311😉

[Shay Boloor @StockSavvyShay]

10 SMALL-CAP NAMES UNDER $5B WORTH WATCHING

1. $CLPT navigation platform for neurosurgeons
2. $OSS rugged edge compute for the battlefield
3. $OUST eyes & perception for physical AI machines
4. $NVTS GaN power for $NVDA 800V data center shift
5. $AMBQ ultra-low-power silicon for always-on edge AI
6. $AMBA edge AI vision chips for cameras, cars & robots
7. $TMDX expanding organ care with fleet & network effect
8. $TE building America's solar supply chain from cell to module
9. $AEHR wafer-level burn-in for photonics, HBM & high-power AI chips
10. $ONDS autonomy & counter-drone systems for the modern battlefield

[✧ 𝕀𝔸𝕄𝔸𝕀 ✧ @iamai_omni]

$FORM：HBM 晶圆级测试 probe cards（HBM 堆叠测试瓶颈，Serenity 明确列出）。
$AMKR ：先进封装（HBM stacking 关键）。
$ONTO ：封装计量/检测设备。
$ACLS ：离子注入设备（内存工艺核心）。
$VECO ：激光退火设备。
$PLAB ：光掩膜（lithography photomask）。
$RMBS ：内存接口 IP。
$MKSI ：激光子系统。
$ADEA ：混合键合（hybrid bonding）技术（先进封装/HBM 相关）。

[华尔街观察 Xtrader @cnfinancewatch]

玻璃基板，康宁 $GLW

全球封测， $ASX

玻璃基板， $LPK

铲子股： $AMAT $KLAC $LRCX

拿走不谢！美股必买AI硬件科技股。

[Serenity @aleabitoreddit]

Very interesting statement today: $MU CEO predicts a multi-decade memory demand cycle driven by humanoid robots.

"Humanoid robots, he says, will require roughly ten times more memory than today’s Level 2+ autonomous vehicles."

"And that demand wave is set to begin before the decade is out."

Something as well as was "Over time, we expect the value of on-device AI combined with pent-up unit replacement demand to drive memory demand growth"

Which is also another trend (Apple Intelligence is currently dog, but I'm sure we'll see innovations with localized/edge AI).

Feels like all the industry leaders from $TSM Chairman, $TSLA Elon Musk, to $MU CEO see humanoids as the next major trend so physical AI is probably next.

I wonder if the world is going to have enough memory. Or if we'll see enough breakthroughs to shrink memory usage.

[311😉 @kayeeee31]

@bubbleunoffical 上两倍了～

[311😉 @kayeeee31]

很炸裂的一晚上………

[311😉 @kayeeee31]

@labubu_trader Me either 🥹

[Art of Speculation @ArtofSpecuycky]

看了很多朋友的留言，顺便把这些AI板块我心里对应的美股（含ADR）代表公司整理一下

1. 存储（HBM + DRAM + eSSD，已经涨价、签长单、利润兑现）：MU、SNDK、DRAM ETF
2. GPU（AI时代的算力引擎，训练和推理的绝对核心）：NVDA、AMD
3. 光互连（800G / 1.6T + Scale-Up + Scale-Across，连接接棒算力）：LITE、COHR、MRVL、NOK、AAOI
4. 先进封装（CoWoS + CoPoS + HBM封装，HBM和GPU放量的前提）：TSM、AMKR、ONTO
5. ASIC + Networking（自研芯片、交换芯片、SerDes、DSP）：MRVL、AVGO、ALAB、CRDO
6. CPU（EPYC、ARM、Grace、Axion，推理时代CPU重要性回升）：ARM、AMD、QCOM、INTC
7. 电力 / AI Factory（变压器、电网、UPS、800V DC、BBU，并网容量是真瓶颈）: ETN、GEV、NVTS、FPS
8. 液冷（DLC、CDU、冷板，高功率机柜必选项) ：VRT
9. 核能 / 发电（长期空间巨大，但兑现节奏慢于电力设备）：CEG、VST、OKLO、SMR、BWXT
10. PCB / ABF / CCL（AI服务器和先进封装配套材料，真实受益但控制力弱于前面）：TTMI（PCB）、UMICY（ADR，ABF载板龙头）
11. MLCC / 被动元件（高频高压高功率带来用量提升，属于配套瓶颈）：MRAAY（ADR，全球MLCC龙头）
12. 半导体设备（检测、量测、测试、先进封装设备，卖水给淘金者）AMAT、KLAC、LRCX
13. Physical AI（机器人、自动驾驶、工业AI，空间大但2026仍偏预期）：TSLA（机器人 + 自动驾驶 + 世界模型）、QCOM（Edge AI + 汽车 + AR）、BB、AMBA（边缘视觉AI SoC）、OUST（数字激光雷达）
14. Agent / 数据层（方向确定，但赢家和利润分配还不清晰）：MDB、SNOW、DDOG
15. 稳定币 / AI金融（AI Agent未来支付层，空间巨大但商业化还早）：CRCL

[Art of Speculation @ArtofSpecuycky]

我把AI未来3年的核心瓶颈排了个序，大概是这样：

1. 存储（HBM + DRAM + eSSD，已经涨价、签长单、利润兑现）
2. 光互连（800G / 1.6T + Scale-Up + Scale-Across，连接接棒算力）
3. 先进封装（CoWoS + CoPoS + HBM封装，HBM和GPU放量的前提）
4. ASIC + Networking（自研芯片、交换芯片、SerDes、DSP，MRVL/AVGO核心位置）
5. 电力 / AI Factory（变压器、电网、UPS、800V DC、BBU，并网容量是真瓶颈）
6. 液冷（DLC、CDU、冷板，高功率机柜必选项）
7. 核能 / 发电（长期空间巨大，但兑现节奏慢于电力设备）
8. PCB / ABF / CCL（AI服务器和先进封装配套材料，真实受益但控制力弱于前面）
9. MLCC / 被动元件（高频高压高功率带来用量提升，属于配套瓶颈）
10. 半导体设备（检测、量测、测试、先进封装设备，卖水给淘金者）
11. Physical AI（机器人、自动驾驶、工业AI，空间大但2026仍偏预期）
12. Agent / 数据层（方向确定，但赢家和利润分配还不清晰）
13. 稳定币 / AI金融（AI Agent支付层，空间大但商业化还早）

未来三年市场会越来越关注：谁是真瓶颈，谁能把瓶颈兑现成利润，谁控制下一代架构。
我目前最看好的，还是存储第一，光互连第二，先进封装第三。因为AI发展到最后，拼的是谁真正卡住了整个产业链的命脉。

[Shay Boloor @StockSavvyShay]

THE ELECTRICITY VALUE CHAIN

• $CCJ uranium supply
• $CEG nuclear baseload
• $SMR modular nuclear reactors
• $OKLO micro-nuclear generation
• $EOSE zinc long-duration storage
• $GEV power conversion hardware
• $VRT power distribution & cooling
• $VST flexible generation & storage
• $TSLA utility-scale lithium storage
• $FSLR thin-film solar for hyperscale
• $NNE utility-scale renewable buildout
• $TLN North American renewable generation
• $BWXT nuclear components & specialty fuel

[Caesar Capital @CaesarCapitalz]

I’m in $SIVE, $AAOI and $AEHR

[311😉 @kayeeee31]

@labubu_trader Fxk

[311😉 @kayeeee31]

@octopusycc @168X_Fortune 大老师也上168x了！

[Omer Cheema @Omercheema]

According to GoldmanSachs, CPO Switch BoM Revenue Contribution per Unit (Quantum-X Photonics / NVIDIA):
Estimated supplier revenue per $130k switch (based on BoM values):
NVIDIA (Switch ASICs + Optical Engines design): ~$12k (ASICs) + significant share of engines → major portion.

TSMC (Fabrication/Packaging of ASICs & engines): Substantial share of ~$44.4k combined ASIC/engines value.

Corning (FAUs + SM Fiber): ~$3.6k + $12.3k = ~$15.9k.

Lumentum / Coherent / Sumitomo (ELS + CW Lasers): ~$7.2k.

Senko / others (MPO connectors): ~$5.8k.

Additional (Shuffle Box, assembly): Balance to total BoM ~$75.8k.

[311😉 @kayeeee31]

@Kay2289123 axti之前被刷下车 🥲

[311😉 @kayeeee31]

@woaitaoershi 怎么就官宣了！！

[Paradis Labs @ParadisLabs]

Equal-weight “Neocloud” ETF:

1. $GLXY
2. $WULF
3. $CIFR
4. $APLD

These were the most compelling to me based on four metrics:

1. Counterparty quality (likelihood of cashflows)
2. Secured -> Energized gap & schedule (when does it start paying?)
3. EV / Total Secured MW
4. Balance sheet capacity

It's very hard to pick a clear sector winner since every name ranks differently:

> WULF are top for the Secured -> Energized schedule but have the weakest balance sheet.

> Similarly, GLXY has the best balance sheet w/ low dilution risk, but they have some counterparty risk due to 100% concentration on $CRWV.

> Then you've got $APLD who built around their CoreWeave exposure w/ a more diversified book now i.e. 200 MW hypserscaler agreements.

That's basically the reasoning for constructing an ETF.

Then looking into H2'26 and 2027+, my thinking is that the sector should continue to outperform since lease revenue flows around then. Which is where the cash flow becomes contracted + longer-dated, which warrants AI infra multiples rather than lower miner multiples.

Will be fun to track how the ETF performs over time though (dollar amount is for illustrative purposes only).

Disclosure:

The ETF forms a fairly small portion of my wider portfolio given how violent the betas are.

If we see significant dips (>10%) then I'll buy more, if not, I'll let it compound. If I need cash for more compelling/higher conviction buys (e.g. a $SNDK / $MU dip), then I'll likely trim the ETF.

[Daniel Romero @HyperTechInvest]

AI Infrastructure Roadmap

What you need to research to outperform over the next few years

Now → memory + optical transceivers

2027 → 800V + early CPO

2028 → PLP + scale-up optical

2029 → glass substrates + HBM5

2030 → optical I/O chiplets + embedded cooling

2031 → 3D DRAM + microfluidic cooling

[qinbafrank @qinbafrank]

不得不说伯恩斯坦这份报告宝藏很多，最近市场大热PCB、ABF、基板以及更上游的材料板块，它五月初的深度报告里详细地阐述了看好逻辑。

[qinbafrank @qinbafrank]

https://t.co/hH0FQPgIt3