The AI building boom is running on borrowed money 🏗️ S&P Global expects the six biggest AI spenders to put about $1.3 trillion into data centres and equipment in 2027. That is roughly the yearly economic output of the Netherlands. A big part of it is borrowed, and more than $450 billion of new debt has been issued this year alone. The rating agency still thinks leverage will level off over the next two years without rating changes. Amazon briefly crosses its downgrade threshold, but not for long enough to lose its rating, while Oracle has already been downgraded. The bigger worry is the debt nobody can see. The large companies often guarantee leases and loans for younger neocloud players, and S&P admits it does not have a full picture of these liabilities. Bond investors already want more for the risk. Hyperscaler credit spreads have widened, and the cost of insuring Meta debt reached a record after its results this week. The shares point the same way: over the past month $CRWV fell 36%, while Microsoft and Amazon were flat. I own CoreWeave, bought on 15 October at $63.06, and it is now about 11% lower at $55.82. I kept the position size sensible because the company runs on borrowed money, and this is the risk I accepted when I bought. As long as the big guarantors stay strong, I am holding. Are you buying AI shares on this dip, or staying away? 💚 I’m Gabor - I Don’t Work for Money, Money Works for Me💰 🔄eToro’s CopyTrader feature lets you automatically mirror my Strategy 🚀 Gabor Fogarasi @fogi70 eToro Pro Popular Investor 🏦 CISI Level 3 Investment Certification med.etoro.com/B22386_A123282… Sign up to eToro to Copy my trades! Your capital is at risk!

A red day almost everywhere 🔴 and AI spending is the reason. The Nasdaq 100 is down about 2%, $META more than 7% after spending more cash than it made. Only Netflix is green on my screen. I hold Meta and I am not selling. How do you handle such red days?

Stablecoins move $400 million an hour, and most of it 🌙 happens after the banks close A new study from $COIN and The Block found that stablecoins moved more than $400 million an hour on average in the third quarter. Over the 12 months to July, about 73% of the volume settled outside US banking hours, which means evenings, weekends and holidays. The other numbers are just as interesting. Sending a $200 remittance in USDC on Base costs about 86% less than through traditional channels. Some 293 million blockchain addresses now hold stablecoins, more than the combined cardholders of American Express and Discover, although one person can own many addresses. Seven of the ten largest US banks have already tested stablecoins or tokenized deposits, and the other three plan to. I have written before that round-the-clock trading is bad for investors, because being on all the time wears people down. Payments are a different thing. A worker sending money home on a Sunday should not wait until Monday, and a business should not have cash stuck in transit over a weekend. That is the real use case here, and it does not need anyone to watch charts at midnight. 👉 I saw this myself recently. After a dinner out with a friend, he offered to pay me back his share in crypto instead of a bank transfer. I was happy to accept it, and the money was in my wallet within seconds, late in the evening, with no bank involved. Would you use stablecoins to send money abroad? 💚 I’m Gabor - I Don’t Work for Money, Money Works for Me💰 🔄eToro’s CopyTrader feature lets you automatically mirror my Strategy 🚀 Gabor Fogarasi @fogi70 eToro Pro Popular Investor 🏦 CISI Level 3 Investment Certification med.etoro.com/B22386_A123282… Sign up to eToro to Copy my trades! Your capital is at risk!

Anthropic's revenue grew 15 times in a year 📈 but the cloud giants still collect the rent Yesterday I wrote about Anthropic's shrinking IPO window. Now an early draft of its prospectus has leaked to the Wall Street Journal, and for the first time we can see real numbers from a top AI lab. Revenue for the first half of 2026 came in at $16.3 billion, up from $1.1 billion a year earlier. Spending on computing power rose from $2.3 billion to $9 billion over the same period. So revenue is growing faster than costs, which is the good news. The rest is harder to like. Anthropic has signed $450 billion of cloud contracts over the next seven to ten years that it cannot cancel, and it warns total spending could reach $750 billion. The net loss for the half year was $80.7 billion, although $72 billion of that was a non-cash accounting charge on convertible notes, and roughly a third of the document is risk factors. Every dollar of those contracts lands with the companies that rent out the chips and data centres: Amazon, $GOOG, Microsoft and CoreWeave. Anthropic has to keep growing to pay its bills, while the landlords get paid either way. That is why I own Alphabet and CoreWeave rather than waiting for the IPO, and why I will read the final prospectus line by line before I put any money into Anthropic itself. Would you rather own the AI lab or the companies it pays? 💚 I’m Gabor - I Don’t Work for Money, Money Works for Me💰 🔄eToro’s CopyTrader feature lets you automatically mirror my Strategy 🚀 Gabor Fogarasi @fogi70 eToro Pro Popular Investor 🏦 CISI Level 3 Investment Certification med.etoro.com/B22386_A123282… Sign up to eToro to Copy my trades! Your capital is at risk!

Muse is off to a strong start, but Meta 💸 spent more cash than it made $META beat expectations last night with $64.6 billion of revenue, up 26% from a year ago, and guided the fourth quarter to $73 to $76 billion. Ads did the heavy lifting again, with the average price per ad up 14%. Muse also had a strong start, as more than 10 million people in the US and Canada set it up in its first six weeks and about a quarter of them use it every day, mostly for shopping. Profit came in at $3.83 a share, or $7.40 without the $9.89 billion teen safety settlement. The cash side looks very different. Meta spent $34.9 billion on data centers and equipment in the quarter, more than the $32.8 billion its business brought in, so free cash flow turned negative at minus $2.1 billion. In the same quarter last year it kept $10.6 billion. Zuckerberg also said Muse loses money for now because the computing costs are so high, and Meta is creating a new class of non-voting shares from January, which works like a two-for-one split and lets the company raise money without weakening his control. 👉 I hold Meta shares at an average price of $532.65, so I'm about 30% up at today's $691. The ad business still pays the AI bill comfortably, so I'm holding, but I want to see free cash flow turn positive again before I add. Would you rather Meta slow down its AI spending, or keep going while ads can carry it? 💚 I’m Gabor - I Don’t Work for Money, Money Works for Me💰 🔄eToro’s CopyTrader feature lets you automatically mirror my Strategy 🚀 Gabor Fogarasi @fogi70 eToro Pro Popular Investor 🏦 CISI Level 3 Investment Certification med.etoro.com/B22386_A123282… Sign up to eToro to Copy my trades! Your capital is at risk!

Should the Fed meet less often? 🗓️ Not exactly breaking news, more of a thought experiment 💭 The Federal Reserve sets interest rates eight times a year, on a schedule that dates back to 1981. In July, chair Kevin Warsh asked his colleagues to consider meeting less often, so more data can build up between decisions. Economists at Deutsche Bank now back the idea, and their main complaint is timing. The September PCE inflation figure, the Fed's own target measure, comes out after next week's decision. That is the third time this year it has missed a meeting, and CPI reports also tend to land too late or too close to a meeting to be used properly. Their proposal is six meetings a year. The gaps would be about two weeks longer, but every meeting would have two jobs reports, two CPI reports and two PCE reports to work with, and the data would be about 30% more recent. The Bank of England already cut from 12 meetings to eight in 2016 and nothing broke. For investors, fewer meetings means fewer days when the market can be surprised, although each surprise could be bigger. I would accept that. A rate decision built on fresher inflation data is less likely to be reversed a few weeks later, and there is less time wasted guessing the next move every six weeks. Would you prefer six Fed meetings a year or eight? 💚 I’m Gabor - I Don’t Work for Money, Money Works for Me💰 🔄eToro’s CopyTrader feature lets you automatically mirror my Strategy 🚀 Gabor Fogarasi @fogi70 eToro Pro Popular Investor 🏦 CISI Level 3 Investment Certification $SPX500 med.etoro.com/B22386_A123282… Sign up to eToro to Copy my trades! Your capital is at risk!

📢 BITCOIN Alert: JPMorgan expects Bitcoin to beat stocks and gold for a decade JPMorgan has put a number on $BTC for the first time. In its long-term outlook for 2027, the bank's asset management arm expects Bitcoin to return about 10% a year over the next 10 to 15 years. That is more than it expects from US large-cap stocks (7.5%), small caps (7.7%), property trusts (8.2%), bonds (5.6% to 6.6%) and gold (5.5%). Starting from roughly $83,000 in late September, 10% a year would take Bitcoin to about $215,000 in ten years. The bull case of 17% a year points to around $386,000. The bear case is a loss of 9.7% a year, so the range is very wide. The bank is still careful. It calls Bitcoin a speculative asset and a hedge against central banks creating more money, not a diversifier, and says the big price swings keep it out of many portfolios. Still, this comes from the bank whose CEO Jamie Dimon called Bitcoin a "fraud" in 2017, so the change in tone is big. Bitcoin trades around $84,500 today, so the market has not moved much on the report. For me, a ten-year forecast is a reason to hold a steady allocation and buy the dips, not to chase the price. 👉 I hold Bitcoin on eToro, bought in three steps at an average of about $78,100, and I am up around 8%. I keep buying the dips rather than chasing rallies. Do you think Bitcoin beats the S&P 500 over the next ten years? 💚 I’m Gabor - I Don’t Work for Money, Money Works for Me💰 🔄eToro’s CopyTrader feature lets you automatically mirror my Strategy 🚀 Gabor Fogarasi @fogi70 eToro Pro Popular Investor 🏦 CISI Level 3 Investment Certification med.etoro.com/B22386_A123282… Sign up to eToro to Copy my trades! Your capital is at risk!

Is $SMCI the most underrated name in the AI buildout? 🤖 Early September-quarter numbers beat guidance and the shares jumped 17%, yet they still sit about two-thirds below the March 2024 peak. Full results come on November 3. Do you trust Super Micro now?

Anthropic's IPO window is getting smaller ⏳ This was supposed to be the year of giant AI listings. SpaceX went first and raised a record $85.7bn. Anthropic, the company behind Claude, has now moved its own listing from September to November, which leaves a narrow gap between the US midterms on 3 November and Thanksgiving. Betting markets put the chances that it happens in November at about 60%. The numbers behind the deal have cooled since the summer. Revenue estimates now sit around $74bn a year, below the $100bn+ some of its investors talked about. Anthropic also offered Blackstone a 10% coupon on $9bn of new debt, part of a $60bn package to pay for chips, and it has signed up for at least $518bn of infrastructure spending over roughly ten years. That is why I compare it with $CRWV rather than Nvidia. Nvidia pays for its growth from its own cash flow. CoreWeave runs on borrowed money, and since mid-July its shares have moved about three times as much per day as Nvidia's, rising 38% to an August peak and then falling 40%. For early owners, a listing like this is mainly a chance to sell to the public. I bought CoreWeave on 15 October at $63.06 and keep the position small for exactly this reason. If Anthropic lists, I will read the prospectus and the debt section before I buy a single share. Would you buy the Anthropic IPO on day one? 💚 I’m Gabor - I Don’t Work for Money, Money Works for Me💰 🔄eToro’s CopyTrader feature lets you automatically mirror my Strategy 🚀 Gabor Fogarasi @fogi70 eToro Pro Popular Investor 🏦 CISI Level 3 Investment Certification med.etoro.com/B22386_A123282… Sign up to eToro to Copy my trades! Your capital is at risk!

Barclays says the CoreWeave scare 🧯 is not the end of the AI trade $CRWV lost about 24% last week as yields on some of its bonds jumped to around 15%, and the fear spread across the AI trade. Barclays said on Monday that investors are making the same mistake they made with Silicon Valley Bank in 2023, when trouble at a few regional banks was read as a crisis for the entire US economy. CoreWeave's unsecured bonds are a small part of its funding, and its spending is a small part of total AI investment. If CoreWeave stopped spending completely, total AI spending would drop by about 4% to 5%. If it halved its budget, the more realistic downside in Barclays' view, the cut would be around $20 billion out of more than $1 trillion expected in 2027. Most of the money comes from Alphabet, Microsoft, Meta and Amazon, and they don't finance their data centers the way CoreWeave does. All four say they still need more computing power than exists today. J.P. Morgan added today that most global banks have limited exposure to CoreWeave, which carries about $50 billion of debt. 👉 I bought CoreWeave on 15 October at $63.06, and at around $58 today I'm about 8% down. I kept the position at a sensible size because the company runs on borrowed money. I also hold Alphabet and Meta, so I own both the borrower and the big spenders. Do you think CoreWeave's debt problem stays a CoreWeave problem, or is it the first crack in the AI trade? 💚 I’m Gabor - I Don’t Work for Money, Money Works for Me💰 🔄eToro’s CopyTrader feature lets you automatically mirror my Strategy 🚀 Gabor Fogarasi @fogi70 eToro Pro Popular Investor 🏦 CISI Level 3 Investment Certification med.etoro.com/B22386_A123282… Sign up to eToro to Copy my trades! Your capital is at risk!