Summary: Fed Policy, Hyperscaler CapEx Chaos & Trading Compute Like Oil

In this live deep dive with Steve Hou (Research Lead at Silicon Data), the hosts dissect the market's reaction to Big Tech earnings and the fundamental economics of the AI infrastructure buildout. The central question is: Is the current AI investment boom built to last?

Hyperscaler Earnings: A CapEx Hangover

• Growth vs. Expectations: Google and Microsoft reported explosive cloud growth (Google Cloud <100% YoY). However, their capital expenditure (CapEx) fell short of market expectations.
• Focus Shift: The stock market punished companies that didn't raise their CapEx guidance. This signals a shift in the debate from the volume of spending to the realization of tangible ROI.
• Cloud Profitability: Google's operating margins expanded from 23% to 36%, proving the cloud growth is not just top-line driven but highly profitable.

The Hard Economics of GPUs and Data Centers

• EBITDA vs. Net Income: While data center EBITDA (e.g., CoreWeave) is massively positive, enormous depreciation expenses on expensive hardware consume a huge portion of the profits. The gap between these metrics is historically wide.
• GPU Rental Prices are Soaring: Spot prices for H100, B200, and B300 chips are rising almost vertically. This indicates a severe demand overhang despite continuous new capacity coming online.
• Token Deflation: Simultaneously, the price of tokens is falling. Bears see this as a race to the bottom and commodification. Bulls see it as the natural proliferation of intelligence as a cheap resource.
• The Circular Demand Problem: A significant portion of demand comes from AI labs themselves subleasing capacity. Is this a sustainable flywheel? Anthropic's S-1 filing provides a strong bullish data point with 88% gross margins and 8% operating margins in Q2.

Trading Compute Like a Commodity

• Contango vs. Backwardation: The implied forward curve for GPU rental pricing currently shows contango (rising future prices). This is unusual for compute and implies expectations of persistent scarcity rather than a glut.
• The Risk of Long-Term Contracts: Companies like CoreWeave locked in cheap long-term contracts before the recent price surge. While this guaranteed revenue, it prevents them from capitalizing on the current high prices, echoing the hedging dilemmas seen in oil and gas markets.

Model Economics & Enterprise Adoption

• Closed vs. Open Source: The debate on switching to cheaper models is more complex than just price. Despite lower costs, enterprise switching remains slow.
• Stickiness and Friction: The iPhone analogy (superior specs vs. seamless user experience and brand loyalty) highlights that intangible hurdles and compliance concerns often make switching economically unattractive, even when cheaper open-source alternatives exist.
• Low Spending Share: While many enterprises use Anthropic models, spending accounts for ~0.1% of their budget. This can be interpreted as either massive room for growth (bullish) or limited current utility (bearish).

Macro Context & Consolidation

• Fed Policy: The Fed's interest rate policy is currently a secondary factor. As long as the returns on AI investments remain high, loans will be secured despite rising rates.
• M&A Trend: Marvell's acquisition of Lumentum (optics) highlights the trend towards vertical integration in the semiconductor supply chain, seeking pricing power and scale in an environment of persistent scarcity.

The Challenge of Large V60 Brews

The creator historically disliked V60 doses over 15-18g, believing linear scaling didn't work. This video introduces a new framework for a 30g brew that focuses on sweetness and body without sacrificing clarity.

The "2+ Cup" Approach (30g)

Parameters:

• Coffee: 30g
• Grind: Slightly coarser than for 15g (about half a number coarser on an Option-O 80 with SSP Brew Burrs)
• Water Temp: 85–88 °C (soft water is recommended to accentuate natural acidity)
• Hardness: 20-30 GH, 5-10 KH

The Pour Structure (The Key)

The secret is a gentle "triple bloom" to saturate the deep bed, followed by a fast final pour that acts like an immersion:

• 0:00: 90g gentle pour (spiral to center, ~5g/sec)
• 0:20: 70–80g pour
• 0:40: Top up to 200–210g
• 1:00: Fast center pour up to 450g total. This increases the perceived sweetness significantly.

Drawdown: 2:00 – 2:30 minutes.

Taste Profile & Results

• 30g Batch: Much sweeter (honey, ripe pear), fuller body, buttery texture. Lower floral notes, richer mouthfeel.
• 15g Batch: Higher floral complexity, sharper acidity, slightly thinner body.
• Target: ~16.5% extraction, ~1.25% TDS.

Key Takeaways

• This is a framework, not a rigid recipe. It uses the deeper bed to its advantage.
• Don't worry about grounds sticking to the walls—it helps filter fines.
• The final center pour is the key to unlocking sweetness.

Conclusion: A revolutionary method for anyone looking to brew a truly delicious big batch (2+ cups) on a V60.

The Attack Timeline

• On October 29, starting at 20:05 UTC, on-chain trackers reported unusual outflows from Binance hot wallets. At 20:24 UTC, the Ethereum wallet Binance 14 was drained in a single transaction worth about $2.5 billion.
• The attack then expanded to hot wallets on Bitcoin, Tron, Arbitrum, Optimism, Avalanche and Solana – stablecoins, BTC, ETH and many other tokens were stolen.
• Binance paused deposits and withdrawals worldwide. Former CEO CZ confirmed the hack, Co-CEO Yi He said user funds were safe, and Binance.US was not affected.

Attack Vector

• Initial findings suggest the attack started with the compromise of a DevOps engineer's laptop. The attackers gained access to a service container in Binance's backend infrastructure and used it to authorize unauthorized withdrawals.
• On Ethereum, the attackers used EIP-7702 to turn a hot wallet into an automated “sweeper” that drained its balance and also redirected new deposits to the attacker.
• Binance says there are no signs that cold storage or master keys were compromised; the exact transaction authorization process is still under investigation.

Perpetrators and Damage

• Blockchain analysis firms such as TRM Labs, Elliptic and Chainalysis suspect North Korea's “Trader Traitor” unit (part of the Lazarus Group). This has not yet been officially confirmed.
• At roughly $6 billion, this is the largest exchange hack in history – the previous record was $1.5 billion (Bybit, February 2025).

Freezing and Mitigation

• Binance plans to cover all losses from its SAFU fund and its own balance sheet. The loss equals about 4% of its roughly $160 billion in assets under management.
• Tether froze $1.6 billion in USDT, Circle froze over $800 million in USDC, Ripple froze about $200 million in RLUSD, and Aave paused affected lending pools.
• In total, $2.5 to $3 billion (40–50%) of the stolen funds were frozen. However, over $570 million in ETH has already moved through Tornado Cash, Uniswap and cross-chain bridges.

Market Impact

• Bitcoin fell from around $81,000 to a low of about $76,400 – the weakest level since mid-September. At the time of writing, BTC was recovering near $78,500.
• Within 24 hours, about $1.5 billion in derivative positions were liquidated, nearly 90% of them longs.
• BNB dropped 5–8% to around $680, and Ethereum fell about 5%.
• Despite the biggest hack in crypto history, user funds remain protected and the industry is actively working to contain the damage.

📊 Bitcoin Price Predictions: Bears vs. Bulls

• Jason Calcanis (Bear): Bitcoin will never hit another ATH. "Dump it all now." AI investor on a victory lap. Prediction: Bitcoin falls below $57k.
• Willy Woo: 10,000 Monte Carlo simulations → Median $1.5 million (2034). 2028 top between $240,000 and $413,000.
• Bitwise: Base case $1.6 million per coin (2036). Bull case: $3.6 million. Bear case: $131,000.
• Alex Thorn / Michael Nado: Bottom confirmed. Focus on 50-week moving average (~$76k). Nado target: $200k - $375k for this cycle.

Our Take: Four-year cycles are dead, but the store of value thesis is intact. Current bears make us bullish.

🏦 Macro: Fed & Inflation

• Fed Decision: Rates unchanged (10:2 vote, 2 for hike). Polymarket sees 80% chance of a December rate hike.
• Sticky Inflation: PCE at 3.6% (3 months trending up). Oil at $111/barrel – Iran escalation risk could push inflation higher. Consumer sentiment low, spending high.

🚀 Hyperliquid & Tokenization Conquer Wall Street

• ICE (NYSE) & Trade XYZ building on Hyperliquid: A CFTC-regulated perpetual futures venue in the US. Hyperliquid hits all-time highs (+17%).
• Industry Giant Support: Stanley Druckenmill's first tweet in over 8 years! Elon Musk retweets.
• Tokenized Stocks: ICE/OKX list stocks (Nvidia, Apple). 30 firms file SEC veto to delist. WSJ ad warns CEOs ("Common Sense Leadership Fund"). Robinhood lists them offshore anyway. "Price discovery" is increasingly migrating onchain (e.g., after hours).

🎰 Robinhood Chain: Hype or Bubble?

• Q3: $50M Revenue - BUT: September earnings collapsed from $8.3M/day to $32K/day (99.6% drop).
• Vlad pledges investment. Memecoin holders demand more chain support and listings. Debate about casino economics vs. sustainable growth.

⚖️ Zcash: The Treasury Curse

• Block Reward Mechanic: 20% of issuance (8% committee, 12% coinholder vote) goes to grants.
• Treasury Explosion: $125M war chest, spending ~$6M/year, inflow ~$68M.
• Spending Drama: $550k for a booth at a "Sexpo" convention. Severe political infighting leading to calls for ending the subsidies (terminate grants).

🔥 Venice & Buyback Economics

• Record Burn: Over 1M VVV burned in October. 1% daily burn growth.
• Hybrid Buyback Model: From subscriptions, API credits & profit. Praised for value accrual, criticized for diverting from reinvesting in growth (Amazon vs. dividend debate).

💡 Quick Hits & Coin-Tool

• Athena: 12% yield on USD in the Athena Pay app (leveraged, caution advised).
• Anthropic IPO: $1.5 - $2 trillion market cap? Ryan is not a buyer (AI bubble, many competitors like DeepSeek, token commoditization akin to L1 wars).

In the Halloween episode of Empire, the hosts and guest Haseeb Qureshi (Dragonfly) discuss whether crypto’s bull market is entering a new phase. The trigger: Stanley Druckenmiller and Intercontinental Exchange (ICE) investing in TradeXYZ and Hyperliquid.

Key Points

• Druckenmiller effect: The move is compared to Paul Tudor Jones calling Bitcoin the “fastest horse” in May 2020. That removed career risk for fund managers; this could now accelerate onchain volumes and perpetual markets.
• ICE & Hyperliquid: ICE appears to be building a US-regulated 24/7 perps market using Hyperliquid, potentially leading to a major fight with CME and incumbents.
• Data as a product: Like Polymarket, the most valuable output may be market data – especially pre-IPO pricing for Anthropic, OpenAI, SpaceX and similar names.
• Bull market phases: The group sees the market in an early to mid stage: prices are rising, but euphoria hasn’t taken over. A possible late-cycle signal is Y Combinator again courting crypto startups, which reminds them of 2021.
• YC’s startup list: AI capital raising, stablecoins, trading and agents are welcomed in spirit, but many are considered naive or already failed ideas (e.g., “Bloomberg for crypto”, tokenizing random assets).
• Google MPC custody: Google Cloud launched an MPC-based key-management product, but it’s a backend building block, not a full custodian like Fireblocks.
• AI vs. crypto: AI mostly benefits incumbents with massive capex, while crypto historically created value for startups. Political risks post-midterms could make crypto an easy target if Democrats push corruption probes.

Outlook

The ICE-Druckenmiller news is seen as one of the most bullish catalysts for perps and onchain capital markets, but regulatory fights and political headwinds remain.

The Core Thesis: The Bear Market Has Bottomed Out

Alex Thorn, Head of Research at Galaxy, argues that Bitcoin bottomed in June/July 2025. His thesis rests on four main pillars:

• Technical Indicators (Cycles): The 200-week moving average has historically served as the floor, while the 50-week moving average (around $75k) is the ceiling of bear markets. Bitcoin has been trading above the 50-week MA for weeks, which has historically signaled the definitive end of every bear market.
• Fundamental Thesis: The core reason to own Bitcoin (fiat debasement, 'Digital Gold') is stronger now than ever before.
• Positioning: The holder profile has fundamentally shifted (fast money out, slow money in).
• Market Sentiment: Everyone is fixated on stablecoins and tokenization—the perfect backdrop for accumulating Bitcoin.

The 'Slow Money' Shift (Positioning)

This is the most compelling part of the analysis:

• Sellers ('Fast Money'): Macro hedge funds like Brevan Howard, Millennium, and Capula have drastically reduced or fully exited their Bitcoin ETF positions (13F Q2 data). Their trading is highly sensitive to macro noise (e.g., AI trade fears).
• Buyers ('Slow Money'): Morgan Stanley Wealth Management and sovereign wealth funds like Mubadala (Abu Dhabi) are accumulating significantly. This is patient, long-term capital that doesn't trade on daily headlines.
• Declining Selling Pressure: The awakening of 'old coins' (OG whales selling after years of dormancy) has fallen sharply from 67k BTC/month (2024) to ~34k BTC/month (2026).

Fundamental Thesis: Digital Gold Has Become Unignorable

• National Debt: US debt has nearly doubled since 2020 to $40 trillion. The incentive to print money has never been higher.
• AI Infrastructure: The massive credit expansion for data centers is effectively private money printing. If the AI market crashes, government bailouts will follow. Both scenarios constitute debasement that benefits Bitcoin.
• Censorship Resistance: In an era of AI surveillance and programmable money (tokenization), Bitcoin remains the only major bearer asset without an issuer – an 'exit' from the system.

Prediction & Risks

• Price Target: Conservative estimate of $125,000. Longer-term potential of $250,000 by end of 2027 (forecast from late 2024).
• Primary Risk 1: A sustained breakdown below the 50-week moving average would invalidate the technical thesis.
• Primary Risk 2: A severe macro-economic shock (depression) or the emergence of a cryptographically relevant quantum computer.

Conclusion: Thorn sees a historical turning point. The thesis for Bitcoin's existence is stronger, the sellers have exited, and patient capital is building. He believes this represents an optimal risk-adjusted entry point.

Live Show Summary: Bitcoin Q4 Dump

In this live stream, Ben analyzes the current Bitcoin price correction within the context of midterm years. He emphasizes that such corrections are typical for this cycle and discusses two main scenarios for the bottom.

Current Market Situation

• Bitcoin Price: ~$77,000, exactly at the 21-week EMA (Bull Market Support Band).
• Lowest Price: $76,642 (matching the 21-week EMA at $76,612).
• Context: Bitcoin is approaching the US midterms (early November), a historically recurring pattern for Q4 lows.

Two Scenarios for the Bottom

Scenario 1: The swift collapse ("rip the band-aid off")

• Similar to the FTX crash (Nov 2022): Bitcoin fell 25-30% in just 2 days.
• The bottom could occur before the midterms if panic erupts (e.g., due to an exchange hack).
• Indicator: A massive volume spike at the low.

Scenario 2: The delayed correction

• Bitcoin finds temporary support at the Bull Market Support Band (73-77K USD).
• The final bottom occurs shortly after the midterms (historical trend: all midterm years since 2014).
• Comparison with the SpaceX IPO correlation (October structure resembles current movement).

Macro Factors

• Tech stocks correcting: Google (-12%), Microsoft, Amazon (-8% after hours), Intel (-23% after hours). The NASDAQ is down for 3 consecutive weeks.
• Exchange hack: According to recent news, $5.9 billion were stolen, $3 billion frozen. Ben warns such hacks are common in midterm years.
• Commodities: Oil (Brent) at $118, target $120; silver and gold are also seeking their Q4 lows.

Investor Psychology

• DCA strategy: Historically safe: buy Bitcoin in Q4 of the midterm year and sell in Q4 of the halving year.
• Retail interest: 90% fewer YouTube views compared to 2021 – the "altcoin season" is over.
• Whale activity: No significant increase – typical for the sideways phase before the bottom.

Recommendations

• Stablecoins: Ben advises using real USD instead of stablecoins (risk of de-pegging, e.g., USDC incident 2022).
• Crypto on exchanges: "Treat exchanges like public restrooms – go in, do your business, and get out."
• Conference: ITC Premium Event (November 20-21 in Miami) – tickets increase starting November 1st.

Conclusion

Ben is uncertain which scenario will play out but remains open to both:

• If Bitcoin fails to hold the 21-week EMA, the bottom could come before the midterms.
• If the EMA holds, the low is more likely after the midterms.

Key message: "Until it breaks, let's not assume it's broken." Despite the pain, this is a recurring pattern – Bitcoin has always recovered in midterm years.