Anthropic S-1 Shocker, Greatest Business Ever? Google's $500B Wipeout, Midterm Predictions

What It Covers

Anthropic's S-1 filing reveals $16.3B revenue in H1 2026, 85% gross margins, and $112B cash on hand, positioning it as potentially the largest IPO in history at a $1.5–2T valuation. Google loses $500B in market cap after search revenue decelerates to 12% growth, and the hosts make midterm election predictions.

Key Questions Answered

• •Anthropic Capital Efficiency: Anthropic raised approximately $110B total but spent only $9–10B building the business, retaining over $100B in cash. This refutes the widely circulated $80B loss narrative, which was a phantom accounting figure from convertible notes tied to valuation increases — not actual cash burn. Investors evaluating AI companies should scrutinize whether reported losses reflect real cash outflows or non-cash accounting adjustments from convertible instruments.
• •AI Compute Cost Curves: Anthropic achieves roughly 10x cost reduction per task annually through algorithmic improvements called "compute multipliers." They pass ~50% of savings to customers via lower token prices while capturing the remainder as margin expansion, pushing gross margins to 85%. Enterprises evaluating AI vendor sustainability should track whether margin improvement accompanies price cuts — a sign of genuine efficiency rather than subsidized pricing.
• •Customer Penetration Runway: Anthropic has 11,900 customers spending $100K+ annually, yet the global addressable base of businesses for whom $100K represents under 1% of total spend exceeds one million companies. Current penetration sits near 1%. Enterprises still in pilot or sub-$100K spend phases represent the next major revenue tier, and vendors with negative churn dynamics — where existing customers expand spend — compound this growth significantly.
• •AI Complex vs. Treasury Competition: The combined debt and capital-raising needs of roughly ten major AI companies — Meta, Alphabet, Microsoft, Anthropic, OpenAI, and others — equals approximately 30% of projected U.S. Treasury issuance through 2027. This creates a structural market risk: either AI firms pay higher borrowing costs, compressing returns across the stack, or Treasury yields rise, increasing borrowing costs economy-wide. Investors in AI equities should monitor bond auction demand as a leading indicator.
• •Google Search Revenue Dynamics: Google Search revenue grew 12% year-over-year to $63B, but its share of total Alphabet revenue fell below 50% for the first time — driven by Google Cloud doubling to $30.7B in revenue with operating income rising from $3.5B to $11B in one year. Alphabet's diversification away from search dependency is accelerating. Investors should weight Google Cloud's trajectory separately from core search when modeling Alphabet's forward earnings multiple.
• •Jevons Paradox in Token Economics: As Anthropic cuts token costs, enterprise consumption accelerates rather than stabilizes — a textbook Jevons paradox. Agents consume 50–100x more tokens than simple prompts, and new use cases like continuous cybersecurity monitoring compound demand further. Enterprises building AI cost models should assume token consumption will grow non-linearly as prices fall, making fixed-budget AI procurement strategies structurally inadequate within 12–18 months.