something i'm watching closely over the next couple of weeks is $DRAM ETF
its biggest weightings are SK Hynix, Micron, Samsung so its more of a broad exposure to the memory trade rather than betting on a single company
i've spent the past few weeks researching memory and i don't think we're anywhere near the top. people keep calling memory cyclical and saying demand has peaked but i think they're missing what's actually driving this cycle.
there are only three companies that matter at scale: sk hynix, micron, and samsung and together they effectively control the industry. building a competitor isn't something you do overnight. it takes billions in capital, years of r&d, and an incredibly complex supply chain.
the reason i'm bullish is because every major hyperscaler is still spending aggressively, HBM demand remains extremely strong and memory is becoming a serious bottleneck given that there is no AI without memory just like GPUs
then there's what comes after ai: robotics. if robotics scales anywhere close to expectations then memory demand expands into yet another market entirely. robots will need massive amounts of fast memory to process vision, reasoning, planning, inference, etc.
major memory companies still trade at relatively modest forward multiples compared to other ai beneficiaries. if the market starts viewing memory as structural instead of purely cyclical then I believe there's room for a meaningful rerating.
could i be wrong? absolutely. the biggest thing that would make me reconsider is evidence that HBM demand is slowing materially, hyperscalers are cutting AI capex, or memory pricing starts rolling over
until i see that i think the broader memory thesis remains intact
its biggest weightings are SK Hynix, Micron, Samsung so its more of a broad exposure to the memory trade rather than betting on a single company
i've spent the past few weeks researching memory and i don't think we're anywhere near the top. people keep calling memory cyclical and saying demand has peaked but i think they're missing what's actually driving this cycle.
there are only three companies that matter at scale: sk hynix, micron, and samsung and together they effectively control the industry. building a competitor isn't something you do overnight. it takes billions in capital, years of r&d, and an incredibly complex supply chain.
the reason i'm bullish is because every major hyperscaler is still spending aggressively, HBM demand remains extremely strong and memory is becoming a serious bottleneck given that there is no AI without memory just like GPUs
then there's what comes after ai: robotics. if robotics scales anywhere close to expectations then memory demand expands into yet another market entirely. robots will need massive amounts of fast memory to process vision, reasoning, planning, inference, etc.
major memory companies still trade at relatively modest forward multiples compared to other ai beneficiaries. if the market starts viewing memory as structural instead of purely cyclical then I believe there's room for a meaningful rerating.
could i be wrong? absolutely. the biggest thing that would make me reconsider is evidence that HBM demand is slowing materially, hyperscalers are cutting AI capex, or memory pricing starts rolling over
until i see that i think the broader memory thesis remains intact
🔥4❤3
Should be an interesting day for equities.
Sentiment has flipped pretty aggressively overnight. First we had Apple lobbying to use Chinese memory, then Meta talking about selling excess compute capacity, and now the latest narrative is that memory efficiency somehow means less memory demand. To me this is just another version of the same fear we've seen throughout every major tech cycle.
I think this is a classic case of Jevons Paradox. If memory gets cheaper and more efficient we'll just end up using more of it
Let's see what today brings🤗
Sentiment has flipped pretty aggressively overnight. First we had Apple lobbying to use Chinese memory, then Meta talking about selling excess compute capacity, and now the latest narrative is that memory efficiency somehow means less memory demand. To me this is just another version of the same fear we've seen throughout every major tech cycle.
I think this is a classic case of Jevons Paradox. If memory gets cheaper and more efficient we'll just end up using more of it
Let's see what today brings
Please open Telegram to view this post
VIEW IN TELEGRAM
❤6
This media is not supported in your browser
VIEW IN TELEGRAM
brutal day for stocks all around
going to touch some grass and remember: volatility goes both ways. see you all tomorrow in the books 🫡
going to touch some grass and remember: volatility goes both ways. see you all tomorrow in the books 🫡
❤3
don’t really have much to say on the markets today. still holding through the blood, and i’d rather not force commentary just for the sake of it. instead, here’s something i’ve been thinking about lately.
i’ve always found it interesting how almost every ancient tradition speaks in cycles. hermeticism, stoicism, eastern philosophy, alchemy, etc. nothing grows forever and nothing dies forever. everything expands and contracts and eventually transforms into something else entirely.
one of the ideas i keep coming back to is “as above, so below. as within, so without.” whether you interpret it spiritually or simply as a philosophy I really think the principle is fascinating. it suggests that the same patterns repeat across different scales of reality. the individual reflects society. society reflects nature. the micro mirrors the macro.
the same emotions that drive a single person can drive an entire market. the same cycle of accumulation, excess, decline and renewal appears in civilizations, companies, relationships and even in people.
modernity has become obsessed with predicting the future when understanding the present is already difficult enough. that’s why traditions focused less on prediction and more on observation. if you can recognize the pattern you’re living through today you probably don’t need to spend so much time trying to predict or worrying about tomorrow.
there’s immense value in reading ideas that have survived for hundreds or even thousands of years because they simply teach you to ask better questions
daily reading here: https://esotericaauthentica.com/en/as-above-so-below-meaning/
i’ve always found it interesting how almost every ancient tradition speaks in cycles. hermeticism, stoicism, eastern philosophy, alchemy, etc. nothing grows forever and nothing dies forever. everything expands and contracts and eventually transforms into something else entirely.
one of the ideas i keep coming back to is “as above, so below. as within, so without.” whether you interpret it spiritually or simply as a philosophy I really think the principle is fascinating. it suggests that the same patterns repeat across different scales of reality. the individual reflects society. society reflects nature. the micro mirrors the macro.
the same emotions that drive a single person can drive an entire market. the same cycle of accumulation, excess, decline and renewal appears in civilizations, companies, relationships and even in people.
modernity has become obsessed with predicting the future when understanding the present is already difficult enough. that’s why traditions focused less on prediction and more on observation. if you can recognize the pattern you’re living through today you probably don’t need to spend so much time trying to predict or worrying about tomorrow.
there’s immense value in reading ideas that have survived for hundreds or even thousands of years because they simply teach you to ask better questions
daily reading here: https://esotericaauthentica.com/en/as-above-so-below-meaning/
Esoterica Authentica
As Above So Below: Meaning and Origin in Hermeticism
Discover As Above So Below meaning in Hermeticism, including its origin and connection to the Emerald Tablet and Hermes Trismegistus.
❤8
This media is not supported in your browser
VIEW IN TELEGRAM
huge day for Korean memory
samsung closed up 8% and sk hynix finished the session up 11% after yesterday’s panic lows
based on today’s closes the $DRAM NAV is now roughly $70 while the perp is still sitting around $64 that’s an implied discount of roughly 7% assuming the numbers hold it should peg next session/few sessions
meta clarification also reinforces what i’ve been saying all week. the AI race is still on and the industry isn’t where it wants to be yet which to me means more iteration, more infra and ultimately more AI spend
monday will be the first opportunity for the US market to digest everything
enjoy the long weekend 🫡
samsung closed up 8% and sk hynix finished the session up 11% after yesterday’s panic lows
based on today’s closes the $DRAM NAV is now roughly $70 while the perp is still sitting around $64 that’s an implied discount of roughly 7% assuming the numbers hold it should peg next session/few sessions
meta clarification also reinforces what i’ve been saying all week. the AI race is still on and the industry isn’t where it wants to be yet which to me means more iteration, more infra and ultimately more AI spend
monday will be the first opportunity for the US market to digest everything
enjoy the long weekend 🫡
❤5
pretty nice to see $ANSEM hit $200m. appetite for the trenches seems to be slowly coming back along with general crypto sentiment
haven’t played memes in a long time but it might be worth keeping an eye out in case coins with similarly strong narratives emerge
that said I’m still cautious. trenches evolved to the point where most of the old edges have been arbitraged away. there’s very little edge left
haven’t played memes in a long time but it might be worth keeping an eye out in case coins with similarly strong narratives emerge
that said I’m still cautious. trenches evolved to the point where most of the old edges have been arbitraged away. there’s very little edge left
❤3
new personal goal with @Thresearch
going to dive deep into learning options and really build a solid understanding from the ground up. it’s a skill I wanted to learn
it’s going to be a long journey with plenty to learn but i’m excited for it
going to dive deep into learning options and really build a solid understanding from the ground up. it’s a skill I wanted to learn
it’s going to be a long journey with plenty to learn but i’m excited for it
❤7
your mind is capable of far more than you can imagine
the kabbalah teaches that this world is not the beginning of creation but its finality. what is revealed is only the last expression of what was first concealed. before there is form there is thought, before there is action there is will, before something exists below, it must first exist above.
have you ever noticed how someone can ruin an entire day before it even begins?
they wake up convinced the meeting is going to go badly. convinced they’ll say the wrong thing. convinced people won’t like them. nothing has actually happened, yet in their mind their reality had already happened. eventually, the world obeys their will and so it becomes true
by the end of the day they smirk and say “I knew this would happen!” but did they really know or did they will it to reality? they had already accepted that future long before the day even began
if the soul can descend into fear before fear ever arrives then why couldn’t it also ascend toward something greater before it is visible
perhaps this is why God gave mankind imagination. man is created in the image of God. not because we resemble him physically but because we participate in creation. God creates worlds and man creates the world he inhabits
unknowingly we sometimes use this gift to build prisons. but what if imagination was never given to us for fantasy and what if it was given to us so we could perceive what has not yet descended into this world
the greatest tragedy is that we imagine too little of what God made us capable of becoming
the kabbalah teaches that this world is not the beginning of creation but its finality. what is revealed is only the last expression of what was first concealed. before there is form there is thought, before there is action there is will, before something exists below, it must first exist above.
have you ever noticed how someone can ruin an entire day before it even begins?
they wake up convinced the meeting is going to go badly. convinced they’ll say the wrong thing. convinced people won’t like them. nothing has actually happened, yet in their mind their reality had already happened. eventually, the world obeys their will and so it becomes true
by the end of the day they smirk and say “I knew this would happen!” but did they really know or did they will it to reality? they had already accepted that future long before the day even began
if the soul can descend into fear before fear ever arrives then why couldn’t it also ascend toward something greater before it is visible
perhaps this is why God gave mankind imagination. man is created in the image of God. not because we resemble him physically but because we participate in creation. God creates worlds and man creates the world he inhabits
unknowingly we sometimes use this gift to build prisons. but what if imagination was never given to us for fantasy and what if it was given to us so we could perceive what has not yet descended into this world
the greatest tragedy is that we imagine too little of what God made us capable of becoming
❤8🔥2👏1
weekend memory recap + what i’m watching this week.
despite last week’s panic, the memory thesis strengthened over the weekend.
- Samsung is reportedly pushing for up to a 20% QoQ increase in Q3 DRAM ASPs (and over 20% for LPDDR). suppliers generally don’t push hikes like this unless they believe demand is there
- Samsung foundry is reportedly in talks with Meta on a ~$6.5B AI chip contract. different business than memory but another reminder that hyperscalers continue investing aggressively in AI
- Alexander Wang’s clarified Zuck’s previous comment: the industry still has a long way to go on agentic capabilities. simply means the AI race continues and has more juice
- Bernstein also reiterated that AI capex remains intact and continues to see hyperscaler spending staying elevated despite recent market fears
key dates this week for memory:
- tonight: Korea opens first. Samsung and SK Hynix will be the first read on sentiment.
- tuesday july 7: Samsung preliminary Q2 earnings guidance. Probably the biggest catalyst of the week for memory.
- friday july 10: SK Hynix begins trading on Nasdaq, giving US investors direct access to the company
the AI race isn’t slowing down as pricing power appears intact, infra spending continues, and this week should give us some hard catalysts
despite last week’s panic, the memory thesis strengthened over the weekend.
- Samsung is reportedly pushing for up to a 20% QoQ increase in Q3 DRAM ASPs (and over 20% for LPDDR). suppliers generally don’t push hikes like this unless they believe demand is there
- Samsung foundry is reportedly in talks with Meta on a ~$6.5B AI chip contract. different business than memory but another reminder that hyperscalers continue investing aggressively in AI
- Alexander Wang’s clarified Zuck’s previous comment: the industry still has a long way to go on agentic capabilities. simply means the AI race continues and has more juice
- Bernstein also reiterated that AI capex remains intact and continues to see hyperscaler spending staying elevated despite recent market fears
key dates this week for memory:
- tonight: Korea opens first. Samsung and SK Hynix will be the first read on sentiment.
- tuesday july 7: Samsung preliminary Q2 earnings guidance. Probably the biggest catalyst of the week for memory.
- friday july 10: SK Hynix begins trading on Nasdaq, giving US investors direct access to the company
the AI race isn’t slowing down as pricing power appears intact, infra spending continues, and this week should give us some hard catalysts
👍5
blasting metal and waiting for US market open
michael burry mode
its ironic because he's short and I'm long
anyway UBS operating profit estimates for Samsung Electronics and SK hynix
1. Samsung Electronics
2026: KRW 441T ($287.9B)
2027: KRW 837T ($546.5B)
2028: KRW 902T ($588.9B)
2. SK hynix
2026: KRW 327T ($213.5B)
2027: KRW 623T ($406.8B)
2028: KRW 667T ($435.5B)
if those estimates are right
> samsung trading at 2.4x 2027 PE
> skh trading at 2.6x 2027 PE
michael burry mode
its ironic because he's short and I'm long
anyway UBS operating profit estimates for Samsung Electronics and SK hynix
1. Samsung Electronics
2026: KRW 441T ($287.9B)
2027: KRW 837T ($546.5B)
2028: KRW 902T ($588.9B)
2. SK hynix
2026: KRW 327T ($213.5B)
2027: KRW 623T ($406.8B)
2028: KRW 667T ($435.5B)
if those estimates are right
> samsung trading at 2.4x 2027 PE
> skh trading at 2.6x 2027 PE
I mean yea I like the guy but he called 50 of the last 2 market crashes so that should tell you something
MU to $1500 with haste
MU to $1500 with haste
picked up a small amount of $SHAZ today, thesis will follow when I’m done researching
this is an extremely high r/r trade, don’t just tail blindly it can gap down hard
this is an extremely high r/r trade, don’t just tail blindly it can gap down hard
> $SHAZ is building AI infrastructure. think GPU cloud and AI factories and compute capacity for enterprises and AI labs rather than training foundation models themselves
> currently sits at a ~$2.8B market cap with a 30D average volume of ~1.28M shares. still a relatively underfollowed name compared to the larger AI infra plays
> $2.2B+ of contracted backlog today
> $1.25B / 5 year AI cloud agreement with ESDS, followed by another ~$950M take or pay agreement with a global APAC technology customer
> recently closed $1.6B strategic financing removing what was arguably the company’s biggest bottleneck being capital
> Leopold Aschenbrenner’s fund disclosed a 19.9% stake making it one of the company’s largest shareholders.
management plans to scale AI infrastructure to 100MW+ supporting 20,000+ GPUs over time
the bear case is execution. the company has to actually deploy the capital, build the infra, and convert these contracts into recurring revenue. it’s also a capital intensive business meaning any delays or cost overruns or customer issues could weigh on the stock. some investors have also questioned the quality so that’s something i’ll continue digging into
there’s speculation around additional hyperscaler and AI lab relationships including Anthropic but I haven’t found enough evidence to treat those as anything more than rumors for now
GPU capacity remains scarce so if SHAZ executes on even a majority of what’s already under contract i don’t think a $10B market cap is an outrageous destination. also worth noting SHAZ was just listed on HL today. liquidity is still thin and the order book is empty so i’d be careful with placing orders until volume builds up.
will keep digging this week
> currently sits at a ~$2.8B market cap with a 30D average volume of ~1.28M shares. still a relatively underfollowed name compared to the larger AI infra plays
> $2.2B+ of contracted backlog today
> $1.25B / 5 year AI cloud agreement with ESDS, followed by another ~$950M take or pay agreement with a global APAC technology customer
> recently closed $1.6B strategic financing removing what was arguably the company’s biggest bottleneck being capital
> Leopold Aschenbrenner’s fund disclosed a 19.9% stake making it one of the company’s largest shareholders.
management plans to scale AI infrastructure to 100MW+ supporting 20,000+ GPUs over time
the bear case is execution. the company has to actually deploy the capital, build the infra, and convert these contracts into recurring revenue. it’s also a capital intensive business meaning any delays or cost overruns or customer issues could weigh on the stock. some investors have also questioned the quality so that’s something i’ll continue digging into
there’s speculation around additional hyperscaler and AI lab relationships including Anthropic but I haven’t found enough evidence to treat those as anything more than rumors for now
GPU capacity remains scarce so if SHAZ executes on even a majority of what’s already under contract i don’t think a $10B market cap is an outrageous destination. also worth noting SHAZ was just listed on HL today. liquidity is still thin and the order book is empty so i’d be careful with placing orders until volume builds up.
will keep digging this week
❤4👏2