Rittenhouse Research

AI infrastructure, data centers, neoclouds, software, fintech.

[Rittenhouse Research @RHouseResearch]

AI Data Centers (Powered Shell Providers): Q2 2026 Earnings Recap

https://t.co/wkX7AcsUNJ

$APLD $CIFR $CLSK $CORZ $GLXY $HUT $RIOT $WULF

[Rittenhouse Research @RHouseResearch]

@PythiaR Not sure I’d agree on thin skin but light on details is a feature, not a bug

[Rittenhouse Research @RHouseResearch]

@kieranwgoodwin Why would an increase in rates impact the value of the loan when it’s floating rate?

[Rittenhouse Research @RHouseResearch]

@ethanrkho Comps. Ilya raised at $32B for SSI, Mira Murat raised at $12B for Thinking Machines, etc. so can use those as benchmarks and adjust based on how good the founder is perceived to be.

[Rittenhouse Research @RHouseResearch]

@ChillBanking I can't believe I forgot how clear, stable, and consistent China's laws and regulations have been over time...

[Rittenhouse Research @RHouseResearch]

There is no way I heard Gary Marcus just say that companies will "..use the Chinese models because at least (they) know what the rules are"..

In what world have these people been living in? The rules are clear in China..? My brain hurts.

[Rittenhouse Research @RHouseResearch]

I mean all else equal if there weren’t constraints around power generation, grid transmission capacity, labor, wafer capacity, etc. etc., there’d inevitably be more debt-financed data centers and chips.. Probably better for everyone involved over the longer-run if the build-out is financed with operating cash flows?

[Rittenhouse Research @RHouseResearch]

@jukan05 Gartner probably knows a bit about customer churn tbf

[Rittenhouse Research @RHouseResearch]

Has Satya gotten anything right in AI other than the OpenAI investment? And even there, could argue the one-foot-in, one-foot-out nature of Microsoft’s relationship with OpenAI is why they find themselves in such a poor position relative to where they started in 2022/2023..

[Rittenhouse Research @RHouseResearch]

@dalibali2 @hsri07 Time to capex also delayed. And I thought capex was entire reason for the -50% move in the first place, lol.

[Rittenhouse Research @RHouseResearch]

@junkbondinvest Kinda insane that the lenders would even fund this project prior to plan to power the data center being finalized..

[Rittenhouse Research @RHouseResearch]

The FT in a rare piece of good reporting pointing out that 73% of Nscale's 2025 revenue was from.. ByteDance. Pretty egregious that this wasn't disclosed in the S-1.

[Rittenhouse Research @RHouseResearch]

Just about through this Nscale $NSCL S-1 and man.. the only bull-cases I can come up with are that Jensen seems to like them a lot and they have a lot of big names on their board..

[Rittenhouse Research @RHouseResearch]

Amusing how every new neocloud is a self-described "vertically-integrated" operator.. Acquiring land, configuring BTM generation, securing grid connections, and building the powered shells supposedly results in a real "competitive advantage".. "Speed" and "lower costs" are often cited as the reason to self-build rather than lease data center capacity.

Yet only Oracle and CoreWeave have achieved 1GW+ of operating capacity and $10B+ of "ARR".. And both lease virtually all of their data center capacity. It's almost as if making the capital-and-operationally intensive cloud business even more capital-and-operationally intensive by self-building data centers might not be all that optimal..

$ORCL $CRWV

[Rittenhouse Research @RHouseResearch]

@JerryCap Ah yes, another middle man taking a cut, exactly what we need!

[Rittenhouse Research @RHouseResearch]

@HighyieldHarry 3 and 4 not much better either. They just absolutely butchered it.

[Rittenhouse Research @RHouseResearch]

@AltayCapital https://t.co/Zmserjxhtq

[Rittenhouse Research @RHouseResearch]

If the only thing we get from trillions in AI capex is an Expedia $EXPE bankruptcy.. so be it.

[Rittenhouse Research @RHouseResearch]

@OnodaCapital “TPU supply chain”.. AI bottlenecks”… “Muse ecosystem”.. etc. etc.

[Rittenhouse Research @RHouseResearch]

"CoreWeave remains the industry's sharpest, sturdiest, most proactive neocloud.. CoreWeave clusters are exemplary in all crucial categories." $CRWV

[SemiAnalysis @SemiAnalysis_]

ClusterMAX 3.0: The Industry Standard GPU Cloud Rating System Returns
In gory detail: reliability, performance, support, pricing
—and, of course, security—
in our most thorough analysis of GPU cloud providers globally
https://t.co/PYQ5uFpvor

[Rittenhouse Research @RHouseResearch]

@JordanNanos