OpenAI's Identity Crisis, Datacenter Wars, Market Up on Iran News, Mamdani's First Tax, Swalwell Out

What It Covers

The All-In hosts, joined by Travis Kalanick, analyze OpenAI's strategic identity crisis against Anthropic's 10x annual growth rate, the accelerating data center permitting collapse across 30 states, New York City Mayor Mamdani's proposed 3.9% annual pied-à-terre tax on properties over $5M, Eric Swalwell's congressional resignation amid coordinated allegations, and market dynamics with the S&P hitting all-time highs despite ongoing Iran conflict.

Key Questions Answered

• •Anthropic vs. OpenAI Growth Divergence: Anthropic is growing at roughly 10x annually versus OpenAI's 3-4x, scaling from $1B to $10B ARR in one year and projecting $80-100B by year-end. Enterprise coding tokens billed like electricity — metered, scalable, uncapped — drive this gap. Consumer subscribers cap at $20/month all-you-can-eat plans with only 3-4% conversion rates, making enterprise the only revenue model that compounds at the scale needed to justify frontier lab valuations.
• •Compute Dependency as Existential Risk: Both OpenAI and Anthropic built their businesses on hyperscaler compute from AWS, GCP, and Azure, which now represents a strategic chokehold. Hyperscalers control 60% of all compute globally. As frontier labs hit capacity ceilings, they must build proprietary data centers — but years of doomer-aligned lobbying against data center construction has salted the regulatory earth they now need to build on, creating a self-inflicted infrastructure crisis.
• •Data Center Permitting Collapse: Approximately 100 data centers are currently contested across the U.S., with roughly 40% getting canceled — a rate that has more than doubled year-over-year. The total economic value of contested projects reaches $162B. Opposition comes from three coordinated sources: utility ratepayer fears, well-funded doomer groups reframing AI risk as water/energy consumption, and Anthropic's political alliances with NIMBY coalitions that now obstruct the very infrastructure Anthropic itself requires.
• •Pied-à-Terre Tax Demand Destruction: New York City's proposed 3.9% annual tax on non-primary residences valued above $5M targets the most price-elastic segment of the real estate market — owners who can place capital anywhere globally. London's equivalent stamp duty reform produced measurable high-end market collapse and redirected wealthy buyers to Zurich, Lugano, and Milan. A $10M New York unit becomes a $20M effective purchase after a decade of compounding tax, eliminating investment rationale entirely.
• •Enterprise AI ROI Still Unproven at Scale: Despite exponential model-layer revenue growth, no large enterprise has publicly demonstrated scaled profit improvement attributable to AI deployment. Change management — not model capability — is the primary bottleneck, as complex undocumented processes inside large organizations resist rapid transformation. Founder-led public tech companies report faster feature deployment cycles, but the productivity gains visible in startups like TaxGPT (serving 6-7% of all U.S. accountants) have not yet translated to measurable bottom-line impact at Fortune 500 scale.
• •Capital Subsidy vs. Revenue Flywheel: Travis Kalanick frames the OpenAI-Anthropic race through the Uber-Lyft network effects lens: whoever scales usage through contribution-margin-positive revenue builds a compounding flywheel that capital subsidies cannot permanently replicate. OpenAI's $122B raise — the largest private round in market history — buys time but not structural advantage. Once token costs get passed through to enterprise customers rather than subsidized, organizations will scrutinize AI output quality, and "vibe-coded slop" from poorly governed agents will face elimination from budgets.
• •Stock Market as Trump Policy Barometer: The S&P 500 recovered all Iran-conflict losses by Tuesday and hit fresh all-time highs by Thursday, pricing in conflict resolution before any deal was signed. Kalanick's framework: Trump uses equity market performance as his primary policy feedback mechanism, tolerating volatility only within a defined band before pivoting toward resolution. Traders have internalized this pattern — sell the escalation, buy the de-escalation — making the market itself a real-time prediction instrument for geopolitical outcomes under the current administration.