Pinned Tweet
Hello, I’m MT.
1,000+ of you are following now. Thank you! ❤
We all know 1,000 is a small number. I see it as a starting point: a place to learn crypto and money together, in public.
A bit about me: 9 years as an investor, community manager and builder in crypto. Here I share my research notes on crypto and money, plus a few shitposts.
What you’ll find:
- How I research projects and opportunities
- What I did with my own money, wins and losses
- Lessons on how to make, manage and keep your money
- The mindset and psychology behind money decisions
I’d love to connect with:
- web3 builders and project teams
- investors and active players
- creators
- friends on X growing financially and personally
Now it’s your turn. Say hi and introduce yourself: what are you learning or building right now? I promise to read and reply to every comment.
All my notes are for education and information only. Not financial advice.
Keep your money first. Make money second.
– MT 📝
Keep your money first. Make money second.
I learned this the expensive way.
I started learning about crypto in 2017. On January 2, 2018, I opened my Binance account. $BTC had already peaked. The whole crypto market peaked five days later.
In the 2018 winter, I lost 90% of my capital.
My first big loss came from one project. I fell for its rosy picture of the future. I married the story and ignored the numbers.
Nine years in, here's who I am:
- Most of what I've earned came from early-stage opportunities: token sales, Launchpool, airdrops, DeFi
- 2020-2023: marketing team of a crypto project, plus community management
So I know what a project shows you, and what to check instead.
A poker lesson I use every day: if you can't spot the easy money at the table, you are the easy money.
The crypto table is full:
- insiders and early funds with tokens to unlock
- bots that react before you can
- accounts paid to sound sure
- and your own FOMO
That's why this profile is my public crypto diary: Binance Alpha, airdrops, token launches and sales, and the trades I take.
How it works:
- Transparent: I rate each opportunity from S to F before I join, then share the result. Trades get posted when they close, with the trade card and my reasoning. Wins and losses both go up.
- Experience-based: everything here is what I did with my own money. My money, my responsibility.
- Work and learn: I show the process and keep improving it, so you can learn the method, not copy the trade.
- Open feedback loop: you comment, I read and reply. Good ideas from the comments often become my next post.
- Only wins and lessons: a loss is only wasted if you learn nothing from it.
Skipping is a skill too. It's how you keep your money.
Your turn: what are you most tempted to buy right now? Tell me in the comments.
Monk's soul, predator's focus.
MT Lab Research
Experience, not financial advice.
Nvidia's board says NO to Huang's return. Three of its own directors reportedly say YES.
The company line is clear: on the Nov 17 earnings call, chair Harvey Jones said directors unanimously accepted Huang's resignation and do not intend to reinstate him. But Semafor reports 3 directors are pushing for his return anyway. That report rests on unnamed sources and is unconfirmed.
Quick catch-up: Nvidia's founder and CEO of 33 years resigned on Nov 12 after reports of an affair with a woman US officials suspect of ties to Chinese intelligence (China denies it). An outside investigation is running and CFO Colette Kress is interim CEO. Huang himself said he would be honored to come back.
Why a split matters even if it never reaches a vote? It makes the job harder to fill. Semafor says Arm's CEO Rene Haas, a name the board has reportedly considered, has voiced concern about taking the seat while some directors still push for Huang. Who wants to be hired by a divided board?
Who decides? The board, not a petition with 180,000+ signatures and not backers like Elon Musk or Trump. Per Al Jazeera, a majority of directors can remove a CEO, and the board had 11 seats before Huang left. By my math, 3 votes is far short of a majority. Huang owns about 3.5% (per Business Insider), so no control there either.
Same night, Nvidia reported revenue up 103% to $115.8B and shares rose about 5% after hours. An analyst said that helps the board's hand but settles nothing about who leads.
Still unknown: what the investigation found, who the next CEO is, and what Huang says at his House hearing on Dec 1.
- Sort every headline into "said on the record" and "said by sources". Only the first is a fact
- A split inside the room can matter before any vote happens
- Ask who holds the vote before you ask who is loudest
Founder back or clean break, where do you land?
Experience, not financial advice.
180,000 signed to bring Jensen Huang back. Nvidia's board isn't weighing it.
Quick catch-up: Huang, the man who built Nvidia, resigned as CEO on Nov 12. He admitted an affair with a Chinese businesswoman who is suspected of ties to Chinese intelligence (Beijing denies it). The investigations are still open.
Since then a "Stand With Jensen" petition took off. Over 180,000 signatures. Elon Musk signed. Trump said he wouldn't mind seeing Huang back. Thousands of people say they work at Nvidia.
Now the part that matters. Per Reuters, three people familiar with the matter say the board is not considering bringing him back.
Why would a crowd that big not move a board? Because a petition is not a vote.
- The petition itself warns that job titles and shareholder claims are self-reported and unverified. Anyone can type "shareholder".
- The people who decide are the directors, plus the big funds that can lean on them. One large investor told Reuters it would be wary of a return while the investigations are open.
- Huang owns only about 3.6% of Nvidia, and Nvidia has no special founder shares that lock in control (Reuters, citing LSEG data). He is not the one holding the votes.
- Per Reuters, Nvidia has lost about $800 billion in market value since the news broke, and shareholders have already sued. A board in that seat does not take its cues from a comment section.
None of this says he never returns. If the investigations clear him and the big holders change their minds, the board can change its mind too. If worse facts land, the door shuts harder. Nobody knows yet, including the people on the petition.
Next time a loud crowd shows up in a headline, ask one thing first: who actually has the vote?
Experience, not financial advice.
The scariest word in the Fed minutes isn't "hike". It's "couple".
Bitcoin is stuck near $63K, and the minutes came out on Nov 18 at 19:00 UTC. Here is the decoder, because the Fed counts officials with words, from smallest to biggest: a couple, a few, several, some, many, most, almost all.
"Almost all" officials said more rate hikes will likely be needed to bring inflation back to 2%. But look at the smallest word. A couple of them said a hike bigger than 25 basis points (a quarter of one percentage point) could be considered in December if inflation data come in bad. The same day, Fed Governor Christopher Waller told CNBC the Fed may need a 50 basis point hike on Dec 9.
Why the small word matters more? Traders already give about 97% to a 25 basis point hike (CME FedWatch). A move everyone expects cannot surprise anyone. The gap between 25 and 50 is the part nobody has priced. By my math, that is one extra quarter point.
Meanwhile the dollar index hit a one-year high of 103.91. A stronger dollar and higher rates tend to hit assets that pay no yield, like bitcoin, and crypto's Fear and Greed Index sat at 11, extreme fear, for a sixth day.
Some scenarios we can have here:
- If Dec 9 brings 25 and no promise of more, that is the expected case, and the hike alone is not the shock.
- If it brings 50, that is a surprise, and surprises are what reprice markets.
- If it brings 25 but a signal of another in January, the market gets a faster pace without the headline number (Waller floated this too, per CNBC).
What I learnt from this:
- Learn the wording ladder. It shows which case is gaining support before the vote does.
- Priced in means the number is known. The size of the surprise is not.
- Nobody knows which case wins, and a hawkish Fed does not set a price.
Hike or surprise, which one worries you more?
Experience, not financial advice.
Radar #9 update: Anthropic's IPO price went up. For Binance Wallet buyers, that's not all good news...
Anthropic, the AI company behind Claude, priced its IPO at $83 a share, above its $70-78 range. It also sold more shares than planned: 980 million instead of 825 million. That raised $81.3B, the biggest IPO on record, above SpaceX's $75B (per Bloomberg and CNN). Nasdaq lists it as ANT on Nov 19.
Great for Anthropic. Here is what it means if you subscribed to the tokenized version, ANTx, on Binance Wallet:
- Your break-even moved up. $83 plus the 5% fee is $87.15 per token. At the old $78 top it was $81.90.
- The cushion shrank. Before pricing, crypto traders on Hyperliquid valued one share near $96.61 (per Bitquery). By my math that is about 11% above the new break-even, down from about 18% at $78.
- The pool filled in hours. The 500M USDC cap was hit early on Nov 18, so expect a small fill. A small fill keeps both the gain and the loss small.
- You still can't sell at the open. Tokens come at least 2 hours after Nasdaq opens at 14:30 UTC, so 16:30 UTC at the earliest, and big IPOs often start trading later.
Why price higher and sell more? Demand. Orders were close to 7x the shares on offer. When demand is that strong, the company keeps more of the first-day gain for itself. What Anthropic collected at $83 is partly what early buyers won't see at the open.
What I plan to prepare:
- If ANT opens and holds above $87.15, ANTx subscribers are in profit after the fee
- If it opens strong and fades before the tokens arrive, the gain can shrink to nothing
- If it opens below $87.15, the fee turns a flat debut into a small loss
Radar #9 was a bet on capped downside, not a guaranteed pop. $83 trims the upside. The math still works only above $87.15.
Above or below $87.15, where do you think ANT opens?
Experience, not financial advice.
I will join Anthropic IPO on Binance Wallet! Can I make money with it?
Radar #9, full guide in this article
Experience, not financial advice.
🏦 Russia's biggest private bank just got a crypto green light. Look who was already in.
Alfa-Bank was added to Russia's official list of firms allowed to trade crypto with customers (per TokenPost, Nov 18). Sounds huge, right? Here is the plot twist: Sberbank and VTB got on that list back in October, and 17 firms were already on it by Oct 22. Alfa is joining a crowded room, not opening a new door.
Being on the list is like passing the driving test. It does not mean the car is in your driveway yet. Nobody has said:
- When Alfa customers can actually buy?
- Which coins?
- Whether Alfa can also store coins for customers (that needs a second license)?
What the rules say for normal people in Russia?
- Up to about $3,800 a year per bank, after passing a test
- Only the most liquid coins
- Paying for groceries with crypto is still banned
A simple habit for any "bank gets approved" news. Ask:
- Which list is it?
- What does it allow?
- When does it launch?
Would you keep your coins at a bank?
Experience, not financial advice.
A click you made months ago can still empty your crypto wallet!
On Nov 16, Canada and the US launched Operation Niagara against approval phishing: scams that trick you into "approving" a scammer's access to your tokens.
It's led by the Ontario Securities Commission, with Canadian police, the US Secret Service and blockchain analytics firms like Chainalysis, TRM Labs and Elliptic.
How the scam works, in plain words:
> A fake pop-up that looks like it comes from a trusted app asks you to "approve" something. It feels like a login or a claim.
> In reality you just gave a scammer's address permission to move your tokens. No password stolen, no hack.
> You signed it, so the chain treats the theft as allowed. And once the money moves, it generally can't be reversed.
What's new, per the reports? The operation tries to spot attacks in near real time, even when access was stolen but not used yet, and to reach victims before the money leaves.
What we don't know? The OSC hasn't shared how many victims, scam networks or dollars are involved. So treat it as a start, not a result. Its March predecessor, Operation Atlantic, reported over $45M in crypto tied to fraud, $12M frozen and 120+ scam domains shut down.
What it teaches?
- Police can warn you after the signature. Only you decide before it.
- Read what a wallet prompt gives away, and to whom. If you can't tell, don't sign.
- Check your open approvals with a revoke tool (revoke(.)cash is a common one) and cancel the old ones you don't use.
- Keep long-term coins in a wallet that never signs approvals.
- "Officials" who contact you first? Verify through the agency's own official number, never a link in a message.
When did you last check your approvals?
Experience, not financial advice.
Nvidia has the chips. Does the world have the power, memory and money?
Its data center sales (AI chips and the gear around them) hit $107.6B in the quarter ending Oct 25, up 110% from a year ago. It guided to $130B in total sales next quarter (one extra week included) and has pointed to roughly 70-75% growth next year.
So can it keep going? Demand is not the shaky part. In August Nvidia itself said supply, not customers, was the cap on its growth. What decides the next quarters happens after the chip ships. Nvidia books the sale at shipping, but customers only reorder if the last batch is up and earning.
Three things have to hold:
1. Power. AI racks draw far more electricity than old servers and grids connect slowly. Gartner expects 40% of AI data centers to be power-limited by 2027, and one Seaport analyst (via Barron's) guesses nearly half of the capacity announced for 2027 won't get built. Power on time means orders become running machines. Late means chips wait.
2. Memory. Each AI chip needs scarce, pricey memory. Nvidia guided its gross margin (what it keeps from each sales dollar after costs) to 71-72% next quarter, down from about 75%, and blamed memory costs in August. If memory loosens, growth can run. If not, it caps the climb.
3. Money. Microsoft, Alphabet, Amazon and Meta spent a combined $201.9B on data centers and equipment in the Sept quarter, about double a year ago (Walnut's tally of their filings), and more of it is now paid with borrowed money. If lenders keep funding, orders hold. If funding tightens, the next order slips first.
The quarter ended before this month's CEO change, so that is not in these numbers either.
Next time an Nvidia headline hits, ask:
- Is the power there?
- Is the memory there?
- Who is paying?
Sold is not switched on.
Which one slows it first, power, memory or money?
Experience, not financial advice.
NVIDIA Earnings: Sales doubled to $115.8B. The stock rose 2-4%. Why so small?
The market grades a beat against what it already expected, and Nvidia also guided its profit margin down for next quarter. Here is the part the headline skipped.
The numbers: Nvidia makes the chips behind most AI. On Nov 17 it reported $115.8B in sales for the quarter ending in October. That is up 103% from a year ago and about 6% above what Wall Street analysts expected. Its own forecast (called guidance) for next quarter is about $130B, also above analyst forecasts. A clean beat on both.
What the headline skipped?
- The $150B buyback (the company buying its own shares) is not new. Nvidia's release says it was approved in September.
- Margin is shrinking. Gross margin is the share of each sale left after the cost of making the product. It was about 75% this quarter. Nvidia guides 71-72% next quarter, and operating costs rise to about $11B from $9.2B. More sales, but each dollar keeps less.
- This is the first report since CEO Jensen Huang resigned last week, with the CFO running the company for now. Nobody knows exactly why a stock moves, but this is a likely list.
Crypto cares too. It has been trading like a tech stock: The Block reported about $800M of crypto positions liquidated on Nov 13 as Nvidia turmoil and Fed worries hit.
Things I am watching NVIDIA:
- If sales keep beating while margins keep slipping, the market may keep shrugging at beats.
- If margins steady and the CEO question gets answered, the same beat could land harder.
- A 2-4% move after hours is a first reaction, not a verdict. It can flip by the open.
When doubling your sales barely moves your stock, expectations have gotten expensive.
Beat or not, which number do you trust more, sales or margin?
Experience, not financial advice.
I will join Anthropic IPO on Binance Wallet! Can I make money with it?
Radar #9, full guide in this article
Experience, not financial advice.
Anthropic's IPO got more expensive before anyone could buy a share?
The company behind Claude marketed its shares at $70-78. Now The Wall Street Journal reports it plans to price above $80 after huge orders from hedge funds. That would value it at about $1.5-1.6 trillion, vs about $1.4 trillion at the top of the old range. Nothing is final until it prices, expected Nov 18, with the first trade on Nov 19.
Who is asking? Altimeter's founder Brad Gerstner says his firm requested $5 billion of stock. Reports put total orders at roughly 6-7 times the shares on offer.
Two prices, in plain words:
- The offer price is what investors who get an allocation pay before trading starts.
- The open price is what everyone else pays once shares trade on Nasdaq (ticker ANT).
- If the stock opens above the offer, that first gain went to the people who got shares at the offer.
Everyday buyers can get up to 20% of the offering, about double the usual share. But with demand this big, many will get a small allocation or none.
Why a hot IPO is not a promise?
- Cerebras, an AI chip company, priced at $185 in May and closed its first day up 68%. Five months later it traded below $185.
- A higher offer price means less room for the next buyer to be right.
- Steve Eisman, the Big Short investor, says the bigger risk is AI infrastructure spending slowing, not this IPO.
By my math, even $80 is only about 3% above the top of the old range. The size of the demand says more than the price, it shows how much good news is already in it.
What I can say:
- If demand stays this strong after the open, later buyers pay the pop price, not the offer.
- If AI spending slows, a $1.5 trillion value has less cushion.
- If you cannot get shares at the offer, buying at the open is a different bet with different math.
Which price do you trust more, the offer or the open?
Experience, not financial advice.
Anthropic's IPO has plenty of ANT-icipation, but still no price. 🐜
Here is why! "Approved to list" is one box on a longer checklist. Anthropic's updated SEC filing on Nov 17 says Nasdaq approved its shares under the ticker ANT, its stock symbol. The same filing says the registration is not yet effective, so the SEC has not given its final go-ahead to sell shares, and there is still no final price.
An IPO is when a company sells its shares to the public for the first time. Here is where this one stands, per the filing and press reports:
✅ Done: paperwork filed, price range set at $70-78 a share, Nasdaq approval
⌛ Not done: the SEC go-ahead, the final price, the first trade
📅 Reported timing: price set around Nov 18, first trade around Nov 19 (neither Nasdaq nor the SEC has confirmed)
💰 Size: 825 million shares, so about $64B at the top of the range by my math
Why everyone is shouting? Bloomberg and Reuters say orders hit about 6x the shares on offer. That is the FOMO fuel. But "oversubscribed" only means more people asked than there are shares, so most buyers get cut back. Per Aiera, Klarna and StubHub were also heavily oversubscribed and still traded below their offer price later. Demand at the offer price is not the same as demand once everyone can trade.
One more thing newbies miss: the scary $80.7B loss in the first half of 2026. Per the WSJ and Reuters, about $72B of it is paper accounting on convertible notes (loans that can turn into shares), not cash burned. Revenue for the same half was $16.3B.
Here are some scenarios to think about it:
- If the final price lands above $78, demand beat even the top of the range
- If it lands at $70 or below, the room was cooler than the headlines
- If you can only buy after the open, you are not paying the IPO price. You are paying whatever the crowd pays that minute
Do you plan to buy after $ANT IPO?
Experience, not financial advice.
NVIDIA Earnings: Sales doubled to $115.8B. The stock rose 2-4%. Why so small?
The market grades a beat against what it already expected, and Nvidia also guided its profit margin down for next quarter. Here is the part the headline skipped.
The numbers: Nvidia makes the chips behind most AI. On Nov 17 it reported $115.8B in sales for the quarter ending in October. That is up 103% from a year ago and about 6% above what Wall Street analysts expected. Its own forecast (called guidance) for next quarter is about $130B, also above analyst forecasts. A clean beat on both.
What the headline skipped?
- The $150B buyback (the company buying its own shares) is not new. Nvidia's release says it was approved in September.
- Margin is shrinking. Gross margin is the share of each sale left after the cost of making the product. It was about 75% this quarter. Nvidia guides 71-72% next quarter, and operating costs rise to about $11B from $9.2B. More sales, but each dollar keeps less.
- This is the first report since CEO Jensen Huang resigned last week, with the CFO running the company for now. Nobody knows exactly why a stock moves, but this is a likely list.
Crypto cares too. It has been trading like a tech stock: The Block reported about $800M of crypto positions liquidated on Nov 13 as Nvidia turmoil and Fed worries hit.
Things I am watching NVIDIA:
- If sales keep beating while margins keep slipping, the market may keep shrugging at beats.
- If margins steady and the CEO question gets answered, the same beat could land harder.
- A 2-4% move after hours is a first reaction, not a verdict. It can flip by the open.
When doubling your sales barely moves your stock, expectations have gotten expensive.
Beat or not, which number do you trust more, sales or margin?
Experience, not financial advice.
NVIDIA Earnings Today: Great Chips, Great Sales, Great Expectations. What Could Go Wrong?
The report lands at about 21:20 UTC, its first since founder Jensen Huang quit. Analysts expect about $109B in sales. Nvidia itself guided to $108B. The real questions are elsewhere.
Quick setup. Huang resigned as CEO on Nov 12 and the release gave no reason. CFO Colette Kress is interim CEO. Analysts also expect $2.47 profit per share (LSEG, per CNBC).
Why a near-double can still disappoint? It is priced in. The market already expects it, like a straight-A student getting another A. In August Nvidia beat by about $4B and the stock still slipped after hours at first (Invezz). What moves it is what is new.
Three things CNBC says Kress will face:
1. China. Nvidia's August outlook assumed no China data center chip sales. Short seller Culper Research alleges about 20% of last fiscal year's compute revenue was China-driven. That is an allegation, not a finding.
2. Neoclouds. These are new AI cloud firms, like CoreWeave, that rent out Nvidia chips. Nvidia has put money into some of them, and they spend it on more Nvidia chips. Critics call it circular financing: if the boom slows, both sides hurt.
3. Leadership. Shares fell 4.3% on Nov 9, about $190B of value, on reports of a probe into Huang's ties (MarketPulse). The CEO search has no timeline, and until it ends Kress runs both the company and the numbers.
What to prepare?
- If the outlook for next quarter is strong and the answers are clear, the numbers can carry the day.
- If sales beat but China and neocloud answers are vague, a beat may not be enough.
- If the leadership answers are thin, expect the questions to follow the stock, not the sales.
Earnings show the past. These questions are about what comes next.
China, neoclouds or the CEO search, which one are you watching?
Experience, not financial advice.
South Korea's stock market has lost almost half its value in 5 months!
On Nov 17 its main index, the KOSPI, closed at 4,968, down 2.59%. It is the first close under 5,000 since Feb 2 and the fourth drop in a row. The record close was 9,114 on Jun 22, so the index is down about 45% (Seoul Economic Daily).
Here is what is pressing on it:
- Foreign investors sold about 1.4 trillion won of shares in one day, roughly $1.1B by my math
- Doubts about AI chip demand, with Samsung down 3.3% and SK hynix down about 2%
- US 10-year yields back above 5%, which pulls money out of riskier markets
- Nvidia's founder-CEO Jensen Huang resigned on Nov 12, and Nvidia reports earnings at about 21:20 UTC on Nov 17, after Seoul closed
Why it is more than a bad day? On paper the market looks cheap: about 3.5 times expected earnings next year (the "forward P/E", price divided by profit forecasts). But that number only holds if the forecasts hold.
The usual dip buyers are running low too. Korean retail cash waiting on the sidelines is about 87 trillion won, down from the 130 trillion range at the peak. And Samsung and SK hynix were about half the index's value in July (Forklog), so two stocks can drag everything.
A 45% drop sounds like a discount. By my math, getting back to the June record needs a gain of about 83%.
What scenarios can we expect?
- If foreign selling slows and US yields ease, the bleeding can slow
- If yields keep rising or AI chip news stays shaky, the pressure stays
- Nobody knows which. Brokers that saw 10,000 in June now mostly target about 8,000 for next year
Three habits I practice from this:
1. Cheap is a forecast, not a floor
2. Ask who is buying the dip, and how much cash they have left
3. When two stocks are half an index, it is really one big bet
Foreign selling, US yields or Nvidia, which one would you watch first?
Experience, not financial advice.
Anthropic's IPO has plenty of ANT-icipation, but still no price. 🐜
Here is why! "Approved to list" is one box on a longer checklist. Anthropic's updated SEC filing on Nov 17 says Nasdaq approved its shares under the ticker ANT, its stock symbol. The same filing says the registration is not yet effective, so the SEC has not given its final go-ahead to sell shares, and there is still no final price.
An IPO is when a company sells its shares to the public for the first time. Here is where this one stands, per the filing and press reports:
✅ Done: paperwork filed, price range set at $70-78 a share, Nasdaq approval
⌛ Not done: the SEC go-ahead, the final price, the first trade
📅 Reported timing: price set around Nov 18, first trade around Nov 19 (neither Nasdaq nor the SEC has confirmed)
💰 Size: 825 million shares, so about $64B at the top of the range by my math
Why everyone is shouting? Bloomberg and Reuters say orders hit about 6x the shares on offer. That is the FOMO fuel. But "oversubscribed" only means more people asked than there are shares, so most buyers get cut back. Per Aiera, Klarna and StubHub were also heavily oversubscribed and still traded below their offer price later. Demand at the offer price is not the same as demand once everyone can trade.
One more thing newbies miss: the scary $80.7B loss in the first half of 2026. Per the WSJ and Reuters, about $72B of it is paper accounting on convertible notes (loans that can turn into shares), not cash burned. Revenue for the same half was $16.3B.
Here are some scenarios to think about it:
- If the final price lands above $78, demand beat even the top of the range
- If it lands at $70 or below, the room was cooler than the headlines
- If you can only buy after the open, you are not paying the IPO price. You are paying whatever the crowd pays that minute
Do you plan to buy after $ANT IPO?
Experience, not financial advice.
NVIDIA Earnings Today: Great Chips, Great Sales, Great Expectations. What Could Go Wrong?
The report lands at about 21:20 UTC, its first since founder Jensen Huang quit. Analysts expect about $109B in sales. Nvidia itself guided to $108B. The real questions are elsewhere.
Quick setup. Huang resigned as CEO on Nov 12 and the release gave no reason. CFO Colette Kress is interim CEO. Analysts also expect $2.47 profit per share (LSEG, per CNBC).
Why a near-double can still disappoint? It is priced in. The market already expects it, like a straight-A student getting another A. In August Nvidia beat by about $4B and the stock still slipped after hours at first (Invezz). What moves it is what is new.
Three things CNBC says Kress will face:
1. China. Nvidia's August outlook assumed no China data center chip sales. Short seller Culper Research alleges about 20% of last fiscal year's compute revenue was China-driven. That is an allegation, not a finding.
2. Neoclouds. These are new AI cloud firms, like CoreWeave, that rent out Nvidia chips. Nvidia has put money into some of them, and they spend it on more Nvidia chips. Critics call it circular financing: if the boom slows, both sides hurt.
3. Leadership. Shares fell 4.3% on Nov 9, about $190B of value, on reports of a probe into Huang's ties (MarketPulse). The CEO search has no timeline, and until it ends Kress runs both the company and the numbers.
What to prepare?
- If the outlook for next quarter is strong and the answers are clear, the numbers can carry the day.
- If sales beat but China and neocloud answers are vague, a beat may not be enough.
- If the leadership answers are thin, expect the questions to follow the stock, not the sales.
Earnings show the past. These questions are about what comes next.
China, neoclouds or the CEO search, which one are you watching?
Experience, not financial advice.
Goldman holds $1.28B of Bitcoin ETFs. Goldman also sold Bitcoin ETFs this year. Both are true!
Here is how. In filings called 13Fs (quarterly holdings reports big money managers send the SEC), Goldman Sachs showed about $1.28B and Millennium, a giant hedge fund, about $1.21B in spot Bitcoin ETFs on Sept 30.
Galaxy Research also named both among 9 firms behind almost all institutional ETF selling from Sept 2025 to June 2026 (about 68,000 BTC, roughly $5.4B).
How can a seller hold? Both cut in Q2 and rebuilt in Q3, per BeInCrypto. And a holding is not always a bet. Galaxy blames mostly a basis trade closing: buy the ETF, sell a futures contract priced higher, keep the gap. The gap shrank from about 7.9% to 4.8% a year, barely above safe Treasury bills, so desks left. Its point: arithmetic, not a loss of faith in Bitcoin. A 13F cannot show which shares are hedged.
Now the part that protects you: the date. Sept 30 is about 7 weeks old. Bitcoin was near $83,700 then and about $64,500 now (it slid toward $62,000 on Nov 15). By my math, the same shares would now be worth about 23% less: roughly $0.99B for Goldman and $0.93B for Millennium.
In the past two weeks about $2.35B left Bitcoin ETFs (SoSoValue), and that data cannot say whose it was. From public data, nobody can tell whether these two held, added or sold. The next filings, due Feb 16, will show.
What can we prepare?
- If both still show big holdings after this drop, that says more than the Sept 30 headline did. It still would not show hedges.
- If they show big cuts, the $1.2B headline told you nothing about today.
- If all you have is the headline, assume you know the past, not the present.
What I learnt from this?
- A big name's holding is a photo. Selling is a movie.
- Check the date before the dollar amount.
- Ask who is on both lists before you follow a name.
Photo or movie, which one do you need more?
Experience, not financial advice.
Berkshire bought $689M of one stock. Should you copy? Check one date first.
The date is Sept 30. Berkshire's latest 13F, filed Nov 16, shows one new buy: Bending Spoons, about $689M as of Sept 30. That snapshot is 47 days old. Then Berkshire bought about $330M more in October. The filing cannot show that. The picture had already moved.
A 13F is a report that big money managers (over $100M in US-listed holdings) must file with the SEC about 45 days after each quarter ends. It lists what they held on the last day of the quarter. Nothing more.
So it is a great place to get curious and a risky place to copy from. Here is what it can and cannot tell you:
- It shows longs only. The SEC says shorts are not reported and are not subtracted from longs. A big "buy" can be fully hedged and you would never know.
- Look at share counts, not dollars. IBIT rose about 41% in Q3, so a bigger dollar stake can just mean a higher price (Cointelegraph).
- A trim is not always doubt. Berkshire cut Bank of America by 5.7%, about 483M shares to 456M (24/7 Wall St). The filing never says why, and a trim can be plain risk control.
Where 13Fs get useful is the crypto trend. Among 1,389 early filers, banks held $258M of spot Bitcoin ETFs, 4.9% of the total, up from 3.2% in Q2 (Bitwise, preliminary). Barclays lifted its IBIT stake 48%. Citi opened a $119M FBTC position. That is a slow shift across many names, a better signal than one stock. Still, it covers Sept 30, before Bitcoin's later drop and about $2.1B of Q4 ETF outflows (SoSoValue).
If you use a 13F, here are some valuable lessons:
- If many unrelated funds add the same name over several quarters, study why.
- If one fund shows a new stake, that is one view, and you do not know its entry price.
- If the price has moved a lot since Sept 30, the filing no longer says what they would do today.
Smart money leaves footprints. Footprints are not directions.
Experience, not financial advice.
180,000 signed to bring Jensen Huang back. Nvidia's board isn't weighing it.
Quick catch-up: Huang, the man who built Nvidia, resigned as CEO on Nov 12. He admitted an affair with a Chinese businesswoman who is suspected of ties to Chinese intelligence (Beijing denies it). The investigations are still open.
Since then a "Stand With Jensen" petition took off. Over 180,000 signatures. Elon Musk signed. Trump said he wouldn't mind seeing Huang back. Thousands of people say they work at Nvidia.
Now the part that matters. Per Reuters, three people familiar with the matter say the board is not considering bringing him back.
Why would a crowd that big not move a board? Because a petition is not a vote.
- The petition itself warns that job titles and shareholder claims are self-reported and unverified. Anyone can type "shareholder".
- The people who decide are the directors, plus the big funds that can lean on them. One large investor told Reuters it would be wary of a return while the investigations are open.
- Huang owns only about 3.6% of Nvidia, and Nvidia has no special founder shares that lock in control (Reuters, citing LSEG data). He is not the one holding the votes.
- Per Reuters, Nvidia has lost about $800 billion in market value since the news broke, and shareholders have already sued. A board in that seat does not take its cues from a comment section.
None of this says he never returns. If the investigations clear him and the big holders change their minds, the board can change its mind too. If worse facts land, the door shuts harder. Nobody knows yet, including the people on the petition.
Next time a loud crowd shows up in a headline, ask one thing first: who actually has the vote?
Experience, not financial advice.
The wildest thing at Nvidia this weekend wasn't its stock. It was the CEO's empty chair!
The stock barely moved. Nvidia stock futures were little changed on Sunday evening (Yahoo Finance). The chair is where the fight is. Silicon Valley wants Huang back. Washington wants an outsider. The board wants quiet. Nothing is decided.
Three camps want three different things:
1. Huang's allies: investors Altimeter and Atreides published an open letter asking the board to keep all options open, including his return. Elon Musk backs it, and allies talk of a "war room" to lobby the board and the White House (per the FT)
2. Washington: Treasury Secretary Bessent said on Fox the board should hire an outsider, and wants answers from the board and its outside counsel within 10 days on chip export compliance
3. The board: it calls guessing about candidates premature. It reportedly approached Arm's CEO René Haas, and CFO Colette Kress is interim CEO
Huang resigned on Nov 12 after reports about his personal conduct, and an independent inquiry is still running. Nvidia has not publicly answered the return talk.
The one price that did move was crypto's. With Nasdaq closed, Nvidia perps on Hyperliquid (crypto contracts that track the stock) rose up to 4.5% after the FT said Huang is discussing a comeback. A perp is a derivative that trades 24/7, so it can move first on a report and be wrong.
The next dated test is earnings on Nov 17 at about 21:20 UTC, the first report without him.
What I leanrt from this:
- "Discussing" is not "decided". Only the board decides
- A weekend price can follow a report before anyone with a vote has spoken
- Three camps pulling on one seat is a story, not an outcome
Who do you think should run Nvidia?
Experience, not financial advice.
Bought XRP the day the first US spot ETF launched? You're down about 59%!
XRP touched $0.89 on Nov 14, 2026, its lowest since Nov 2024. Nearly every headline holders waited for came true, yet the price gave most of it back. Dates in UTC:
> Nov 2024: Trump wins the election on Nov 5; on Nov 21 SEC chair Gary Gensler says he will leave. XRP goes from $0.49 on Nov 3 to $2.91 on Dec 3.
> Jul 18, 2025: the peak, $3.66, as the GENIUS Act, the first US stablecoin law, is signed.
> Aug 7: the SEC and Ripple end their case.
> Oct 10, 2025: after Trump threatens 100% tariffs on China, about $19B of leveraged bets are liquidated (force-closed by the exchange, adding selling). XRP wicks from $2.84 to $1.25.
> Nov 13, 2025: the first US spot XRP ETF launches with about $250M of inflows, and XRP closes 7% below its open. The fund takes coins in-kind: holders deposit them instead of buying on exchanges.
> Aug 19 to 21, 2026: the Treasury doubles its bond buybacks, yields fall, and shorts get squeezed. A short is a bet on a falling price, so a rise forces shorts to buy back, pushing it higher. Over $3.3B of shorts closed and XRP rose 45%, from $1.01 to $1.45.
> Sep 15, 2026: the Senate vote to advance the CLARITY Act (crypto market rules) fails 49 to 50, 60 needed. XRP drops about 10%.
> Nov 4 to 14, 2026: $1.34 slides to $0.89. On Nov 14, 94% of liquidated positions were longs (bets on a rising price), so the drop forced more selling. XRP ETFs lost $19.13M in the week of Nov 9 to 13, the most since March.
Everyone had a reason:
- Headline buyers expected laws and ETFs to bring a flood of new money.
- Leveraged longs borrowed to amplify gains, so falling prices closed them out.
- August shorts bet on drift, then had to buy back higher.
What this chart teaches me:
- Good news does not buy coins. New money does.
- Leverage speeds the climb and the fall.
- Ask who is forced to act: longs, shorts, ETF deposits.
Is XRP good to buy now?
Experience, not financial advice.
Bitcoin's six red weeks: buy the dip, become the dip. $1.31B of longs did it in one day.
$BTC is down 24.5% since Mon Oct 5, the longest weekly losing run since end-February. Week six alone lost 12.7%. The headline changed every week. The sellers did not. ETF holders redeeming and leveraged longs liquidated (forced out when margin runs out).
> Oct 7: Oil above $100, the 10-year yield above 5.3%. US spot ETFs (funds holding real BTC) pulled $485M, the biggest outflow since June. BTC fell from $87k to $80.4k.
> Oct 22: ETFs had added $1.08B in three sessions and BTC hit $84.9k. Then Tesla and Meta earnings revived AI doubts, ETFs pulled $635M in a day and breakout longs were flushed to $80.9k.
> Oct 29: The Fed held rates with two votes for a hike. Binance then said attackers drained about $6.3B through a compromised engineer laptop. It froze about $3B and promised to cover all losses. BTC slid from $80.8k to $76.3k, $1.54B liquidated, mostly longs.
> Nov 5: A chip-stock rout (semiconductor index -12.2%) took BTC from $78.3k to $75.6k in one hour. $1.03B liquidated, 89% longs. ETFs lost $1.28B that week.
> Nov 8: A Sunday flash from $75.7k to $73.5k in 15 minutes: Nvidia's CEO probe and stalled Iran talks hit a thin weekend market.
> Nov 12-15: Nvidia's CEO resigned, hot inflation lifted December hike odds to 88.5%, ETFs sold seven sessions in a row ($1.91B). Below $70k, $1.31B of longs went in a day. Below $64k, $671M of longs went in one hour. Low: $62k.
Everyone had a different reason for their own action:
- ETF holders: cut exposure as price fell
- Leveraged longs: bought dips, stops stacked under $82k and $70k
- The Fed: inflation first, two officials wanted a hike
- Binance: needed users to stay, so it paid
What six red weeks taught:
- A bounce is not a trend until buyers keep paying
- News picks the day, positioning picks the size
- After a flush, funding flips negative: shorts now pay longs
Will we have seven red weeks?
Experience, not financial advice.
3 giant banks, 1 yen stablecoin, and a $300 billion elephant in the room.
The elephant is the dollar. Dollar coins are about 99.5% of a market above $300 billion (per DefiLlama and CryptoQuant). All yen coins together: about $180 million (per DIA). By my math, under 0.1%. A stablecoin is a token that stays worth a fixed amount, here 1 yen.
On June 10, 2026, MUFG, SMBC and Mizuho said they will issue one coin together, with live business use targeted by the end of March 2027 (per The Block). Today that is an agreement and a council, not a launched coin.
How it would work, per their joint statement? The three banks set up a trust, and a separate trust bank holds the reserve. One shared coin instead of three rival ones.
Who it is for? Companies, not you. The three banks serve more than 300,000 business clients (per Nikkei, via The Block). The first test run is Mitsubishi Corp moving money between its Japanese and overseas offices, with support from Japan's financial regulator.
What to take note from it:
- Planned is not live: nothing is issued yet, and the target is March 2027
- The banks bring customers, not coin supply: the market stays small until companies really use it
- Still unknown: issue size, blockchain and retail access, none given in the official statement (per crypto(.)news)
Would you hold a stablecoin issued by a bank?
Experience, not financial advice.
The wildest thing at Nvidia this weekend wasn't its stock. It was the CEO's empty chair!
The stock barely moved. Nvidia stock futures were little changed on Sunday evening (Yahoo Finance). The chair is where the fight is. Silicon Valley wants Huang back. Washington wants an outsider. The board wants quiet. Nothing is decided.
Three camps want three different things:
1. Huang's allies: investors Altimeter and Atreides published an open letter asking the board to keep all options open, including his return. Elon Musk backs it, and allies talk of a "war room" to lobby the board and the White House (per the FT)
2. Washington: Treasury Secretary Bessent said on Fox the board should hire an outsider, and wants answers from the board and its outside counsel within 10 days on chip export compliance
3. The board: it calls guessing about candidates premature. It reportedly approached Arm's CEO René Haas, and CFO Colette Kress is interim CEO
Huang resigned on Nov 12 after reports about his personal conduct, and an independent inquiry is still running. Nvidia has not publicly answered the return talk.
The one price that did move was crypto's. With Nasdaq closed, Nvidia perps on Hyperliquid (crypto contracts that track the stock) rose up to 4.5% after the FT said Huang is discussing a comeback. A perp is a derivative that trades 24/7, so it can move first on a report and be wrong.
The next dated test is earnings on Nov 17 at about 21:20 UTC, the first report without him.
What I leanrt from this:
- "Discussing" is not "decided". Only the board decides
- A weekend price can follow a report before anyone with a vote has spoken
- Three camps pulling on one seat is a story, not an outcome
Who do you think should run Nvidia?
Experience, not financial advice.
Anthropic's IPO gold rush has a twist: the shovel sellers are buying in!
Everyone calls this IPO a test of investor appetite for AI. Before you read it as "the whole market wants in", look at the order book: $26B of it is pre-sold to suppliers.
Per CNBC and Yahoo Finance, seven suppliers (Google, Nvidia, AMD, Broadcom, Micron, Microsoft, TSMC) will buy $26B of special non-voting shares at the IPO price. Google and Nvidia take about $10B each at the midpoint. These are the companies that sell Anthropic its computing power, and the proceeds are meant partly for compute and infrastructure. That is a loop worth watching.
The rest, per the filing: 825M shares at $70-$78 on Nasdaq under $ANT, up to about $64B raised. Pricing is set for Nov 18, trading Nov 19. Bloomberg says orders are already several times the shares on offer. Nothing is final until it prices.
Then the valuation, which comes in two sizes. On the 13.09B shares the filing says will exist, the range is about $916B to $1.02T. Counting options and other securities (fully diluted), CNBC says up to about $1.4T. By my math, the $74 midpoint on the basic count is about $969B, close to the $965B private round in May. The $2T talk faded.
About its business, $33.2B revenue in 9 months, up from $2.5B a year earlier. Net loss $104B, but $94.3B of that is a non-cash accounting charge. The operating loss, what the business itself lost, was $1.26B.
What to watch?
- Priced above $78 and holds after day one: demand beat the range. Priced at the bottom or trades under the price: appetite had limits.
- Always check which valuation a headline uses. The two differ by almost $400B.
- A packed order book proves people want shares at this price. It does not prove AI pays for itself.
At what price would you buy Anthropic's stock?
Experience, not financial advice.




