John Lee (@niker8202) header
John Lee (@niker8202) profile picture

Focused on optics, semiconductor substrates & Japan-related material stocks.

Joined October 2025
Photo by John Lee (@niker8202)Photo by John Lee (@niker8202)Photo by John Lee (@niker8202)Photo by John Lee (@niker8202)Photo by John Lee (@niker8202)Photo by John Lee (@niker8202)Photo by John Lee (@niker8202)Photo by John Lee (@niker8202)Photo by John Lee (@niker8202)Photo by John Lee (@niker8202)
I am so bullish on this one. Thanks to @yongwoonkang Santec Holdings (TSE:6777) is a Japanese optical device maker founded in 1979, producing tunable lasers, optical measurement instruments, and optical components. Nvidia and Huawei are among its major customers as AI infrastructure demand surges. Revenue over the past year hit 31.5 billion yen, up 31.1% year-over-year, with an operating margin of 32.8% — five straight years of record profit. It's essentially the "pick-and-shovel" play on the AI boom: as data centers switch internal wiring from copper to fiber optics, demand for optical components and the instruments that test them is exploding, and Santec sits right at that chokepoint.
2
1
7
2,157
I covered this earlier in Jan regarding PCB bottlenecks with photonics + $AMZN Kuiper / $SPCX starlink supply chains. But as an update, it like other industries outside of AI are really starting to feel the pain now too. Global Electronics Association: "Is issuing this Industry Alert on escalating shortages across the printed circuit board (PCB) supply chain. Companies in automotive, industrial, medical, defense, and telecommunications report longer lead times and difficulty securing supply as AI absorbs scarce capacity, effectively putting production at risk" The association meaningful relief unlikely before mid-2028. The association urged buyers to plan ahead for supply constraints and manufacturers. Hyperscalers like $AMZN already planned ahead of time taking stakes in Gold Circuit Electronics (2368) for supply procurement... But can't say the same about other industries procurement (medical industry is probably not used to competing with hyperscalers over the same supply lol). The resulting shortages of critical materials, including CCLs and fibreglass cloth, are squeezing supplies for consumer electronics. TLDR: same bottleneck, same timeline for 2028. It's just interesting to see association warnings about it impacting medical to defense industries. So there's incremental information that it's become an industry wide allocation issue. I also think it's helpful to see escalating shortages + broadening... so it might show the selloff is starting to look very disconnected with fundamental demand.
41
13
1
361
142,119
You can’t call that bottom’s in on a falling knife. These things can keep on falling. You first need to see some certain things happening so you can call bottom: 1. Selling Exhaustion comes first. We may see a selling climax as well, with stocks/indices on extreme dips, but then we have to see them stop making significant lows. 2. We have to successfully retest previous lows with buyers responding more aggressively into buying the dips. 3. Maybe see a Wyckoff spring with price briefly breaking the established support, but with sellers failing to sustain another breakdown. 4. We need to see a base accumulation forming without new lows. 5. Downtrend structure to start breaking with price breaking above previous lower highs. !! 6. Bullish Volume confirmation is needed. You can’t call bottom with significant selling volume! 7. Price to start trending above the accumulation range. 8. RSI Improves with stock outperforming the market or its benchmark. 9. Price reclaiming important Moving Averages. 10. Broader market confirmation with indices recovering as well, just as institutional buying becoming more evident again. 11. Most important for me, is to see a successful retest on previous breakout level. You see a stock holding its breakout as support, then you’re good to go and call bottom in. TLDR: 1. Possible bottom : Selling exhaustion + initial support + bullish divergence. 2. Probable bottom : Successful retest + accumulation + improving relative strength. 3. Confirmed trend reversal : Higher low + break above significant swing high + strong volume + sustained upward structure. You can’t call bottom when : ❌ RSI below 30 ❌ MACD bullish crossover alone ❌ A stock falling 70–90% ❌ One large green candle ❌ Bullish divergence alone ❌ Touching historical support ❌ A short-term oversold bounce Have that in mind. ✍🏻
21
35
6
456
52,415
In June 2024, my firm published our prediction for Nvidia $NVDA reaching a $10 trillion market cap by 2030, stating, “If this estimate materializes, Nvidia’s data center segment will be at $320 billion in 2027, up from data center run rate of $90 billion today.” In November 2025, my firm raised our outlook to $20 trillion by 2030, explaining that “Nvidia is well on its way to reaching my firm’s brand-new updated estimate for a $230 billion data center quarter by Q4 of 2030 or $930 billion for the full year.” Fast forward to today, and Nvidia’s TTM Data Center revenue has reached $334 billion, reaching our original prediction from two and a half years ago – and if Nvidia can maintain Q2’s and Q3’s growth pace at 110%+ YoY, it could hit that $230 billion quarterly Data Center milestone before the end of 2028. In this week’s analysis, we take a step back from Nvidia’s latest numbers, which continue to reaffirm that trajectory, and instead present to you the Top 5 Reasons that the AI Trade has Only Begun. Link in bio.
112
89
9
1,033
151,382
We are thrilled to expand our collaboration with @GlobalFoundries. In addition to deploying Spot across their global manufacturing network, GF will be the first semiconductor manufacturer to test humanoid robots as part of Atlas early adopter program. GF has worked with Boston Dynamics and Spot for over four years and recently expanded the program to Singapore, where two Spot robots have traveled nearly 900 kilometers and completed more than 90,000 inspections in sub-fab areas since April, achieving a 99% mission success rate. bosdyn.co/4jmxibq
39
46
18
343
18,922
“Move Europe/Japanese equities to overweight.” I'm very bullish on Japanese stocks. This correction just made the setup even better.
Morgan Stanley sends a note to clients using the term: “Max Overweight” for global equities. Arguing that AI capex cycle remains intact and sustains corporate earnings into 2027. They told clients: - hold less cash, trim commodities and exposure to bonds - move Europe/Japanese equities to overweight - Global stocks trade ~15x fwd earnings (MSCI ACWI), the cheapest since 2023 - AI theme is expected to spread beyond the hyperscalers - Global earnings growth is forecast at 13% in 2027 and 8% in 2028 But flagged: - Inflation could force more tightening than the two further Fed hikes - AI capex cycle may not last as long as assumed Earlier this week they argued: “The hedge fund deleveraging unwind has been brutal since October. While retail investors remain calm and long-term oriented, levered institutional hedge funds have acted as ‘chickens with their heads chopped off’ (IMO), moving daily stock prices around at nonsensical speeds.” Anyway the most recent time I found this term used was 14 years ago June 8th, 2012. Where they called “maximum” overweight on emerging-market equities. Usually I treat this as noise but TIL “max overweight” was a term used.
2
147
John Lee retweeted
대만 증시 역대급 떡락 터짐… 외인·기관 1,140억 대만달러 패대기 16일 대만 증시가 엔비디아 칩 국가안보 논란 터지면서 하루 만에 2,635.04포인트(-6.03%) 급락한 41,040으로 마감함. 역대 두 번째로 큰 낙폭이고, 패닉셀 쏟아지면서 거래대금만 8,450.57억 TWD 터졌음 수급 보면 메이저들이 작정하고 물량 던짐. 기관 총 순매도만 1,140.55억 TWD인데, 뜯어보면 외인 및 외국계가 701.89억 TWD, 투신 179.62억 TWD, 딜러가 259.03억 TWD씩 전부 매도 폭탄 날림 원인은 젠슨 황 사임 이후 미국 안에서 반도체 국가안보 이슈가 계속 커진 탓임. AI 산업 앞날이 불투명해졌다는 공포감이 시장 전체를 덮쳤는데, 대만 증시 자체가 AI 하드웨어 하청·부품 밸류체인 몰빵 구조라 직격탄 맞고 투매가 쏟아진 거임 대장주 TSMC부터 5.74% 빠지면서 2,215 TWD에 마감함. 미디어텍 약 8%, ASE 테크놀로지 8%, 액튼은 9% 넘게 개박살 남. 델타일렉트로닉스, 유니마이크론, 난야 PCB, 아시아바이탈컴포넌츠 같은 핵심 기술주들은 아예 하한가 직행함. 특히 AI 데이터센터 핵심인 CPO,PCB,방열, 후공정 관련주들이 제일 심하게 두들겨 맞음. 이날 상장·상장예정 종목 통틀어 무려 168개 종목이 하한가 처맞았고, 대만 증시 주가 상위권인 고가주 10개 중 6개도 하한가로 잠김 그동안 기세 좋던 유리섬유 소재주도 단체로 급락함. 그나마 전통산업 쪽에서는 내수 방어주 성격인 식품이랑 건설 업종만 상승하면서 피난처 역할을 함. 해운주는 매도 압력 받으면서 에버그린이랑 양밍은 하락 전환했고, 완하이 혼자 약 2% 오르면서 버팀.
5
13
2
62
10,604
The "REAL" Bubble is Consumer Staples.
Consumer Staples P/E has soared to 24.2x... that's the highest since 1999 @yardeni
12
10
79
8,240
John Lee retweeted
摩根士丹利:AI获利了结之后,下一轮投资主线正在浮现 2027年开始,AI产业链的投资机会可能进一步向上游材料和核心零部件扩散,ABF载板、CCL覆铜板、PCB等领域有望迎来新的供需变化。
51
30
7
445
468,600
John Lee retweeted
You do understand why this could ignite the BIGGEST RALLY in AI optics yet, right? 🚨Bernstein forecasts InP supply to surge 60% in 2027 and another 67% in 2028, with $AXTI delivering the majority, yet supply still can't catch demand. In the same research, Bernstein estimates $NVDA Rubin Ultra shipments will expand $NVDA's optical ecosystem by 91% in 2028, from roughly $24.5B in 2027 to $46.8B in 2028, driving 70% to 90% annual InP optical demand growth. You can already see that wafer scarcity in $AXTI's financials: Revenue climbed from $26.9M in Q1 2026 to $47.6M in Q2 and $73.4M in Q3, pushing GAAP gross margin from 29.6% to 56.0%, with commercial backlog above $500M and Q4 revenue guided to roughly $90M. Downstream US laser makers are paying years ahead to lock in supply. $LITE posted a 42.3% Non-GAAP operating margin in Q1 FY27, landed over $5B in NPO and CPO purchase commitments, and wired $AXTI a $43.5M wafer prepayment. $COHR signed a 5-year $20B PhotonLink deal anchored by 400mW CW lasers, with Bernstein putting its Rubin Ultra opportunity at $9.4B in 2027 and $36.1B in 2028. What makes China central to this trade is that it sits on both ends of the bottleneck. $AXTI's main production base is Beijing Tongmei in China, which signed $285M to $545M in private equity financing to expand InP capacity, and wafers shipped inside China require zero export permits. Riding that local demand, Chinese InP laser maker Yuanjie Technology (688498 SH) grew 9M 2026 revenue 320% YoY to RMB 1.61B (~$241M) and net profit 867% YoY to RMB 1.02B (~$154M) at an 81.7% gross margin. Every layer of this chain is expanding into pre-committed orders, with customers helping fund the buildout upfront. $AXTI $COHR $LITE $NVDA
9
13
3
143
28,162
$NVDA <The Quiet Removal> The faction war probably began the moment he pushed his daughter to the FRONT LINE. It was not just one seat. Daughter: Senior Director, Physical AI Marketing Son: Director, Robotics Son-in-law: VP, Global AI Infrastructure All in the core business, all at once. In Asia, using status to secure seats for your children is NOTHING NEW. But doing it this openly is the FASTEST way to make enemies. Everyone inside can see exactly who got the seat, and why. And look at how he was removed. So quietly, so instantly. That tells you there were many inside who hated him and envied him. Of course, the affair was wrong. But that alone does not explain this speed.
140
🚨Qualcomm HBC = HBM killer? $Qcom Short answer: not yet. It's coexistence. 1. HBM is for training. Expensive, only 3 makers, sold out through 2027. HBC is for inference. Stacks cheap LPDDR used in phones, with a compute chip underneath. 2. So here are Qualcomm's bragging numbers. 18x bandwidth vs LPDDR, 5x efficiency per watt vs HBM4. Data doesn't travel far, it just goes straight down, so it uses less power. 3. But HBC still needs DRAM. Who makes LPDDR? #Samsung, $SKHY , $MU It's not stealing HBM demand, it's creating new demand. 4. Also, 133TB/s is an internal combined number. A GPU has 6-8 HBM stacks, so real world is 20-27TB/s, so the gap is more like 5-7x. And it's still a 2-layer test chip, production needs 4 layers. Yield is the key. Bottom line: Not the end of HBM. As inference grows, the memory market is splitting in two. AI250 sampling in 2027, AI300 in 2028 will be the real test. Video by @yongwoonkang youtube.com/watch?v=rQfz1Y6Z…
3
380
John Lee retweeted
📘【三井金属(5706)決算|営業利益45%増!AI向け銅箔好調、1,900億円の新工場投資】 AIサーバー向け銅箔の需要が拡大! 上期営業利益は575億円(+44.7%)。 通期営業利益予想を940億→990億円に上方修正。 さらに、マレーシアで約1,900億円の銅箔新工場建設を発表しました。 📊 2027年3月期 上期決算 ・売上高:4,311億円(+18.3%) ・営業利益:575億円(+44.7%) ・経常利益:609億円(+55.7%) ・純利益:395億円(+107.6%) ・営業利益率:13.4%(前年10.9%) 📌 ① 機能材料|AI向け銅箔が好調 ・売上高:2,118億円(+40.7%) ・営業利益:464億円(+65.4%) ・経常利益:451億円(+70.6%) 半導体パッケージ向け極薄銅箔「MicroThin」と、AIサーバー向け高周波基板用銅箔「VSP」が増販。 販売数量の増加や価格改定、在庫要因の改善が大幅増益に寄与しました。 📌 ② 金属|増収も営業減益 ・売上高:2,115億円(+31.2%) ・営業利益:157億円(▲11.3%) ・経常利益:188億円(+2.2%) 亜鉛・銅などの金属価格上昇と円安で増収。 一方、製錬コストや操業面の負担が重く、営業利益は減少しました。 📌 ③ 銅箔新工場に約1,900億円を投資 11月10日、マレーシアでの新工場建設を決定。 ・投資額:約1,900億円 ・対象:MicroThin、VSP ・稼働開始:2031年4月から順次 既存拠点も含めた能力増強により、銅箔事業の営業利益は2025年度比で2030年度3.2倍、2033年度5.5倍を想定。 ただし、新工場の収益寄与は2031年度以降です。 📌 ④ 通期業績予想を修正 ・売上高:8,500億→8,750億円 ・営業利益:940億→990億円 ・経常利益:1,000億→1,040億円 ・純利益:800億→780億円 銅箔の販売好調を受けて営業利益は上方修正。 一方、退職給付制度改定に伴う特別損失64億円を計上し、純利益予想は20億円引き下げました。 📌 ⑤ 配当方針を引き上げ ・DOE目安:3.5%→4.0% ・期末配当予想:14円→18円 ・年間配当予想:28円→32円 累進配当方針は継続。 ※年間配当は10月の1株→10株の株式分割後換算。 📌 ⑥ CFは利益成長と対照的 ・営業CF:+149億円(前年+320億円) ・投資CF:▲144億円 ・FCF:+5億円 ・自己資本比率:59.2% 棚卸資産の増加や税金支払いなどで営業CFが減少。 今後は銅箔事業の増産投資と現金創出力のバランスが重要になります。 🧠 今回の決算の核心 AI・半導体向け銅箔の販売拡大が、機能材料事業の大幅増益を牽引。 会社は約1,900億円の新工場投資を決定し、中長期の成長に向けて供給能力を拡大します。 一方、通期純利益は下方修正され、営業CFも減少。 旺盛な需要と利益成長に加え、投資回収の進捗を確認する必要があります。 🧭 今後の確認点 ・MicroThin・VSPの販売数量 ・銅箔事業の利益率 ・新工場の建設・稼働スケジュール ・非鉄金属価格と在庫評価損益 ・営業CFと設備投資負担 ・増配方針の持続性 今後は、AI向け銅箔の成長を継続的な利益と現金創出につなげられるかに注目です。
📘【本日11/10(火)|日本企業の注目決算🇯🇵】 本日はソフトバンクG・フジクラ・ホンダ・スズキ・川崎重工・住友金属鉱山・三井金属など、各業界の主力企業が決算発表! AI・光通信・半導体材料・自動車・防衛関連の業績見通しに注目です。 📅 決算発表予定時刻(日本時間) 🕘 09:00|西松建設(1820) 🕚 11:00|雪印メグミルク(2270)      日清紡HD(3105) 🕦 11:30|川崎重工業(7012) 🕛 12:00|大気社(1979) 🕧 12:30|三井海洋開発(6269) 🕐 13:00|東レ(3402)      神戸製鋼所(5406)      SUBARU(7270) 🕑 14:00|注目決算! ・フジクラ(5803) ・東京応化工業(4186) ・三井E&S(7003) ・スズキ(7269) ・太平洋セメント(5233) 🕝 14:30|住友金属鉱山(5713) 🕒 15:00|ふくおかFG(8354)      オプトラン(6235)      筑波銀行(8338) 🕞 15:30|主力企業の決算が集中! ・ソフトバンクG(9984) ・ホンダ(7267) ・三井金属(5706) ・日本製鋼所(5631) ・アルバック(6728) ・浜松ホトニクス(6965) ・アズビル(6845) ・三菱ガス化学(4182) ・エーザイ(4523) ・資生堂(4911) ・シスメックス(6869) ・DeNA(2432) ・ネクソン(3659) ・住友不動産(8830) ・加賀電子(8154) 🕓 16:00|りそなHD(8308) 📌 特に注目したいポイント 🔹ソフトバンクG:AI投資・Arm・投資事業の損益と財務 🔹フジクラ:データセンター向け光通信・電線需要 🔹東京応化:先端半導体向けフォトレジスト材料 🔹三井金属:AI半導体向け電子材料・銅箔 🔹住友鉱:銅・ニッケル価格と製錬・材料事業 🔹アルバック:半導体・電子部品向け製造装置 🔹浜松ホト:光デバイス・半導体関連需要 🔹日本製鋼所:産業機械・防衛関連の受注 🔹川崎重工:防衛・航空宇宙・エネルギー事業 🔹ホンダ・スズキ・SUBARU:北米・インド需要、関税・為替影響 本日は14:00にAIインフラ・自動車関連、15:30にはソフトバンクGをはじめ主要企業の決算が集中! 業績の市場予想比、通期見通しの修正、受注・利益率・営業CFの変化に注目です。 ※発表時刻は予定。変更・前後する場合があります。 #日本株 #決算発表
1
6
1
26
14,465
Insane 2-day bounce. Limit-up yesterday at +22.73%, still green today. Dip buy was a win.
Replying to @santec_corp
@santec_corp Checking one of my holdings, Santec's first-half results were insane. Revenue up 65%, net profit almost doubled. Optical components sold well thanks to North America, and measuring instruments sold well thanks to optical communications. So the company sharply raised its full-year outlook to ¥50B (approx. $315M) in revenue and ¥18B (approx. $114M) in operating profit. The reason is simple, demand for optical communications is still strong, and they can sell even more from the second half as production expands. I actually see the pullback as an opportunity. In the end, the stock price follows earnings.
127
John Lee retweeted
【6315TOWA】 半導体モールディング装置(樹脂封止装置) ICチップやLED、パワー半導体を樹脂で包んで保護する装置、世界シェアNo.1 シンギュレーション装置(切断・分離装置) 封止された半導体パッケージを高精度に切り分ける装置 超精密金型 極めて高い精度の半導体用樹脂封止金型、超硬工具の製造
6315 TOWA 2,185円 S高 +400 +22.41% 決算受けてS高 短信↓ towajapan.co.jp/jp/wp-conten… 機関投資家・アナリスト向け説明資料↓ towajapan.co.jp/jp/wp-conten…
1
1
1
5
4,129
Santec (6777) LIMIT-UP BUY IMBALANCE 🔥 Massive earnings + dividend upgrade on strong optical components demand!!
Replying to @santec_corp
@santec_corp Checking one of my holdings, Santec's first-half results were insane. Revenue up 65%, net profit almost doubled. Optical components sold well thanks to North America, and measuring instruments sold well thanks to optical communications. So the company sharply raised its full-year outlook to ¥50B (approx. $315M) in revenue and ¥18B (approx. $114M) in operating profit. The reason is simple, demand for optical communications is still strong, and they can sell even more from the second half as production expands. I actually see the pullback as an opportunity. In the end, the stock price follows earnings.
285
【生産量は「ラピダスの40倍」 マスク氏半導体工場、調達部隊が極秘来日】 イーロン・マスク氏の半導体工場計画「テラファブ」が動き始めた。日本の製造装置メーカーに発注計画を示し工場稼働を急ぐ。半導体業界は新たな顧客出現を歓迎する一方、常識外れともいえる要求に困惑する声もあがる。 「いま半導体業界で一番の話題はキオクシアでもラピダスでもない。イーロン・マスクだ」。サプライヤー幹部は興奮気味に話す。 出典:日本経済新聞 nikkei.com/article/DGXZQOUC0…
4
94
21
368
50,319
John Lee retweeted
💫 Small announcement from my side: This Fear & Greed Index will be part of the Meridian expansion to become a truly OWN WEBSITE. When I started the Meridian project, it was clear to me that it wouldn't stay a TradingView indicator forever. That was my dream, and my plan. The recent TradingView restrictions then accelerated the need (and also desire) to be independent of other platforms, and to have my own instead. Of course, there will be more scores to assess the current market situation at one glance. All of them will feature statistics and a chart to understand what happened after prior such readings - totally transparent. And the best part...there's going to be a tier introduced not including the Meridian system itself, but including both the charting platform and those market scores - and much more! Thanks to everyone for their positive feedback on the Meridian system - I'm so hyped to get this to the NEXT LEVEL now!
Meridian Research's Fear & Greed Index is sitting at 16, reminiscent of prior near bottom or bottom scenarios. $SPY
10
8
52
12,179
John Lee retweeted
SK Hynix $SKHY: “The volume already secured under LTAs and the demand currently being discussed with customers taken together means that our DRAM and HBM volumes for the next 3 years through 2029 are effectively sold out.” $MU $NVDA $AMD
40
56
10
672
114,601