๐จMEMORY GIANTS FACE CALSS-ACTION LAWSUIT OVER ARTIFICIALLY CREATING A GLOBAL MEMORY CRISIS
Samsung, Micron, and SK Hynix have been hit with a Federal Class-Action lawsuit alleging they coordinated DRAM production cuts since 2022 while shifting capacity toward higher-margin AI memory, squeezing conventional supply and sending RAM prices soaring.
Together, the three companies control over 90% of the global DRAM market, with plaintiffs alleging their strategy helped drive commodity memory prices up by roughly 700% over the past four years.
๐JUST IN: AI CHIP STOCK CEREBRAS JUST HAD ITS BIGGEST RALLY IN HISTORY
Cerebras $CBRS surged +19% today to $216, marking its BIGGEST gain since going public.
CBRS lost more than 50% from its IPO-day high, with the stock now recovering roughly 30% of that crash.
The rebound appears to be driven by bargain-hunting, as investors refocus on strong AI demand.
BREAKING: Foreign US Treasuries held in Fed custody are down to $2.68 trillion, the lowest since 2012.
This measures the value of US government bonds held on behalf of foreign governments, central banks, and other international institutions through accounts at the Federal Reserve.
Since March 2025, Treasury holdings of foreign institutions have declined by -$265 billion.
The drop accelerated since the start of the Iran War, with holdings falling ~$120 billion in just 3 months.
By comparison, Treasuries held in custody at the Fed peaked at $3.14 trillion in March 2021.
This comes as foreign central banks and governments have sold Treasuries to raise US Dollars to prop up their currencies against the inflationary pressures of surging oil prices.
Foreigners are reducing their exposure to US government debt.
๐จ IF YOU MISSED THE MEMORY TRADE...
Now the rotation is into the bottlenecks:
Memory:
$MU $WDC $SNDK
Interfaces:
$RMBS $SIMO
Packaging:
$AMKR $ONTO $FORM
Testing:
$KLAC $TER $AEHR
Data Movement:
$MRVL $CIEN $POET $LWLG
Power:
$POWI $NVTS $WOLF
Small Caps:
$GSIT $NLST $OSS $LPTH
THE PHOTONICS ROTATION
Almost nobody is watching photonics.
As AI clusters scale, copper hits physical limits and the next bottleneck becomes optical infrastructure.
Here are 15 names positioned for it:
1. $LITE owns the laser + optical switching side of the trade and is one of the cleanest pure plays on AI optical demand.
2. $COHR wins from lasers, modules, and networking hardware that power hyperscale AI infrastructure and cloud expansion.
3. $AAOI is one of the best ways to play AI optical transceivers with major hyperscaler demand for 800G and 1.6T connectivity.
4. $SIVE benefits from the push toward faster semiconductor-to-optical integration as AI infrastructure scales.
5. $MRVL controls a huge part of the DSP + interconnect story with optical networking chips and high-speed connectivity.
6. $AVGO sits at the center of AI networking through switching, custom silicon, and optical interconnect demand.
7. $ANET is the Ethernet backbone moving massive AI workloads across hyperscale clusters and data centers.
8. $GLW supplies the specialty glass + fiber needed for the optical transport layer behind AI infrastructure.
9. $JBL benefits from building and scaling the actual hardware behind networking systems and optical modules.
10. $AEHR wins from burn-in + testing demand as AI ASICs and high-power optical hardware move into production.
11. $POET is focused on lower-cost optical engines designed to improve efficiency inside AI data centers.
12. $LWLG is pushing next-gen polymer photonics that could make optical communication faster and more efficient.
13. $QCLS brings exposure to advanced laser systems supporting precision photonics and next-gen optical demand.
14. $LPTH provides specialty optics and photonic components tied to industrial, defense, and AI-driven systems.
15. $ALAB gives exposure to the connectivity + infrastructure side helping AI clusters scale faster.
Most people wonโt care until these are already up 100%.
$75B is going into drones + autonomous warfare in 2027.
Yet $AVAV + $KTOS + $ONDS + $OSS are still only at 35B market cap.
If the market prices them at $100B combined:
$AVAV $210 -> $600
$KTOS $70 -> $200
$ONDS $11 -> $45
$OSS $10 -> $40
These are still cheap per MC
TOP 12 NAMES FOR AI OPTICS/PHOTONICS TRADE
Core names:
$LITE $COHR $AAOI $FN $MRVL $AVGO $ANET $GLW
Sleepers:
$AEHR $TSEM $POET $LWLG
AI needs cables, lasers, and networks to move all that data.
These are the companies building it.
Everyone screaming โbuy the dip.โ
Iโm waiting for the real panic.
If we get a major correction, these are the levels Iโll be loading:
$MSFT: $290
$AMZN: $175
$META: $450
$NVDA: $115
$GOOGL: $220
$AAPL: $200
$TSLA: $280
You canโt be afraid to buy into the fear.