Sandeep Anand

🚨 $NET just posted its best quarter ever — and it’s the biggest beneficiary of AI !! 📈Revenue $782.5M, +39% YoY — fastest growth in nearly 4 years. EPS $0.38 vs $0.34. Net retention 122% RPO $2.877B (+34%) Q4 guide $845–846M vs $796M street. 🚀FY guide raised to $2.96B / $1.35 EPS — 3rd raise this year (wow… ) GAAP profitability pulled forward to Q4 (was end-2028) 🚨Why Cloudflare is biggest ‘toll-taker’ of AI: -> Non-human traffic just crossed 50% of Cloudflare's network for the first time. -> AI agents are the new customers — and they don't sleep don't churn and keep spending Listen to CEO Matthew Prince talking about this jus 2 weeks back 👇

There are 3 bottlenecks weighing on AI CAPEX: 1/ Dem approvals 2/ Cost of Capital (FCF turning -ve) and 3/ Power availability Example: The chart splits the MAG8 into three camps: 1. Self-funding: $AAPL ($137B FCF on $10B capex) and $NVDA ($127B on $7B) — they sell the picks and shovels, they don't have to buy them. Together they're $264B, 2.4x the other 6 combined. $AVGO ($39B on $1B capex) sits here too. 2. Squeezed: $MSFT still makes $60B, but on $136B of capex $GOOGL's TTM FCF is down to $20B — its last two quarters were FCF-negative (−$9.1B, −$5.9B) $META at $29B, with Q3 FCF already negative 3. FCF-negative: $AMZN at −$42B TTM, with 2027 capex guided "meaningfully more" than $220 billion My take: this is cost-of-capital thesis in one picture. - At 0% rates FCF didn't matter — debt was free. - At a 4.9% 10-year, every capex dollar beyond FCF costs real money or dilution, which is why the tape is now watching the funding, not the growth: $META's up-to-$95B equity raise, MSFT's 25B of bonds, $GOOGL's $6B of bonds.

$NBIS IS AT A SERIOUS ADVANTAGE WITH INTERNATIONAL GW AVAILABILITY. US: Vineland NJ 300 MW + Pennsylvania up to 1.2 GW + Birmingham 300 MW International: Finland 75 MW live on cheap green power, Paris, UK, Iceland, Israel colos. >2.5 GW contracted by end-2026, 5 GW target 🚨 AI spending is a 1% game — the top 1% with GW does the land-grab from here. Power, not chips, is the constraint. And $NBIS clearly headed to be that 1% 📈

🚨 AI spending is a 1% game Ramp AI Index October Summary: In October 2026, Anthropic $ANT captured the largest share of AI token spending among businesses on Ramp, at 60.5%. 📈Total token spending rose 8.9% compared with the same week last month. In September 2026, AI adoption reached 55.9% of US businesses on Ramp, up 0.6 % points from August and 12.1 points from a year earlier Top 1% of firms: $9,567/employee/month on AI. Median firm: $15.70 — a 609x gap. 1% of companies drive 80% of all business spend on OpenAI + Anthropic, and 80% of those whales are VC-backed startups. The "enterprise AI boom" is venture money, not the S&P 493. $SPY $QQQ

🚨2 RESONS WHY $PLTR HAS HIGH P/E ?? 1. High Gross Margins (GAAP) : At 83% !! Palantir sells Gotham/Foundry/AIP as a platform — once the code exists, each incremental license dollar costs almost nothing. COGS is basically cloud hosting, customer support and some deployment cost. 2. Revenue Visibility: Rapid acceleration of Commercial revenue gives ‘visibility ’ of earnings potential . Commercial was a sideshow at $284M in Q4 2023 and has 4x'd to $1,168M — overtaking Government for the first time in Q3 2026 After fall today it’s treading at 104x Fwd earnings ie a PEG < 1

🚨 Anthropic jus published amended S-1: 1/ Key Positives:: -> Q3 revenue $17B, up 12x YoY and +47% QoQ 🚀 -> Zero to $17B a quarter in 3.5 years — fastest enterprise adoption of any technology in history. -> Net loss fell to $23.29 billion in Q3 from $66.3 billion in Q2 2. Key negatives: -> GAAP swung from +$974M operating income in Q2 to a -$567M operating loss -> Compute costs hit $9.9B — 58 cents of every $1 revenue, up from 48 ‼️ -> Demand is running ahead of their contracts. $NBIS $AVGO $CRDO -> $ANT buying $5B of short-term compute at "significantly more" than contracted rates because they can't wait 12–18 months. 3. Key CAPEX item: -> $895B of compute commitments signed, $69.8B due in 12 months, against $131.9B of cash. -> Only $GOOG has bigger commitment than Anthropic !!

🚨 Howard Marks just said— “AI is still in 3rd innings of a long cycle” 🚀 💸UBS report shows — AI Debt is still only 25% of AI infrastructure spend — vs over 50% at the dotcom peak. 📈AI is still being built on equity, not leverage. Bubbles pop when leverage does. 1/ How big is the buildout vs history? AI capex is ~2.5% of US GDP today. Telecom hit 1.2% in 2000. Railroads peaked at 6% in the 1870s–80s. The biggest infrastructure boom in history ran 2.5x hotter than where we are now. 2/ The pipeline: $4.2T of future commitments from hyperscalers + AI labs -> Goldman sees $5.3T through 2030. There is still money to be made before the peak — but railroads made millionaires AND bankruptcies. Howard infirecyly pointed Weaker players in Neoclouds can fall out over the next 3 years. 3/ Where the money will go — key beneficiaries vs all-time highs, this morning: $NVDA -13%, $AVGO -32%, $MU -28%, $GOOGL -29%, $META -22%, $AMZN -12%, $MSFT -3%, $PLTR -12%, $TSM -17% $MRVL -30% , $CRDO -50% $LITE -20% $CRWD -11% 5/ The fallout candidates — leverage and weak balance sheets get crushed first: $SMCI -72%, $CRWV -72%, $IREN -70%

🚨 October scoreboard: Nasdaq -5.3% (worst since Mar '25), Semis -15.3% $SOXX $SMH $CRWV -39%, $NBIS -37% Meanwhile S&P 500 -0.9% ( $SPY) Dow +1.7%. Rotation to “Cash flows, Cash flows and Cash flows”

🚨 $5.6 TRILLION WIPED OUT — AI + TECH STOCKS OFF THEIR ALL-TIME HIGHS ⭐ = MAG 8. 🔻Biggest dollar damage: 1. Cipher Mining, $CIFR: -70% 2. Alibaba, $BABA: -66% 3. CoreWeave, $CRWV: -65% 4. Iren, $IREN: -62% 5. Credo Technology, $CRDO: -43% 6. Sandisk, $SNDK: -40% 7. Astera Labs, $ALAB: -39% 8. Nebius, $NBIS: -36% 9. Vertiv, $VRT: -35% 10. Silicon Motion, $SIMO: -35% 11. Vistra, $VST: -34% 12. Bloom Energy, $BE: -34% 13. Intel, $INTC: -32% 14. Alphabet, $GOOGL ⭐: -29% 15. Broadcom, $AVGO ⭐: -27% 16. GE Vernova, $GEV: -25% 17. Meta, $META ⭐: -21% 18. Marvell, $MRVL: -20% 19. Lumentum, $LITE: -19% 20. Micron, $MU: -19% 21. Tesla, $TSLA ⭐: -18% 22. Amazon, $AMZN ⭐: -14% 23. Cloudflare, $NET: -12% 24. Nvidia, $NVDA ⭐: -11% 25. TSMC, $TSM: -10% 26. Palo Alto Networks, $PANW: -6% 27. CrowdStrike, $CRWD: -4% 28. Microsoft, $MSFT ⭐: -2% 29. Reddit, $RDDT: -50% 30. AXT, $AXTI: -54% Leveraged AI ETFs: Semiconductor Bull 3x, $SOXL: -64% Korea Bull 3x, $KORU: -78% Nasdaq-100 Bull 3x, $TQQQ: -17% If this is not bear market what is it???

🚨 Receipt day for my cost-of-capital call. $GOOGL: FCF -$9.1B, search miss 2 quarters running, $500B market cap gone $MSFT capex $45.2B vs $50.7B expected . Up AH Market has shifted rewarding from who is spending more FCF to who is preserving This quarter could mark ‘beginning of end of AI CAPEX’ race but AI stocks will still rip higher for months to come 📈