Earnings Season Capital

Buyside analyst. It's always earnings season.

[Earnings Season Capital @CapitalCold]

A great long idea setup is when next 12-24 month numbers are obviously going higher but when/by how much/what multiple to pay for that is up for debate. Typically by the time the numbers role in the easy money has been made. Tactically, set-ups like this tend to provide something of a “shield” heading into earnings prints where the market will look through most near-term choppiness given hopes for better things going forward.

[Earnings Season Capital @CapitalCold]

@viggy_krishnan @stevehou Meta last quarter...

[Earnings Season Capital @CapitalCold]

Data center delays are a real risk, but one within the prevailing paradigm (demand well in excess of supply)

[Earnings Season Capital @CapitalCold]

How is everyone just now finding out about optics

[Earnings Season Capital @CapitalCold]

This screenspot-pro score seems crazy impressive… does this mean computer use is here?

[Earnings Season Capital @CapitalCold]

@ContrarianCurse The ads have more utility for me than 90% of posts from people I follow at this point. Shopping idea gen portal that gets me

[Earnings Season Capital @CapitalCold]

More here: https://t.co/sdTFjbMwk3

[Earnings Season Capital @CapitalCold]

Some thoughts post mega-cap tech earnings... I think $AMZN is the most actionable idea of the group.

[Earnings Season Capital @CapitalCold]

@BlameQuants $smci is on its way there

[Earnings Season Capital @CapitalCold]

It seemed clear to me a little under a year ago that reasoning models represented a paradigm changing breakthrough that was severely underappreciated by investors. I saw firsthand the potential for compute usage to go up by orders of magnitude, and appreciated that inference demand was far more likely sustainable than training demand (yes I got wrecked by DeepSeek/Tariffs). I am not sure I see a similar disconnect today as the corresponding inflection in compute usage (and revenue to an extent) on the back of reasoning models has been recognized and has led to a new round of much bigger and more ambitious capex/infrastructure build plans and financing deals. The upside case would be we see another similar breakthrough in the next 12 months or so/the ROI continues to be attractive on much bigger input numbers (truly epic bubble scenario). I honestly don’t know if I would rather bet for or against that at this point. But the key point is it feels like we have gone from tangible disconnect (so many did not update their priors post reasoning models) to having to take things a bit more on faith.

[Earnings Season Capital @CapitalCold]

Thoughts on a few recent prints here: https://t.co/Pyozr60nre

[Earnings Season Capital @CapitalCold]

A great long idea setup is when next 12-24 month numbers are obviously going higher but when/by how much/what multiple to pay for that is up for debate. Typically by the time the numbers role in the easy money has been made. Tactically, set-ups like this tend to provide something of a “shield” heading into earnings prints where the market will look through most near-term choppiness given hopes for better things going forward.

[Earnings Season Capital @CapitalCold]

That setup describes what we have seem in the semicap space recently with surging memory prices and strongly positive leading indicators for leading edge logic demand. I have been a proponent of buying low/selling high in this space (but maintaining long-term exposure given strong secular trends and competitive positions/returns on capital) so trimming here is probably prudent… but I also hold the view that you have to make hay while the sun is shining/maintain material exposure during “harvest” periods that can often overshoot to the upside.

[Earnings Season Capital @CapitalCold]

@ContrarianCurse Or if your biggest customer is falling apart/hires your major competitor’s former CEO that’s probably a bad sign 😂

[Earnings Season Capital @CapitalCold]

@ContrarianCurse Do you agree that a key takeaway is just wait until a company actually reports and guides post closure of a transformational deal?

There was a deal overhang discount on the stock that arguably went away post deal closure but prior to the actual report

[Earnings Season Capital @CapitalCold]

$AMD / $INTC should presumably be key beneficiaries of this trend (the latter so long as key workloads are not "accelerated"). Various infrastructure software vendors (e.g. $DDOG, $MDB, $ESTC, etc.) also likely stand to benefit over time.

[Earnings Season Capital @CapitalCold]

$MU's callout of AI agents driving surging traditional CPU-based server usage (driving an upped market growth view to MSD % from flat) still seems relatively underdiscussed / perhaps underappreciated (at least outside of the direct memory/storage space). The fact that they are calling this out as a material source of upside given how early we are in agentic AI deployments (proof of concept stage really) is very interesting given those workloads have the potential to increase many many orders of magnitude from here.

[Earnings Season Capital @CapitalCold]

I haven't heard any other companies call out this trend (anyone?) and maybe $MU is wrong or being misleading as to the materiality of this specific driver... but the idea makes perfect intuitive sense: these agents ping various apps/APIs/databases etc. far more than humans do.

[Earnings Season Capital @CapitalCold]

@ContrarianCurse The consmer value prop is incredible and probably has a long way to run in terms of penetration, given that view is still non consensus IMO (although def some stated vs. revealed preference dynamic at play here)

[Earnings Season Capital @CapitalCold]

The $hut thesis is actually not terrible.

Very cheap on an implied per watt basis relative to peers with tight supply/demand environment leading to likely unlock catalyst in near future

If there’s a fatal flaw it’s the $ABTC back out math… what is the real fundamental value there if we are being honest?

To be clear I know next to nothing about the actual asset base and what is real vs. phantom power availability for HPC/AI here

[Earnings Season Capital @CapitalCold]

$CDNS / $SNPS downside beta here feels a bit extreme