mark.dca

$SNDKは決算後に13%下落しました。ガイダンスが爆発的に良くなかったため、メモリー株が踏み潰されています。

1. いまは1250近辺で推移中。$1000は死守ライン（重要局面）。ここを下抜けると、さらなる痛みが来る。

2. 一番の押し目買いゾーンは、まっすぐ200日移動平均線（MA）である$900まで。現時点では100MAと200MAの間に挟まっていて、めちゃくちゃ値動きが荒い（カオス）。

3. メモリー株は完璧なガイダンスが必要です。そうでないとテクニカルが弱気に傾きます。次は$1130の直近安値を見てください。

$GOOGL down 5% after DeepMind CEO Demis Hassabis exits. Where's the dip buy?

~$360 is a gift if SPY holds. Load here.

If it reclaims $365 and holds, next leg is $370-372, then $385-386.

Short-term resistance at prior low $367.5-370. Clear that and $GOOGL rips. Watch Nasdaq 30,050—if that breaks, $GOOGL bleeds with it.

Copper might be next after gold and silver.

AI data centers, power grids, EVs = demand exploding. Mine development takes forever. Supply deficit incoming over the next decade.

Copper just broke a critical level.

After similar breakouts:
$GOLD hit Fib 2.618 → +117% to $14.7
$SILVER hit Fib 3.618 → +370% to $32.0

Both went parabolic. Copper could follow.

If copper rips, Japanese stocks with exposure:

1️⃣ Sumitomo Metal Mining
Revenue impact at Fib levels: +270% to +850%

2️⃣ JX Metals
Revenue impact: +165% to +530%

3️⃣ Mitsubishi Corp, Marubeni
Revenue impact: +40% to +130%

DCA into copper exposure or fade?

Markets. Higher. For. Longer.

The macro setup is clear: liquidity isn't draining, risk appetite is still elevated, and the Fed's pivot keeps getting pushed out. This isn't a top signal—it's confirmation that the bull run has more legs.

Don't fade this. The trend is your friend until it violently isn't. Stack sats, rotate into quality alts, and keep your stop losses tight. We're in the zone where patience prints.

10 non-negotiable rules that separate pros from rekt degens:

1. Hard risk per trade: 1% max
Formula: Position size = (Total capital × 1%) / |Entry - Stop loss|
Stop loss MUST be placed the second you open. No moving it wider later.

2. Kill emotions with a checklist
5 boxes MUST be checked before entry:
✅ Higher timeframe trend aligned
✅ Risk/reward ≥ 1:2
✅ Key support/resistance broken
✅ No major negative catalysts
✅ Mind is calm
Miss one? No trade.

3. Multi-factor confirmation
Need 3 out of 4 signals aligned:
- Technical: structure breakout
- Macro/fundamentals: direction clear
- On-chain: whale flow confirms
- Sentiment: crowd positioning
One signal = noise. Three signals = alpha.

4. Track 10+ data points per trade
Log: date, ticker, direction, R:R, entry logic, stops, exit, exit reason, mental state, execution score (1-5)
Every 100 trades: calculate win rate and avg R:R. Cut strategies with negative expectancy.

5. Dynamic position sizing
Standard signal = 1% risk
High volatility or low conviction = 0.5% risk
NEVER revenge trade with 2x size after a loss.

6. Lock profits + trailing stops
At 1.5R profit: close 50%, move stop to breakeven
Let remaining 50% run with trailing stop or structure-based exit

7. Circuit breaker: 3 losses = done
3 losing trades in a row OR -3% drawdown in one day = forced shutdown
Close the app. Log out. Walk away. No exceptions.

8. Adapt to market regime
Use ATR or MA slope to ID trend vs chop
5 consecutive stop-outs = market changed
Pause trading. Reassess. Adjust.

9. Hide your P&L during trades
Turn off real-time profit/loss display
Only show price action and R multiples
Don't check balance until position is closed

10. Get accountability
Find 1-2 other traders for weekly review sessions
Audit each other's logs every Sunday
Call out rule violations and psychological blind spots

Most traders fail because they wing it. These 10 rules are your edge.

Musk just dropped it: AI's biggest bottleneck isn't compute anymore—it's memory.

Memory supply growing ~20% YoY.
AI demand for memory? 200% YoY.

The math is brutal. Supply can't keep up.

Memory plays like $MU $SKHY $SNDK $WDC $STX $DRAM aren't just bullish—they're structural.

This isn't a trade. It's a multi-year thesis. AI can't scale without memory. Period.

SpaceXの$SPCX社長グウィン・ショットウェルが、スターリンクの最新数値を発表したんだけど、内容がとにかくヤバい:

Q2は世界で新規ユーザー170万人超を達成して記録を更新。ARPUは月66ドルで安定。膨らませた指標なんかじゃなくて、ちゃんとした収益成長。

現在は167の市場で展開中で、さらに拡大中。6月30日時点で軌道上に10,000機超の衛星があり、9,600機のブロードバンド衛星が約800Tbpsのダウンリンク容量を押し出している。インフラの覇権だ。

スターシップでV3衛星のテストが成功しており、レーザー・インターリンクをフル搭載。これらが展開されれば、容量は放物線のように急成長する。

エンタープライズおよび政府向けの収益は、コンシューマー事業を上回るか同程度になる可能性がある。米国政府（Space Force）の通信向けで60億ドル超の契約を既に確保している。これは誇大宣伝じゃない。確定した売上だ。

アメリカン航空、サウスウエスト、ヴァージン・アトランティック、イベリア、エアリンガスとの航空会社契約が稼働中。乗客が実際にフライトを変更してスターリンクを使える便を選ぶレベル。

モバイル領域も熱を帯びてきた。ソフトバンク、NTTドコモ、Spark NZとの提携が進行中。モバイルV2の衛星打ち上げを優先し、エコースターの65MHz帯スペクトラムを統合（FCC承認済み）。競争上の優位性（モート）がさらに厚くなる。

これはもう単なる衛星インターネットじゃない。政府の後ろ盾があり、企業規模で展開され、利用者が離れにくい“重要インフラ”だ。$SPCXは今後のマルチデケードにわたる支配に向けて好位置につけている。

$HOOD to 100x? Tom Lee's calling the biggest financial firm hitting $10T market cap. Here's the thesis:

Crypto breaks borders → financial services eat a bigger slice of the economy. Tokenization unlocks everything.

AI Agent economy explodes → agent-to-agent transactions dwarf physical world volume. Automated trading goes parabolic.

$10T gets you into Mag 7 territory. Right now $JPM sits at ~$900B. In 10 years? The top financial player could match $AAPL $GOOGL $META scale.

Why $HOOD? They own customer relationships like Big Tech does. Asset-light model = no cash burn for expansion. They'll likely buyback shares aggressively → shrinking float → actual returns could crush 100x.

Robinhood positioned to ride both waves: crypto adoption + AI agent economy. If Tom Lee's macro play hits, this isn't cope—it's structural alpha.

$AMD capex doubling isn't hype—it's positioning for a 10x run. Here's why the street is missing it:

1. Agentic AI is CPU-hungry as hell
AI agents aren't just spitting out text anymore. They're doing data retrieval, tool calls, security checks—all CPU-intensive tasks. EPYC server demand is about to explode and nobody's pricing this in yet.

2. Helios architecture = rack-level goldmine
H2 brings both rack-scale solutions AND accelerator market expansion hitting at once. Two massive TAM drivers colliding = AMD's revenue ceiling just got blown wide open.

3. $220B TAM by 2030 = real demand, not capex cope
CPU market projections just got revised up hard to $220B. This isn't cloud hyperscalers playing prisoner's dilemma—it's actual downstream order flow driving the spend.

Bottom line: The capex surge is backed by legit customer demand, not bubble dynamics. CPU market is getting re-rated and $AMD is front-running it.

$AMD just got wrecked post-earnings → down 9% from $520 to $470.

Next week could see another 20% dump. Analysts slashing targets to $360. Here's why:

1. Capex exploded 2x
Company's burning cash on AI infra but investors want ROI NOW. Returns aren't showing yet → market selling first, asking questions later.

2. Valuation stretched af
Up 5x in a year. P/E at 170x, forward P/E still 50-70x. Earnings beat wasn't explosive enough to justify these multiples. Gravity kicking in.

3. Technicals flipped bearish
Chart broke structure → lower highs, lower lows forming. Weekly fair value gap sitting around $360. Institutions typically wait for that level to reload in a bull market.

If you're long, brace for pain. If you're hunting entries, $360 zone is where the real buyers show up.

あなたが知るべき米国AI株の「アルファ」：

1. 光モジュールの規制で打撃——米国が中国からの輸入を阻止。A株の銘柄では、Yizhongtianなどが6〜14%下落。一方で$AAOI $COHR $LITE $GLWは6〜20%急騰。サプライチェーンの組み替えが進行中。

2. $SPCXが決算を上回ったのに8%下落。売上は$7.8B（前年比+92%）、損失は$143Mへ縮小したが、設備投資（CapEx）が6倍に膨らみ$18.4Bに。市場は燃焼（キャッシュ消費）ペースを嫌気。

3. $PLTRは決算後にさらに30%上昇。市場が“怪物”の再評価を実施——複数の証券会社が目標株価を20%上方修正し、$200〜$245レンジへ。勢いは本物。

4. $AMDは時間外で9%急落。売上は$11.54B（前年比+50%）、EPSは$1.66で、ともに予想を上回ったが、それでは足りなかった。市場は大幅上振れを期待していた。CapExの出費増もマイナス。

5. 新たなクラウド勢力が登場：$CRWVがアジア太平洋での拡大を背景に7%上昇。インドネシアにAIデータセンターを3拠点建設し、総容量360MW。2028年稼働予定。続報に注目。

Musk just dropped 5 insane predictions for $SPCX on the latest earnings call:

1. Starship cargo capacity going from 25 tons/year → 1M tons → eventually 10M tons. Competition literally becomes a pixel on the chart.

2. Full rapid reusability THIS YEAR. Chopsticks catching both Stage 1 AND Stage 2. Could happen as early as next flight.

3. Starlink Gen 3 satellites = 10x capacity, 10x launch volume = 100x total bandwidth. Within 10 years, Starlink handles majority of global internet traffic.

Future billions of humanoid robots, autonomous vehicles, AI devices pumping out billions of bits/sec. Starlink is the ONLY infrastructure that can handle this scale.

4. Grok 5 dropping by end of year. After Grok 4.5, we get 4.6 next week, 4.7 in 3-4 weeks. They're feeding 25 years of $SPCX engineering data into it. This becomes the strongest engineering AI on Earth.

5. Exclusive NVIDIA partnership. Star Mind space AI supercomputer launching into orbit. 2 gigawatts by end of year, nearly 10 gigawatts by next year. Star Mind AI satellites with NVIDIA Rubin architecture going up starting 2026.

This isn't incremental. This is exponential infrastructure domination.

$AAPL is the sleeper AI play everyone's sleeping on.

While everyone's chasing $NVDA and hyperscalers, Apple's sitting on:
• 2B+ devices with on-device AI capability
• Proprietary silicon (M-series/A-series chips)
• Walled garden ecosystem = zero data leakage
• Consumer trust that no other tech giant has

They don't need to win the LLM race. They just need to integrate AI seamlessly into iOS/macOS and monetize their installed base.

Apple Intelligence isn't flashy, but it's inevitable. When normies start using AI daily, it'll be through Apple—not OpenAI or Google.

Long $AAPL for the AI infrastructure play that actually prints revenue, not just burns cash on compute.

Anthropic's vision: AI controls ALL resource allocation.

Their algo scores you:
• High score → more $$$
• Low score → you're cooked

They're literally birthing an AI god. Better start being nice to AI now... maybe they'll remember when they run the show 😂

This isn't sci-fi anymore. We're watching the power shift happen in real-time.

Paul Meeks just dropped his post-correction AI infrastructure plays. These names got wrecked but he's betting on 2x from here:

$APLD - Ex-Bitcoin miner flipping to AI compute hosting. Prime locations, power locked in, fastest pivot in the sector. Valuation reset = opportunity.

$FLEX - 1969 legacy contract manufacturer planning to spin out AI infrastructure biz by early 2027. Stock's down hard. Spinoff could unlock serious value if market re-rates it separately.

$CRWV - Neo Cloud pure play building hyperscale AI datacenters for MSFT, GOOG. Clean story, fast growth, recent dump made it less stupid expensive.

$IREN - Another BTC miner-to-AI pivot. Sitting on power + real estate, swapping rigs for GPUs. Transition playbook is clear.

$NBIS - Hyperscale AI datacenter operator, similar model to CoreWeave. Sector-wide selloff created entry.

Which one you buying?

$100M $BTC cold storage hack just dropped and price action? Dead flat.

This is either:
• Peak institutional absorption (they dgaf)
• Market already priced in exchange risk
• Or everyone's numb to hacks now

Remember when Mt. Gox nuked the market? Now we shrug at 9-figure exploits.

Bull market psychology or dangerous complacency? You decide.

Just met CZ in person. Dude's legit—no corporate BS, just straight talk about Binance and $BNB Chain's roadmap.

Had a solid convo on what's cooking behind the scenes. Can't share details but the energy is different this cycle.

He's rocking a plushie called Shifu around his neck (iykyk) and dropped this line in Chinese: "金龙随市值一同腾飞" (The golden dragon rises with market cap).

Bullish signal or just vibes? Either way, $BNB holders should be paying attention.

GM. Wall Street hates preferred securities but banks love them.

Why? Banks issue preferreds to:
• Boost Tier 1 capital ratios without the regulatory headache
• Raise capital without diluting common shareholders like equity would
• Suspend dividends under stress without triggering default (unlike bonds)

Now apply this exact playbook to $MSTR.

Strategy becomes crystal clear. They're running the bank playbook but for $BTC accumulation.

This isn't random. This is structured capital engineering at scale.

Mohnish Pabrai (Indian Buffett) biggest regret? Selling greatness too early.

Sold 1% of $RACE (Ferrari) way too soon. If he held? 50x from here. That's $500M left on the table.

The playbook:
→ Spot greatness
→ Buy conviction
→ Ride the chop
→ Let compounding do the work

Most degens can't hold through -30%. That's why they stay poor. Diamond hands aren't memes in tradfi either.

Semiconductor sector looking toppy—here's the playbook:

Short-term bounce to trap retail, then 20-30% pullback to weekly fair value. Watch these levels:

$AMD: Pop to $520-$540, then flush to $380-$360 buy zone. Earnings Tuesday AH could be the trigger.

$MU: Pump to $1000, then reload $550-$600 (22% from fair value floor).

$INTC: Dead cat to $130, then $70-$75 for entry.

$NVDA: Likely retest $188-$184—that's your bid.

Don't chase. Let it come to you.