DanielRomero

The only reason memory is trading lower is that the market is still in denial about the pricing power
“Hyperscalers won’t spend 40% of their capex on memory.”
“Hyperscalers won’t accept memory suppliers earning 85% gross margins.”
It’s cognitive dissonance at this point
There isn’t even an argument or any supply-and-demand analysis
It’s just denial of reality

Open-source models improving gives tremendous power to neoclouds
This is what many people don’t understand
> $NBIS, for example, only offers open-source models through its managed inference platform
> It allows enterprises to fine-tune and deploy open models for internal workloads much more cheaply than using the OpenAI or Anthropic APIs
As open models improve, demand for this infrastructure increases
The GPU supply remains the same. The number of valuable workloads competing for that supply rises
That increases utilization, strengthens pricing power and gives neoclouds, especially leading one

$NBIS just announced that it will fully transition to post-quantum encryption either this year or next
It’ll be interesting to see who its partner will be

> Rubin is on track
> $META is doubling data center deployments in 2027
> Anthropic is close to profitability
> OpenAI 5.6 compute can’t keep up with demand
> $ASML beats estimates, while $TSM raises capex
What is keeping the market down at the moment?
The Iran war? Rate-hike fears?
This doesn’t seem like a structural, fundamentally driven drawdown to me
It looks more like a temporary dip within a stronger upcycle

$SPCX becomes interesting around $70
I don’t know how you justify buying $SPCX instead of $META at the same market cap
Just as a random example, $META generates $130B in annualized OCF
How long will it take $SPCX to generate that?
5 years?

It’s surprising how lightly everyone is taking the fact that trillion-dollar companies are undergoing an almost bear-market-like correction

What could drive memory prices lower?
Some industry insiders believe it's a change in product specifications
“Wu Jinrong, general manager of MicroDrive Technology, warned that supply-and-demand distortions leading end customers to adjust product specifications are more likely to cause memory prices to turn in the second half of 2027 than capacity expansion.”
I agree, yet it’s not happening
There’s no way memory content can be designed down enough for that to occur

Interesting
The market may be greatly overestimating how much new memory supply is coming
“Recent supply chain reports suggest that SK Hynix’s actual new memory production capacity in 2028 may be only one-sixth of the amount initially planned.”

Dumb headline
The ADR has nothing to do with SK Hynix’s current price movement on the KOSPI
The solana:SKHYhSjuRWHgikq8eRKbtBbpABgJSkd7ytQV14i9EQ3 ADR is currently trading at a 32% premium, assuming it stays flat when the US market opens (it won’t)
This dump isn’t “euphoria cooling down” at all

Morgan Stanley names $NVDA a top pick with a $288 price target
Analysts expect the company to generate $20.50 in EPS in FY29, implying a 14× FY29 earnings multiple at the target price

The sentiment shift on $META is ridiculous
It shows that most people form their opinions just based on vibes
Recency bias is at ATHs right now

Who bought the $MU dip?
It got below $900 just two days ago
$MU gives you the returns of a microcap and the volatility of a small cap, all at a $1.1T market cap

While the market fears excess compute after the $META news…
$META is reportedly planning to double its data center deployments to 14GW in 2027, up from around 7GW in 2026, per Reuters
Massive acceleration
Now ask yourself: what kind of demand are they seeing to justify that level of capital investment?

BofA estimates $NVDA annual shipments add up to 99 GW of data center capacity by 2030
That is solely for NVIDIA hardware
Add ASIC and $AMD deployments, and you could get to over 150 GW of installed capacity
So, so bullish for energy and data center plays