lmao i am like struggling to write up this bspdn ting and I'm like why is it taking so long. its like 7000 words, that's why. ugh

sorry* 100million +

god the cost to entry is so freaking high. That's like a 10s of millions of dollars cluster lmao

@dylan522p i mean whatever besi im okay w/ - but like please please please love of god fucking go public but yes dylan i completely get it now

god im like begging for ev group to be public

Wow so they are talking about wafer bonding - and hopefully a new way to scale the business so everyone everywhere is starting to bond wafers together saying that there's a way to have more advanced scaling - yeah let's see!

VERTICAL STACKING SUCKS slide #4 to explain this this is pretty cool presentation

In hindsight - kind of obvious, we linearly scaled cost and that screwed over the economics of the industry This btw is pretty much a comment on semicap unsustainability for memory - wooooooo spicy. Sadly going forward we still don't have an answer other than layers

Just adding layers is one of the worst things you can do more bit growth, at worst cost. LMAO this is a great cost graph architecture, lateral, and other ways to scale have to be the future. "don't keep adding layers, you'll keep adding bits but wont reduce cost"

we scaled bits more than cost reduction - and that is the problem!

the industry is literally taking money, giving it to the capital equipment guys, and doing it again lmao

when we scale - you can't just make more bits, you have to reduce cost as well. If you sell more bits than the market will puke it out You cannot reduce cost without producing more bits.

Price drop is important, but if they don't have cost reduction they will lose money. Cost reductions are *demanded* from memory because of the price elasticity

This is a really cool graph i've never seen before. There is a relationship - drop the price - people will consume more. They have a good understanding of that

I mean its good to take a step back and appreciate this. Data will get bigger, this is an inevitability

despite all of that - data is one of the *most* secular parts of the entire world. You don't see 30% CAGR in anything, and yet here memory is.

"its hard for the industry to know anything, and you have to understand the economics to appreciate this business"

"we have not had a drop like this in the memory market since 1998, and even then it might be faster. the optimist in me says its going to come back even better"

U-curve of increased fabrication cost of semiconductors expressed differently

Whew

Freedom isn’t free. Why semiconductors gonna cost more /s

Okay I have to agree that price vs value difference is so big for semiconductors and I think we have finally moved to value based pricing