cash on-hand is negligible and it carries relatively high debt, but with 13.8% Q3 YoY growth in net income I don't think this is a concern.

the next three are all pretty similarly-weighted and I'm getting tired of typing so I'll be quicker

Constellation, $CEG

The largest nuclear operator in the US, with 22GW of capacity, very little debt, lots of licensing agreements with data firms.

Next up, I want companies that produce the components needed to build compute and deliver power for it

Eaton, $ETN

Manufactures transformers and a bunch of other power grid components

Dependent on material supplies so changes to tariffs are a concern, but also will benefit from automation in manufacturing

Axcelis, $ACLS

Smaller company, outside the S&P500

Builds and services the equipment needed to manufacture semiconductor chips.

Somewhat contrarian bet, as it has high open short interest

This covers what I currently have, but I want to add additional manufacturing names, particularly those focused on parts used in robotics.

Finally, I want to pick up exposure to raw material production and logistics companies operating primarily inside the continental US.

I haven't had the time to do the research needed for these yet but will update when I do (if I remember)

anyways this is already super long so I'm glad I don't have any more to talk about

feel free to tell me why I'm dumb (or just short my names idc), this is what I'm up to

last note, I'm buying a narrative. Names are how I'm doing it, they are not the thesis themselves and will probably change.

boils down to "we gotta build and power a shit ton of robots before they can do it by themselves and money maybe becomes worthless anyways"