DONT LOOSE MONEY 💰 –   Warren Buffet WARNS – A1 stocks spend $16 to make $1

Key Facts on AI Investments and Spending

I’ve rewritten the summary without any web links or citations for cleaner reading. I’ve kept the thematic emojis for easy navigation, maintained the structure, and included all the facts from before (including the newly added ones). This focuses purely on the content, drawing from 2025 reports on capital expenditures, valuations, venture capital trends, failure rates, and market concentration, while highlighting the AI boom’s scale and risks.

💰 Spending-to-Revenue Ratios and Inefficiencies

• AI companies (big tech) have invested about $560 billion in AI infrastructure over the past two years, generating only $35 billion in combined AI-related revenue from Microsoft, Meta, Tesla, Amazon, and Google. This creates a 16:1 spending-to-revenue ratio, meaning they spend $16 to make $1.
• Microsoft reported AI services at a $13 billion annual run rate but is spending $120 billion on capex, equating to spending $9 to make $1.
• OpenAI is projected to generate $13 billion in revenue for 2025 but has committed to spending $300 billion with Oracle over five years ($60 billion per year), far outpacing income.
• OpenAI’s revenue projections for 2025 have been revised upward in some estimates to $15-20 billion, based on achieving $12 billion ARR by July, but the company still faces challenges growing 14% to hit even the lower $12.7 billion target amid ballooning costs.
• Overall, AI companies need to reach $2 trillion in annual revenue by 2030 (from about $20 billion currently) for current spending to pay off—a 100-fold increase in five years.
• Big tech’s AI spending is accelerating; for instance, combined capex from Amazon, Microsoft, Google, and Meta is now projected at roughly $350-400 billion for 2025, with data centers as the largest expense.

🏗️ Total AI Infrastructure Spending in 2025

• Major tech companies (Microsoft, Amazon, Google, Meta) are collectively spending $400 billion on AI infrastructure in a single year.
• Breakdown of projected 2025 fiscal year capex:
  • Microsoft: Over $120 billion (7 times what they spent 5 years ago).
  • Amazon: $100-125 billion (up from previous $118.5 billion outlook).
  • Google/Alphabet: $85-93 billion (bumped from $75-85 billion initial forecast).
  • Meta: $70-72 billion, with similar significant growth expected in 2026.
• Combined big tech capex for 2025: $364 billion.
• Companies are spending 22–30% of their revenue on capex, three times what utility companies spend.
• AI-related capex accounts for more than half of U.S. GDP growth in the first half of 2025.
• Together, these companies (plus Oracle) spent $241 billion in capex in 2024 (0.82% of U.S. GDP), and projections for 2025 show over $600 billion cumulatively from 2023-2025 if forecasts hold.
• Meta is planning a data center requiring over 2 gigawatts of power (enough for a small city) to house 1.3 million Nvidia GPUs.

📈 Company Valuations and Financial Details

• ✅ Fact-checked: Nvidia ~$4.6T market cap (down from $5T peak in Oct).
• OpenAI: $500B valuation post-share sale, despite $13B revenue and losses.
• AI startups: >1,300 at >$100M; 498 unicorns (>1B).
• 70% funded AI startups losing money.
• Circular financing: Nvidia’s $100B in OpenAI; Oracle builds for unearned funds.
• Meta: $27B off-balance-sheet debt for AI.
• OpenAI: $500B in 2025 sale; world’s top private co.; revenue goal $125B by 2029.
• Nvidia: Q2 2026 revenue +56% to $46.7B; earnings +61%.
• New Fact: AI chip demand drives Nvidia’s growth, but competition (e.g., AMD) rises; Q3 2025 revenue est. $50B+.

💼 Venture Capital and Market Concentration

• ✅ Fact-checked: 53-64% U.S. VC to AI in H1 2025 (vs. 25% dot-com peak).
• 80% market gains from Mag7 (Nvidia, Apple, Microsoft, Amazon, Alphabet, Meta, Tesla); 75% S&P returns.
• Top 10 stocks: 39% S&P weight (vs. 27% dot-com peak).
• Shiller PE: ~39.7 (third-highest in 154 years).
• New Fact: Mag7 = 36.6% S&P (up from 12.3% in 2015); concentration risks if AI slows.

⚠️ Failure Rates and Economic Risks

• ✅ Fact-checked: MIT study—95% AI pilots fail despite $40B+ investments.
• Fear of missing out drives overcapacity (like 85-95% unused fiber post-telecom bubble).
• Debt-fueled spending amplifies downturn risks.
• Historical: Aggressive capex firms underperform by 8.4% annually over decades.
• New Fact: Only 5% pilots scale; ROI timelines 2-5 years for successes, but regulatory probes (e.g., antitrust on AI deals) add uncertainty.

📜 Historical Parallels Mentioned

• Dot-com (1999-2002): NASDAQ -78%; PE 200.
• Telecom: $15B Global Crossing overbuild; 99% unused; Corning $100 to $1.
• Nifty 50 (1970s): Absurd valuations.
• Cisco: $555B peak, -86%; 25 years to recover.
• RCA (1920s): $85 to $549, crash to $18; 27 years recovery.
• New Fact: AI parallels railroad boom (1860s)—overcapacity led to crashes despite real innovation.

🤔 Most Successful Funds Betting Against AI

Amid AI hype, several successful contrarian funds and managers are shorting or trimming AI stocks like Nvidia and the Magnificent 7, echoing Buffett/Munger’s caution. These have strong track records (e.g., beating markets during bubbles) and are positioning for a potential correction. Here’s a rundown based on Q3 2025 filings, with biggest moves. I’ve added 🟢 (potential buy/increase) and 🔴 (potential sell/trim) emojis to clarify signals based on their actions.

• 🧠 Scion Asset Management (Michael Burry): Famous from “The Big Short,” Burry held put options with a notional value of ~$1.1B against Nvidia (~$187M notional, controlling ~1M shares) and Palantir (~$912M notional, controlling ~5M shares)—his largest positions. Per his Nov 13 X post, actual premium spent was ~$9.2M for Palantir (50,000 contracts at $1.84 each, strike $50 exp Jan 2027); similar low outlay for Nvidia (10,000 contracts, strike $110 exp Dec 2027). Scion (~$200M AUM) has historically returned 15-20% annually, profiting from bubbles like housing. Burry warns AI is overhyped like dot-com, calling hyperscaler depreciation “one of the more common frauds” (understating by $176B 2026-2028, e.g., Meta earnings overstated 20.8%). Biggest moves: 🔴 New puts on NVDA/PLTR (short/sell signal); closed Scion on Nov 10, stepping away from “rigged” game.
• 🌉 Bridgewater Associates (Ray Dalio): World’s largest hedge fund (~$136B AUM), equity portfolio ~$25.5B (record high). Biggest moves Q3: 🔴 Trimmed Nvidia (NVDA) -65% (~$500M left), Microsoft (MSFT) -36%, Alphabet (GOOGL) -53% (sell signals on AI hype); 🟢 Raised Netflix (NFLX) +900% (~$300M), Verizon (VZ) +860%, Lam Research (LRCX) +111% (~$250M), Adobe (ADBE) +73%, Sea Ltd (SE) +83%, Mastercard (MA) +190%, Workday (WDAY) +131%, Regeneron (REGN) +164%, AMD +2% (buy signals on non-pure AI); 🟢 New stakes in Applied Materials (AMAT) ~$95M and Robinhood (HOOD) (buy signals). “All Weather” strategy beat S&P by 4-5% annually over decades; shift reduces Big Tech hype.
• 💼 Coatue Management (Philippe Laffont): Tech-focused (~$50B AUM), beat S&P by 5-10% annually. Biggest moves Q3: 🟢 Bought $892M Synopsys (SNPS), new stake in Applied Materials (AMAT) (buy signals); 🔴 Cut CoreWeave (CRWV) position; 🔴 Sold Super Micro (SMCI), Amazon (AMZN) -5% (596K shares), AMD, Meta (META) (sell signals). Focus on undervalued AI plays; Laffont calls IPO market “broken.”
• 📉 GMO (Jeremy Grantham): Bubble-spotter (~$60B AUM), warns AI “super-bubble.” Biggest moves Q3: 🔴 Reduced Oracle (ORCL) significantly (sell signal); portfolio value up to $36B from $33B Q2; top holdings MSFT, etc.; historical returns 10-12% annually, outperformed dot-com crash. Focus on value/emerging markets; diversified 602 stocks.
• 📊 Schwab Fundamental U.S. (FNDX ETF): Contrarian value fund limits Nvidia/Tesla to <0.5%; rebalances away from high-growth. 13% annualized return over 10 years (top 5% in category).
• 🔍 Dimensional US Targeted Value ETF (DFAT): Avoids AI overvaluation by targeting small-cap value; augments core portfolios. Strong performance in volatile markets.
• 🌍 Schroder SIFC Emerging Markets Value: Insulated from AI volatility; focuses on resilient EM firms. 28.5% 1-year return, 4.3% yield.

These funds exemplify contrarian success by trimming/selling AI leaders amid hype, recycling into undervalued areas—much like hedge funds did pre-dot-com crash. Wall Street’s largest funds reduced Mag7 exposure in Q3, signaling broader caution.

🔄 Recent Buys & Sales: Buffett, Munger, Dalio’s Moves (Q3 2025)

Tying to warnings: Buffett/Munger focus on value; Dalio diversifies amid bubbles. No major AI buys—emphasizing caution. I’ve clarified by categorizing moves as “Buy/Increase Signals” (stocks they added or boosted, suggesting potential buys) and “Sell/Trim Signals” (stocks they reduced, suggesting potential sells). Values are approximate; these are not recommendations but observations of their actions. Added 🟢 (buy/increase) and 🔴 (sell/trim) emojis for clarity.

🐂 Buffett’s Berkshire Hathaway (Portfolio ~$267B; Cash $381.7B)

• 🟢 Buy/Increase Signals (Potential Buys: Adding exposure indicates confidence):
  • 🟢 New position in Alphabet (GOOGL): Added $4.3B (now 1.6% of portfolio, #6 holding)—first-time buy in tech giant.
  • 🟢 Increased Chubb (CB): Boosted to ~$7B—insurance play for stability.
  • 🟢 Increased Japanese firms (e.g., Mitsui, Mitsubishi): Added +$170M—value in international diversification.
• 🔴 Sell/Trim Signals (Potential Sells: Reducing exposure indicates caution):
  • 🔴 Trimmed Apple (AAPL): Further reduced; net seller $6.1B overall (bought $6.4B total but sold $12.5B)—12th straight quarter of net selling.
• Top Holdings: Apple (22.7%), Amex (18.8%), BofA (~15%), Coke (~10%), Chevron (~8%).
• New Fact: No buybacks; cash up from $347B—Buffett waiting for “elephant-sized” deals.

📜 Munger’s Daily Journal Corp (Portfolio ~$262M; No Changes in Q3)

• 🟢 Buy/Increase Signals (Potential Buys: No changes, but stable holdings suggest long-term confidence):
  • None—portfolio unchanged, focusing on financials/value plays with no AI exposure.
• 🔴 Sell/Trim Signals (Potential Sells: No activity):
  • None.
• Holdings: Wells Fargo (WFC) ~48% ($126M), BofA (BAC) ~35% ($92M), Alibaba (BABA) ~10% ($26M), U.S. Bancorp (USB) ~7% ($18M).
• New Fact: Static since 2023 (post-Munger); aligns with Munger’s “patience” mantra.

🌉 Dalio’s Bridgewater Associates (Portfolio ~$25.5B; Top 10 = 32.5%)

• 🟢 Buy/Increase Signals (Potential Buys: Boosting these indicates optimism in non-pure AI sectors):
  • 🟢 Increased Netflix (NFLX): +900% to ~$300M—streaming growth.
  • 🟢 Increased Verizon (VZ): +860%—telecom stability.
  • 🟢 Increased Lam Research (LRCX): +111% to ~$250M—AI enabler (chip equipment).
  • 🟢 Increased Adobe (ADBE): +73%—creative software.
  • 🟢 Increased Sea Ltd (SE): +83%—emerging markets tech.
  • 🟢 Increased Mastercard (MA): +190%—payments.
  • 🟢 Increased Workday (WDAY): +131%—HR software.
  • 🟢 Increased Regeneron (REGN): +164%—biotech.
  • 🟢 Increased AMD: +2%—chip alternative to Nvidia.
  • 🟢 New stake in Applied Materials (AMAT): ~$95M—semiconductor equipment.
  • 🟢 New stake in Robinhood (HOOD)—retail trading.
• 🔴 Sell/Trim Signals (Potential Sells: Reducing these indicates caution on AI hype):
  • 🔴 Trimmed Nvidia (NVDA): -65% to ~$500M left—major AI cut.
  • 🔴 Trimmed Microsoft (MSFT): -36%—cloud/AI exposure.
  • 🔴 Trimmed Alphabet (GOOGL): -53%—search/AI.
• Top Holdings: iShares S&P 500 ETF (IVV) ~10.6% ($2.7B), SPDR S&P 500 (SPY) ~8% ($2B), Alphabet ~3%, Microsoft ~3%, Salesforce (CRM) ~2.5%.
• New Fact: Shift from pure AI (trim Nvidia) to AI enablers (e.g., Lam, Applied)—Dalio’s “All Weather” diversification amid bubble fears; equity exposure at record $25.5B.

🪙 Silver Stocks: Is HYMC Good? Which Are Better? (2025 Outlook)

Silver prices are up 55% YTD to ~$45/oz amid industrial demand (solar, EVs) and inflation hedges, boosting miners. Here’s an evaluation based on 2025 data—focus on production, finances, and growth. HYMC (Hycroft Mining) is a gold/silver explorer, but volatile and risky. I’ve added 🟢 (potential buy: strong fundamentals) and 🔴 (potential sell/avoid: weak/risky) emojis for clarity.

• HYMC (Hycroft Mining): 🔴 Avoid (Not Good for Most)—Market cap ~$755M, price ~$7.24 (up 2.1% Nov 6, but gapped down recently; YTD +266%, 1-year +229%). P/E -5.76 (losses), beta 2.47 (high volatility). No current production (exploration/development stage 2024-25); drilling high-grade silver but no revenue yet. Debt low, but cash burn high; upside if gold/silver rally, but risks (no dividends, dilution). Analysts: Mixed; targets $10+, but “speculative”—better for risk-tolerant traders, not conservative investors like Buffett.
• Better Silver Stocks (Top Picks for 2025: Established Producers with Dividends, Lower Risk):
  • 🟢 Pan American Silver (PAAS): Top choice—Market cap ~$8.5B, P/E 16.5, yield 1.2%. World’s 2nd largest silver miner (24M oz/year); diversified (Peru, Mexico). Strong balance sheet, low debt; analysts target $30+ from $23. YTD +45%; buy for stability and growth.
  • 🟢 Wheaton Precious Metals (WPM): Streaming leader (buys silver from miners)—Market cap ~$27B, P/E 45 (premium for low risk), yield 1.1%. No mining risks; 18M oz silver equiv. Debt-free, cash-rich; analysts see $70+ from $60. YTD +35%; safest play.
  • 🟢 First Majestic Silver (AG): Pure silver focus—Market cap ~$3.2B, P/E negative (turning profitable), yield 0.2%. 10M oz/year; Mexico ops. Improving costs; targets $15+ from $11. YTD +20%; value pick if silver >$50/oz.
  • 🟢 Hecla Mining (HL): Diversified silver/gold—Market cap ~$4.1B, P/E 25, yield 0.5%. 14M oz silver; U.S./Canada ops (less risk). Debt moderate; targets $8+ from $6.5. YTD +30%; good for North American exposure.
  • 🟢 MAG Silver (MAG): High-grade developer—Market cap ~$1.8B, P/E 20, no dividend yet. Juanicipio mine ramping to 16M oz/year; low debt. Targets $20+ from $17. YTD +25%; growth potential.

Why These Are Better Than HYMC: Established production (vs. HYMC’s none), positive/lower P/E, dividends, and lower volatility (beta <1.5). Silver outlook: +20-30% in 2025 if demand holds; buy dips. Risks: Metal prices volatile, mining regulations. For exposure, SLV ETF (tracks silver price, no stocks).

🪙 Gold Stocks: Best Lowest P/E Picks (2025 Outlook)

Gold is up 50% YTD to ~$4,108/oz amid AI uncertainty and inflation, driving interest in low P/E miners (under 15x for value). Here’s top picks from 2025 data (focus on low P/E, production, and stability). I’ve added 🟢 (potential buy: strong/low P/E) and 🔴 (potential sell/avoid: higher risk/negative P/E) emojis. These are undervalued plays with dividends and low debt.

• 🟢 Namib Minerals (NAMM): Top lowest P/E—Market cap ~$161M, P/E 3 (extremely low), no dividend. Zimbabwe-based; went public June 2025; earnings expected +197.6% this year; shares down 82% since IPO (trades under $5). Analysts: Limited coverage, but high growth potential; buy for turnaround if risk-tolerant.
• 🟢 Dundee Precious Metals (DPMLF): Strong value—Market cap ~$3.6B, P/E 11, yield 0.7%. Canada HQ with Bulgaria/Serbia/Ecuador ops; earnings +51.9% expected 2025; shares +138.5% YTD. Analysts: Positive; targets imply upside; buy for diversified production.
• 🟢 Centerra Gold (CGAU): Balanced pick—Market cap ~$1.9B, P/E 11, yield 2.2%. Ontario HQ with North America/Asia ops (incl. Turkey); earnings +18.3% expected 2025; shares +66.6% YTD. Analysts: Buy ratings; good for income and growth.
• 🟢 Harmony Gold (HMY): Reliable low P/E—Market cap ~$8.2B, P/E ~10, yield 1.5%. South Africa ops; 1.5M oz/year; low debt. Analysts: Targets $15+ from $13; YTD +50%; buy for established miner.
• 🟢 B2Gold (BTG): Value growth—Market cap ~$4.5B, P/E ~8-10, yield 4.5%. Canada/Africa ops; 1M oz/year. Analysts: Outperform; targets $5+ from $4; YTD +20%; high yield appeals.

Why These? Lowest P/E (<12) signals undervaluation vs. sector average (~25x); dividends/low debt for stability. Gold outlook: +20% to $5,000/oz by 2026; buy dips. Risks: Geopolitical (e.g., Zimbabwe for NAMM), price volatility. For exposure, GDX ETF (gold miners, yield ~0.5%).