The Great Unraveling of Anthropic

In this episode, Ed Zitron, known for his deep dive into OpenAI's finances, analyzes the leaked financial data of Anthropic from their confidential S-1 filing for an IPO. The numbers paint a picture even bleaker than OpenAI's.

A Look at the Books: Spending $2.75 to Make $1

• Dire Finances: Anthropic lost a staggering $8 billion in 2025 on a revenue of just $4.6 billion. The company spends $2.75 to generate every $1 of revenue.
• Competitor Dependency: Almost 47% of their revenue comes from reselling their models through Google and Amazon. This is a massive concentration risk, as these partners are also direct competitors.
• The Debt Trap: The company is locked into $413 billion in non-cancellable compute contracts ($252B with hyperscalers, $161B with Broadcom). This debt is baked into the earnings expectations of their biggest partners.

The Deceptive "Run Rate" Metric

Zitron tears apart the industry's favorite vanity metric, the Annualized Run Rate (ARR):

• Daily Extrapolation: Instead of looking at stable subscriptions or contracts, companies often take a single high-spend day and multiply it by 365 to inflate projections.
• No Standard Definition: The lack of transparency allows for wild, unsubstantiated claims. "It's like taking my best day on YouTube and claiming it's my annual viewership."

The Products: Muse, Instinct, and the Hype

Even the new "Agentic AI" products like Claude Muse or OpenAI Dots can't save the business model:

• Useful but Unprofitable: These tools finally deliver on some old promises (booking, email sorting) but at an immense infrastructure cost and with enormous liability risks.
• Missing Mass Market: The industry is a B2B and developer play. No consumer-grade killer feature justifies the massive spending.

The Big Picture: A Financial Crisis in the Making

Zitron argues that AI hasn't boosted productivity but has instead sparked a systemic financial crisis:

• GDP growth is mostly from GPU and datacenter construction itself, not from AI software output.
• The entire edifice is built on a house of cards involving private credit, bond markets (Oracle, CoreWeave), and Japanese banks (SoftBank, MUFG). A bust would be worse than the Dot-com bubble.
• Conclusion: Zitron demands transparency. "Anyone trying to give OpenAI or Anthropic investment grade credit should go to jail." The IPO will likely happen, but the underlying fundamentals are a disaster waiting to unfold.

📊 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.

In this episode of Un(easy) Money, the hosts discuss the riskiest aspects of the crypto market. The focus is on Anthropic‘s IPO and the governance structure they identify as the biggest danger. Additionally, they analyze the increasing security threats from AI-powered phishing attacks and recent incidents at hardware wallet manufacturers.

Anthropic IPO: Billion-Dollar Risk from Governance Experiment

• Anthropic filed for its IPO. Revenue stands at $5B (2025) and $12B (Q2 2026) — much higher than Google’s $2.7B before its 2004 IPO.
• Biggest Risk: The seven founders hold a combined 50% voting bloc in a separate pool and must vote together. The panel sees this as a massive governance risk.
• Comparison with DAOs and multisig problems: The structure is unproven and could fail in case of internal conflicts. There is a danger the founders will trip over themselves.
• The company is going public unprofitable but has potential for enormous profitability once chip costs decline and models become smarter.

AI Progress: Both Blessing and Curse

• Models like Opus 5.5 and Terra 6.1 deliver extremely good results – sometimes better than humans at math problems.
• Concerns: The speed of development and AI‘s growing ability to handle complex tasks are already changing education and everyday life (e.g., children using Claude instead of Google).
• At the same time, phishing attacks are becoming much more dangerous through AI: Scammers use clankers and merged databases to create personalized attacks.

Hardware Wallet Security: Two Major Incidents

• Keystone: Leak of customer data (names, addresses) and the supply chain signing key. Data was supposed to be deleted after 180 days but wasn't.
• Foundation: Hackers gained access to email systems and sent perfect phishing emails containing customers‘ personal data.
• Recommendations: Do not panic, do not engage with unsolicited calls or emails, and do not rush to buy new hardware wallets as further leaks are likely. Better: stick to existing processes and learn long-term.

Ethereum Upgrade: Fast but Safe?

• The “Glamsterdam” upgrade caused a gas price issue that could have locked liquidity positions. The bug was caught in time.
• Critique: Core developers focus heavily on the protocol level and overlook potential errors at the application level. A more systematic approach (e.g., fuzzing the entire chain) would be desirable.

Conclusion

The panel warns: “This is one of the most dangerous times for crypto.” To protect your assets, don't panic, stick to your proven security measures, and never trust unsolicited contact.

Historic Breakthrough: Wall Street Embraces Crypto Perpetuals

• The Big News: ICE (owner of the New York Stock Exchange) has signed a deal to offer 24/7 perpetual futures on the Hyperliquid blockchain in the US – in partnership with market maker tradeXYZ. Final CFTC approval is pending but considered a formality.
• Heavyweights Involved: Stanley Druckenmiller (legendary investor) surprisingly expressed optimism about the partnership. Elon Musk also called the development “interesting,” underscoring the event’s significance.
• Why ICE is Entering: The derivatives market for institutional and especially retail traders is enormous. 98-99% of current business is institutional – retail is seen as a massive growth opportunity. tradeXYZ has already processed over $650 billion in volume in just over a year.
• Hyperliquid (HYPE) in Focus: HYPE is the dominant on-chain derivatives exchange. With only 13,000 users, it generates $2.3 million in daily fees – 9 times that of Ethereum and 4 times that of Solana. 88% of fees are used to buy back HYPE tokens, creating strong upward pressure.
• Impact on Other Chains:
  • Solana (SOL): Benefits indirectly from overall crypto adoption. Projects like Jupiter and pump.fun are also partnering with Hyperliquid, and Solana remains the leader in daily active users (nearly 5 million).
  • Ethereum (ETH) & Bitcoin (BTC): HYPE has outperformed ETH by 1,151% and SOL by 567% since its inception. Activity on ETH and BTC is comparatively lower.
  • Options vs. Perpetuals: Perpetuals do not replace options. Perps are pure leveraged directional bets (no expiry, no time decay), while options remain useful for hedging and volatility speculation.

Conclusion

ICE’s (NYSE) entry into crypto perpetual trading is a historic milestone. Traditional finance and crypto are converging. Hyperliquid offers the cleanest exposure, while Solana provides broad sector beta.