Charles Bice

Solutions Architect

[Charles Bice @charles_bice]

I actually looked at building something similar, found your platform, couldn’t believe how cheap it was, and locked in a subscription.

Let me know if you’re interested in some solutioning for getting those costs under control.. that bill can come down a ways.

Bigger concern is that your negative margins will shrink further as it scales, based on what I can gather about how it is constructed. Managed services can be abusively priced.

[Charles Bice @charles_bice]

@porterstansb lol you think the fed will let BoA fail? We’ve already seen this story once before

[Charles Bice @charles_bice]

Agree, this plus SKHY planning to push 50% of FCF into buybacks/dividends when it’s clear capacity needs to expand is insanely bad capital management.

In addition, they’re squeezing customers with pricing power, which is going to push those customers to look elsewhere and get creative about getting their costs down.

Long term I think this is a critical strategic blunder

[Charles Bice @charles_bice]

@RichardMoglen @Deepvue Man, this looks like a speculative ending to the bull market..

[Charles Bice @charles_bice]

Not sure about the risk on backdrop, this feels like late stage bull. You noted the change of tone after Monday and I think so too.

Not much strength in the biggest growth names likely due to anticipation that we’re near the end of the capex/cyclical earnings move. Concurrently, good things don’t usually follow when commodity prices spike across the board.

I think the move today on long dated treasury easing news says the market is demanding more easing than what the administration is providing in order to keep the bull alive. Feels like a game of chicken between the market and policymakers.

Could be wrong but this feels like the wrong time to get risk on.. hope this isn’t a bull trap for BTC

[Charles Bice @charles_bice]

@zerohedge Because summer is over

[Charles Bice @charles_bice]

Cloud compute costs have not come down though, they have only increased.. margins remain high for hyperscalers who are doing the bulk of the investment

It’s already profitable for them

The issue is primarily the viability of growth and margin expansion (contraction?) for the big AI labs. Not on the demand for intelligence overall. So that’s a far more narrow issue

[Charles Bice @charles_bice]

@dexhorthy Can you imagine describing your preference for second row 6 to waitstaff at a diner

[Charles Bice @charles_bice]

@factor_members Why not close half the position to cover some open profit then let the rest ride?

Profit + stop should cover the risk completely with optionality for upside

[Charles Bice @charles_bice]

@RealSimpleAriel There’s a narrative spreading that cheap Chinese memory is going to put a ceiling on prices and drive up supply

[Charles Bice @charles_bice]

@factor_members

[Charles Bice @charles_bice]

I think the frame is wrong on this topic

When did we start calling external vertical integration circular financing?

Nvidia is investing in their downstream business (inference) and OpenAI is raising capital to finance its own expansion of said business

It would be the equivalent of Standard Oil investing in gas stations and refineries

Back then Standard did it the obvious way, crushed their competition in said downstream businesses, and paid the price for it

XAI is the only company trying to close the loop on (most of) the value chain right now, others are avoiding it and instead keeping the investment indirect through these “circular deals”

More interesting is how antitrust law will adapt to these bottlenecks in the value chain.. that is what we don’t talk about nearly enough

[Charles Bice @charles_bice]

@factor_members My reframing of the question: trying to accurately model the underlying distribution is the most important, and difficult, part.. how do we (accurately) approximate the distribution specific to the markets we trade and keep it (them) accurate over time?

[Charles Bice @charles_bice]

@zerohedge It’s interesting that neither meta nor spacex openly sell cloud compute, yet we are to call them hyperscalers

[Charles Bice @charles_bice]

@levelsio Yeah looking at this I don’t think you were originally far off with your sweet spot.. it looks like you go from good to great starting around 300

[Charles Bice @charles_bice]

@zerohedge lol the only hyperscalers in that data are GOOG, MSFT and AMZN.. their spreads look fine

[Charles Bice @charles_bice]

@levelsio This is not showing the relationship you think it is.. plot a histogram instead of a bar chart to better understand the distribution

Just split the buckets (price bands) into even increments (I would run 2 plots, one at 50 and one at 100) and then plot it

[Charles Bice @charles_bice]

@richardartoul Well the value in the capex comes from the hyperscalers expanding infra, and the appeal of having someone else host your inference is not going to go away.. it’s almost certainly going to continue to grow

just another chapter in the story of cloud compute

[Charles Bice @charles_bice]

@factor_members An additional point to consider for copper fundamentals: it’s not used nearly as much as fiber in modern data centers. It’s also rare to see it in power generation and transmission due to costs. Copper is more of a residential/commercial wiring play vs an industrial use play.