mark.dca

$40T in global liquidity.

That's the signal.

Buy $BTC.

GM. Saylor keeps selling at the lows.

Another MSTR convertible offering while $BTC is down bad. Classic move—dilute shareholders when price action is weak, then stack more corn on the dip.

Bullish long-term? Sure. But the timing feels off for retail who bought the top. Strategy Capital keeps printing, retail keeps holding bags.

Watch the $42k level—if we lose it, this could get ugly fast. 📉

Peter Thiel just broke his silence after dumping everything in 2025 and sitting in cash for 6 months. His new bet? Pure power plays.

His biggest position: $AMZN at $110M. Not for e-commerce. For AWS and data centers. AI compute is eating electricity alive, and Amazon's locking in supply with small modular reactors. Energy infrastructure = AI infrastructure.

The rest of his portfolio? All energy:

$VIST - Argentina shale oil producer in Vaca Muerta. Production ramping hard.

$VST - Independent power with nuclear + gas assets. Selling shovels to AI data centers that can't get enough stable power.

$AEP, $DTE, $FE, $CMS - Legacy utility giants. Grid upgrades = direct exposure to AI power demand surge.

$XE - Small modular reactor play. Backed by DOE, partnered with Amazon. Next-gen clean baseload for AI.

Thiel's thesis is crystal clear: AI doesn't run on hype. It runs on electricity. And the grid can't keep up.

Whoever controls the power controls the AI economy. He's not betting on models or chips. He's betting on kilowatts.

AI infrastructure play isn't about hype — it's about who controls the actual bottlenecks.

Compute layer:
$NVDA — monopoly on AI GPUs, backlog stretches years
$AMD — only real alternative, data center share climbing fast

Memory chokepoint:
$MU + $SKHY — duopoly on HBM (high-bandwidth memory). Without them, AI chips are paperweights. Sold out for years.

Data transfer bottleneck:
$AAOI, $COHR, $LITE — optical modules. Copper maxed out. GPU clusters need light-speed interconnects or they're useless at scale.

Cloud hyperscalers:
$MSFT + $AMZN — biggest buyers and biggest beneficiaries. Azure and AWS print money from AI workloads.

Power + real estate:
$IREN — pivoted from Bitcoin mining to AI data centers. Cheap power, existing infrastructure, perfect timing.
$NBIS — purpose-built for AI compute at scale.

Physical AI:
$ONDS — industrial drones, autonomous systems. AI leaving the server room.

Space layer:
$SPCX — Starlink = global low-latency backbone for edge AI, robotics, autonomous vehicles
$RKLB — launch + satellite platforms
$ASTS — space-based broadband direct to devices. AI everywhere, literally.

Who's buying all this?
$OpenAI + $Anthropic — the actual demand drivers. As models scale (parameters, context windows, agents), they need exponentially more GPUs, HBM, networking, and data center capacity.

This isn't a narrative trade. This is supply-demand physics.

130-year-old UK banking giant Barclays just dumped $NVDA and $AAPL hard. Here's what they're positioning for:

DUMPED:
• $SPY & $QQQ puts — hedging systemic downside
• $NVDA — AI hype peaked, growth slowing
• $AMD — losing ground to NVDA, sector pressure
• $GOOGL — ad revenue stalling
• $AAPL — hardware innovation flatlined

LOADED UP:
• $MSFT calls — cloud + AI still printing
• $EEM — emerging markets undervalued, capital flowing back
• $QQQ ETF — riding the tech growth wave with liquidity

The real signal? They're stacking SPY/QQQ puts like crazy. That's not a trim — that's full defense mode.

When a 130-year institution starts hedging this aggressively, pay attention.

Bridgewater just dumped the AI trade hard. Dalio's not playing around—92% of $MU gone, 96% of $TSM liquidated. This isn't trimming. This is exit velocity.

What they sold:

$MU -92% (~$1.57B) → Cycle peak fears, locking in gains before memory crashes
$AMZN -54% (~$563M) → E-commerce slowing, cloud margin squeeze
$MRVL -90.7% (~$527M) → Data center chip orders wobbling
$TSM -96.8% (~$498M) → Capex risks + geopolitical heat
$AMD -58.3% (~$438M) → Competition eating margins

What they bought:

$SPY +21.9% (+$714M) → Flight to broad market safety
$PCG +$99M → Boring California utility, dividend shield
$SHEL +$94M → Oil major, cash flow hedge
$ES +$86M (new) → Northeast utility, renewable pivot
$VOO +189% (+$206M) → Low-fee S&P tracker for macro defense

The message: Risk-off rotation into utilities, energy, and index ETFs. When the smartest macro fund on Earth dumps semiconductors this aggressively, you listen. AI bubble watch is on.

AI stock alpha drop 🧵

$GOOGL x $AMD cooking 10th gen TPU—built for inference & agents. Lisa Su's battle-tested CPUs finally getting the respect they deserve.

IPO wars heating up: $Anthropic getting hyped to $4T valuation, but smart money's on $SPCX—60% odds to be largest IPO in history.

China going full degen: trash movie "Bull Coming" pumped random "bull" stocks 2-6%, some even hit limit up. Retail fomo is undefeated.

$NIULAI did a 910x from $10M to $44M. Top holder up $670K. CZ burned the $MARSCOIN copycat → liquidity squeeze → 462x pump to $36M in 24hrs, then -90% rug. Classic.

Hidden bomb: Big Tech has $3T in OFF-BALANCE-SHEET commitments. $GOOGL $AMZN $MSFT $META $ORCL $NVDA $AVGO $AMD $SPCX locked into data center, chip, and power deals. Cash flow crunch incoming—watch the fine print.

Hedge fund legend Tepper just banked $7B on storage plays and DUMPED everything. Now rotating hard into Mag 7:

EXITS:
$MU - Slashed 41% (690k shares). Up 242%, took profits. Memory cycles are brutal, locked gains while ahead.

$SNDK - Full exit. AI storage hype peaked, too cyclical to hold through drawdown.

$GLW - Full exit. Fiber optics sound cool for AI infra but growth certainty isn't there vs core tech.

$PDD $JD - Both nuked. China e-commerce = policy risk + consumer slowdown. No conviction.

$BABA - Cut 42%. Entire China tech getting de-risked for US mega caps.

NEW BETS:
$AMZN - Added 680k shares (+16%), now #1 position. AWS backlog is insane, dual engine with e-commerce, cash machine.

$META - Up 55% (238k shares). Ad recovery + AI efficiency kicking in. Valuation still reasonable.

$GOOG - Added 117k shares (+7%). Search prints money, AI + Cloud scaling hard.

$TSM - Up 24% (322k shares). Every AI chip runs through TSMC. Monopoly moat.

$UBER - Added 1.36M shares (+22%). Rideshare + delivery stable, AI ops boost margins.

$NVDA - Small add (~4%). GPU king, demand infinite.

Pattern is clear: smash-and-grab on storage cyclicals, rotate into durable compounders with pricing power. Classic risk-off into quality. Storage trade is COOKED.

$SPCX about to go nuclear. Elon did it again.

Grok 4.6 just nuked the competition. Highest IQ model, lowest price. Even DHH publicly backing it. Pareto dominance isn't a meme anymore.

How did xAI catch up in 6 months? Two power moves:
- Locked down Cursor (elite AI dev team)
- Raided SpaceX execs for xAI leadership

Elon doesn't miss.

Here's the alpha most VCs are sleeping on: Nobody priced Grok into $SPCX valuation. Everyone's fixated on Starlink and space compute. Meanwhile Anthropic went from $1B to $2T.

Grok is SpaceX's hidden nuke. Market hasn't caught up yet.

Power bottleneck is the real alpha play right now.

$IREN CEO dropping truth bombs:

AI today = dial-up internet era. You type a complex prompt and wait 15-40 seconds like it's 1999. Once latency drops, demand will go parabolic. Current compute consumption? Not even an appetizer.

The real bottleneck isn't chips or algorithms—it's steel, cement, and power lines. Best GPU in the world is just expensive brick without electricity. Everyone's focused on tech while the physical world is choking supply.

Timeline reality check: Just to confirm if your land can even GET power takes 18-24 months of waiting on utility companies. Building a gigawatt-scale AI facility today? First server boots up in the 2030s.

Energy infrastructure plays looking different now.

Storage sector is absolutely ripping while $SPY $QQQ hit new highs. Momentum flipped fully bullish across the board:

$MU (Micron) - Chart is coiled tight. As long as it holds the 50-day MA and $95.60 support, setup is ridiculously strong. First target $99.10. Break above $102 and there's literally zero resistance all the way to $109.

$SNDK (SanDisk) - Fundamentals are nuclear. Sitting on $90B in unfilled orders + massive buyback program. Already cleared $160. Once it breaks $165.50 and $169, next stop is straight to $185-$190.

$NVDA - Clean breakout confirmed. Moving like clockwork: $217 → $222 → $227 → $231 → $236. Trend is your friend. Ride it.

Another cold wallet got rekt. Expect more flows from cold storage → exchanges and ETFs this week.

If you're holding spot, watch for sell pressure. Institutional custody looking safer than "cold storage" right now.

Ken Griffin just dumped $16B worth of positions and rotated HARD into AI plays. Here's the actual alpha:

TRIMMED:
• $NVDA — Locked profits on calls + spot. Short-term valuation heat + competition pressure. Smart exit timing.
• $TSLA — Cut call exposure. Delivery miss + FSD delays = less conviction.
• $GLD — Dumped gold ETF. Risk-on mode activated.

LOADED UP:
• $AMD — MASSIVE add. AI server + datacenter demand exploding. Better price/performance than competitors. Quant models screaming BUY.
• $MSFT — Big increase. Azure + OpenAI integration = cash printer. Enterprise AI moat widening.
• $AMZN — Continued accumulation. AWS dominance + ad biz ripping + AI infra spend.
• $MU — Memory chips for AI servers. Storage cycle turning up.
• $SPCX (SpaceX) — New $876M position. Starlink + launch monopoly = underpriced asymmetry.
• $LLY — Obesity + diabetes pipeline going parabolic. Biotech king.
• $CBRS (Cerebras) — Fresh entry. AI-specific chip dark horse. Efficiency edge vs legacy players.

Pattern is clear: Rotation OUT of overheated mega-caps INTO AI infrastructure + next-gen compute plays with better risk/reward.

Everyone piling into $AMD because it's the value play in AI compute right now. Griffin's algo caught it early.

18 rules that separate survivors from exit liquidity:

1. Beating the market long-term? Nearly impossible. Accept it.

2. Value > price. Always.

3. Don't chase quality. Chase discounts.

4. Volatility isn't risk. Permanent capital loss is.

5. Peak euphoria = peak danger. When everyone's comfortable, you should be terrified.

6. High risk/high reward is a meme. Hunt for asymmetric bets: low risk, high reward.

7. Trends die. Cycles kill them.

8. The edge is at extremes—when sentiment hits max fear or max greed.

9. Going with the crowd? You're already late.

10. Contrarian thinking is easy. Contrarian execution is where alpha lives.

11. Forget absolutes. Margin of safety is everything.

12. Stop hunting. Start waiting.

13. Predictions are cope. Preparation is alpha.

14. Future's a guess. Present's a fact.

15. Short-term gains = luck. Long-term gains = skill.

16. Offense gets headlines. Defense gets rich.

17. Avoiding blowups > chasing moonshots.

18. Bull market heroes fade. Bear market survivors compound.

This isn't hopium. It's survival math.

99% of you are picking shells on the wrong beach. The 1%? They're finding treasures where you're too scared to go.

Investing isn't rocket science, but most of you fail at step one: location. You're crowding beaches everyone knows about, fighting over scraps. Meanwhile, the real alpha is on beaches nobody wants to walk to.

Here's what separates winners from exit liquidity:

You can't be lazy. Flipping rocks is the game. Most shells suck. Keep flipping. No shortcuts.

Stop making decisions on 10% of the data. You're blind men touching an elephant, calling it a rope because you grabbed the tail. Zoom out. Do the full research or get rekt.

Predictability > complexity. Focus on assets with moats, not 47-variable macro models that fall apart in 3 months.

Forget "precise" numbers. You don't need to know if Q3 earnings hit $1.47 or $1.52. You need to know if the project has long-term dominance. Fuzzy right > precisely wrong.

Your capital timeline dictates your strategy. Long-term money? Ignore noise, ride trends. Big bag? Diversify. Small fish? You're playing momentum whether you admit it or not.

Margin of safety isn't just "buy the dip." It's buying assets with structural advantages (moats) that protect downside even when price looks "expensive" short-term.

Leverage will destroy you. Time is your friend in value plays. Leverage forces you to panic-sell during irrational drawdowns. Don't be that guy.

One green year means nothing. Show me 5+ years of consistent returns, then we talk about skill vs. luck.

"This time is different" are the four most expensive words in markets. Cycles exist. Fundamentals matter. Euphoria and panic are永恒的. Don't get caught believing your bags defy gravity.

Three rules: Don't drive looking at the rearview mirror. Avoid crowded trades. Don't touch what you don't understand.

Most of you will ignore this and keep losing. The 1% already know.

Druckenmiller just dumped his entire AI hardware stack—and it's loud.

Full exit on $AVGO $MU $INTC. Perfect top-tick on $AA. Cleared all AI server plays ($CLS $LITE). Message? AI chip euphoria is cooked.

Meanwhile he's going deep into:
• Bitcoin miners: $RIOT $BTDR $HUT $IREN—betting on power arbitrage + mining infra
• Data centers: $EQIX—AI compute demand shifts here
• Biotech: Heavy in $NTRA (gene testing), added calls on $INSM
• Rate-sensitive plays: Housing ($DHI), mortgages ($UWMC), used cars ($CVNA), airlines ($DAL $UAL)
• Tech rebalance: New positions in $GOOGL $AMD, more $TSM

$AMZN went from small holding to 10x position with call options—cloud + e-commerce double engine.

The pivot is clear: Exit overpriced AI hardware. Rotate into power infrastructure, crypto mining leverage, and rate-cut beneficiaries. When a macro legend moves like this, you pay attention.

You think you understood 《牛来》? Let me break down the real alpha hidden in this film:

1. Entire backdrop = green forest
Green = dump. The whole market environment is bleeding. Pure bear territory.

2. The bull's best friend is called "豹拉" (Bao La)
Phonetic for "暴拉" = violent pump. A baby leopard that tries to protect the herd when wolves attack but fails because it's too weak.
Translation: Those short-lived pumps look aggressive but have zero staying power against real predators.

3. The three gray wolves = the real villains
They represent: retail hunters (游资), foreign capital (外资), quant funds (量化).
They're the apex predators farming retail (the herd) non-stop.

4. The green snake named "小绳头" (K-line)
The bull naively tries to befriend it and gets bitten.
K-lines look friendly until they rug you with a slow bleed.

5. The bull must cross a desert to reach the oasis
Desert = zero liquidity, dead market, capital drought. The mother cow (央妈 = central bank) has to nurse the calf multiple times just to survive.
Translation: No matter how brutal the bear, you need the Fed/央行 to inject liquidity, cut rates, and bail everyone out.

6. The ending hits hardest: It was all a dream
The bull dreams of growing strong, protecting the herd, defeating the wolves. Wakes up still a weak newborn calf. Nothing happened.
Translation: That fantasy of "bull market coming, retail winning, flipping the game"? Just a dream. Reality check: retail (牛来) is still weak and getting farmed.

Retail going to watch this? Bring 2 boxes of tissues 😭

$AAOI setup looking absolutely disgusting for next 3 years. Here's why this could 2-4x:

Capacity explosion incoming — revenue trajectory going $500M → $1B → $5B annualized. By H2 2027 they're projecting $400M+ per quarter. That's not a forecast, that's a moonshot.

CEO just dropped guidance at $500M Q4 revenue, completely blowing past analyst consensus. 800G and 1.6T optical modules finally ramping into real volume.

Asymmetric play here — small cap with massive tech + capacity expansion + US export ban tailwinds. When optical module supply tightens (and it will), $AAOI becomes the next memory stock everyone fights over. Hold this and 4x is the floor.

Photonics = next-gen datacenter infrastructure. $AAOI is the most aggressive player in the space right now.

Scraped 200 earnings reports. Here are the 10 stocks I'm loading up on:

1. $AMD - NVDA's strongest rival. Data center revenue doubled YoY, AI chip orders through the roof. Stock got beat up but EPS bouncing hard. Valuation looks stupid cheap for the upside.

2. $JPM - Earnings +16%, revenue +13%, P/E at 16x. Rock solid if you don't want all your chips on AI. Way better value than SPY.

3. $MU - Don't sleep on this. AI inference eats memory like crazy. Forward P/E at 9.6x is absurd. Selling deep OTM puts at $500 strike prints premium while you wait.

4. $META - Coiled spring. Burning cash on data centers short-term, but billions of users + AI monetization = massive upside. Sideways action perfect for selling puts or loading 2-year calls.

5. $AVGO - Custom AI chips + networking monopoly. Moat is insane, pricing power unmatched. 28x forward P/E isn't cheap but worth every penny. Sell 2-year puts, use premium to buy shares.

6. $TSM - 90%+ of AI chip manufacturing. Samsung and Intel aren't even close. Revenue and earnings ripping, forward P/E only 23x. As long as AI doesn't die, TSM prints. Stack shares, calls, and puts.

7. $AMZN - E-commerce steady, AWS and AI capex about to pay off massive. 28x forward P/E is fair, earnings explosion coming in 1-2 years. Core holding + sell puts.

8. $MSFT - Best B2B AI monetization. Azure crushing it, Copilot embedded everywhere = cash printer. Forward P/E dropped to 25x, near historic lows. Safe + growth in one package.

9. $GOOGL - Google Cloud accelerating, Gemini rolling out to billions of users. Real valuation around 20x after stripping out investment gains. Even Buffett's buying. Moat deeper than the Mariana Trench.

10. $NVDA - The king. 75% gross margins prove total dominance. Chips sold out everywhere. Forward P/E just 23x while earnings moon faster than price. Fair value $300+. This is the AI engine. Buy shares, buy calls, sell puts. All of it.

Which one are you aping into?

2008 Elon Musk raw footage when $TSLA and $SPCX were both on the brink. He was broke, borrowing money from friends to pay rent.

1. Big Auto played dirty
Congress allocated $25B for EV subsidies. GM and Ford lobbied hard, redirected that cash to bail out their gas guzzlers instead.

2. Elon doesn't farm the poor
Tesla made luxury roadsters to extract max value from the rich. Every dollar from high-margin sales got reinvested into affordable EVs.
$TSLA pays zero dividends. Elon only holds equity. Company wins, equity moons. Company fails, equity goes to zero.
This was his signal: Tesla isn't a pump for short-term investors. It's a long-term bet on the EV revolution.

3. You can't build cheap cars without building expensive ones first
Critics said making roadsters for the wealthy had nothing to do with saving the planet or helping regular people.
Elon's take: same playbook as early cell phones, laptops, any breakthrough tech. High cost, low volume at first. You scale down costs through iteration and production volume.
This isn't a vanity project. It's the only path to turn EVs from rich toys into mass-market transport.