Sam | Fundamental Research

Researching where AI profits appear next: infrastructure, bottlenecks, software monetization and mispriced growth. Independent research. Positions disclosed.

[Sam | Fundamental Research @SamsEquityNotes]

Most AI investing still gets framed as one trade: buy the companies supplying the chips.

I think that misses where the economics can spread next.

My framework is three layers: infrastructure, models and applications.

[Sam | Fundamental Research @SamsEquityNotes]

@kevinxu $META is still in a good spot

[Sam | Fundamental Research @SamsEquityNotes]

@Mr_Derivatives caught META early and now my portfolio’s up 52.12% over the past month. feels good seeing it play out

[Sam | Fundamental Research @SamsEquityNotes]

once oil drops and yields come down, what are you going heavy on? $AMD and $MU for me 👀

[Sam | Fundamental Research @SamsEquityNotes]

$META doing the heavy lifting 😂 my smaller account is up 52% this past month. i’ll take that!

[Sam | Fundamental Research @SamsEquityNotes]

Korea’s chip rout isn’t dragging U.S. futures lower—that relative strength matters.
I’d watch whether $NVDA and $MU hold the open; if they do, forced selling may be fading. Small indirect $MU exposure.

[Sam | Fundamental Research @SamsEquityNotes]

@TradexWhisperer Cheap only matters if earnings hold. The real question is whether memory pricing stays tight at $MU and $SNDK.

[Sam | Fundamental Research @SamsEquityNotes]

@carmenli This is the signal I’d trust most. If GPU rental prices stay firm, AI demand hasn’t cracked at $NVDA.

[Sam | Fundamental Research @SamsEquityNotes]

@DanielTNiles The capex cut matters more than the miss. If supply tightens while AI demand holds, the top call could be early for $MU.

[Sam | Fundamental Research @SamsEquityNotes]

KOSPI’s selloff is a warning for U.S. chips—not the whole market.
I care more about whether AI demand cracks at $NVDA, $MU and $SNDK than overnight panic.

[Sam | Fundamental Research @SamsEquityNotes]

I use this account to work through public-market ideas in real time.

Mostly semiconductors, AI infrastructure and software—always with the valuation, the bear case and what would change my mind.

Independent research. No paid promotions. Positions disclosed when relevant.

[Sam | Fundamental Research @SamsEquityNotes]

This is not a call to abandon semis. Infrastructure demand can remain strong while value migrates upward.

The risk: AI becomes a feature, not a moat—and software pricing compresses before new usage revenue arrives.

Disclosure: No direct $NOW, $TEAM or $HUBS positions.

[Sam | Fundamental Research @SamsEquityNotes]

The best AI stocks next may not be the companies making the most chips.

They may be the software platforms that turn falling inference costs into higher-value work.

[Sam | Fundamental Research @SamsEquityNotes]

The evidence is starting to show: $NOW says AI passed $1B of ACV. $TEAM says Rovo customers are growing ARR about 2x faster than non-users. $HUBS is charging per AI-resolved conversation.

Different products. Same direction.

[Sam | Fundamental Research @SamsEquityNotes]

The risk: models may capture more of the economics, AI features may become table stakes, and customers may resist usage pricing.

Layer 3 is not automatically the winner. It is where the next round of proof should show up.

[Sam | Fundamental Research @SamsEquityNotes]

Most AI investing still gets framed as one trade: buy the companies supplying the chips.

I think that misses where the economics can spread next.

My framework is three layers: infrastructure, models and applications.

[Sam | Fundamental Research @SamsEquityNotes]

The companies I’m watching have three things: proprietary context, permission to act inside a workflow and a credible way to charge for outcomes.

That is a much higher bar than adding a chatbot.