Elon's Anthropic Deal, The Next AI Monopoly?, "FDA for AI" Panic, Trading the AI Boom

Episode Summary

Executive Summary: The episode centers on AI infrastructure, especially Elon Musk’s Colossus lease to Anthropic and the rise of “EWS” as a hyperscaler, then pivots to the policy fight over whether AI should face FDA-style approval. The hosts argue AI demand is real but bottlenecked by compute and power, that current regulation risks regulatory capture, and that AI is already boosting GDP, cloud growth, and market performance, even if labor-market effects remain mixed.

Main Topics: Elon Musk, Anthropic, and the EWS hyperscaler thesis (Priority: 5/5): The hosts discuss Musk leasing all of Colossus One to Anthropic as a major validation of compute-as-a-business, framing it as the emergence of Elon Web Services and a new layer of monetization for SpaceX/XAI infrastructure. AI compute, power constraints, and data-center buildout (Priority: 5/5): They argue AI revenue is constrained by supply of data centers and electricity rather than demand, and that whoever secures power and compute early will dominate the frontier-model race. Regulation debate: FDA for AI vs targeted guardrails (Priority: 5/5): A long exchange covers rumors of a White House review regime for AI models, with Sachs, Brad, and others rejecting a pre-release approval system while accepting narrow cybersecurity coordination and KYC-like controls. Anthropic, OpenAI, and monopoly/competition dynamics (Priority: 4/5): The hosts debate whether Anthropic’s rapid growth could make it a historic monopoly, while also emphasizing that competition among frontier labs remains intense and that regulators should not freeze the market early. AI’s economic effects and market implications (Priority: 4/5): They connect AI to hyperscaler growth, margin expansion, construction booms, and strong public markets, while noting that the full ROI for end users and broad productivity gains has not yet fully shown up. Political backlash, safety rhetoric, and tech wealth distribution (Priority: 3/5): The discussion broadens to tech’s public legitimacy, with calls for more broad-based giving, Invest America-style accounts, better messaging on AI’s benefits, and concern that fear of concentrated wealth is driving regulation.

Key Arguments: AI demand is not the problem; compute and power supply are the binding constraints on revenue growth for frontier labs. Elon’s deal with Anthropic turns infrastructure into a revenue engine, reducing balance-sheet pressure on xAI and reinforcing SpaceX valuation. A pre-approval/FDA regime for AI would be premature, risk regulatory capture, and let Washington pick winners and losers before the market is mature. Narrower safeguards like cyber hardening, API monitoring, and KYC for preview access are more reasonable than full model approval. Anthropic and OpenAI are the only AI companies currently generating substantial frontier-model revenue, which is why their growth trajectories matter so much. AI is already contributing to GDP growth, cloud revenue, and operating-margin expansion, even if the full end-user productivity payoff has not yet been proven. Public concern about AI is driven less by AI’s salience than by fears of job loss, concentrated wealth, and weak communication from tech leaders about societal benefits. The market reaction to AI should be viewed as a signal that companies must demonstrate durable ROI, not just capex and model hype.

Data Points: Anthropic ARR (Jan-Mar 31): ~$10 billion to $30 billion - Sachs cites Anthropic tripling in Q1 as evidence of extreme growth. Anthropic ARR (April): $30 billion to $44 billion - Used to argue growth accelerated further after Q1. Anthropic projected year-end ARR: ~$100 billion - Sachs projects 10x growth exiting the year. Anthropic future projection: $1 trillion ARR in 2027 - Discussed as a possible extreme outcome if trajectory continues. XAI compute at Colossus 2: More than enough compute - Hosts say the lease gives xAI ample training capacity. Anthropic GPU expansion: 220,000+ NVIDIA GPUs - Mentioned as additional compute recently added. Anthropic energy expansion: 300+ megawatts - Tied to its new compute footprint. Nine-gigawatt buildout: ~9 GW expected online this year - Brad says about half is now being protested or delayed. Elon Web Services incremental revenue: $4-5 billion this year - Brad estimates lease revenue on top of prior analyst numbers. SpaceX/Elon infrastructure facilities: 3 facilities: Colossus, MacroHard, MacroHarder - Brad references multiple compute sites and their power mix. Hyperscaler run rates: AWS $150B, Azure $108B, GCP $80B - Discussed as evidence of massive cloud demand and market opportunity. Cloud growth rates: AWS 28%, Azure 39%, Google Cloud 63% - Used to illustrate strong AI/cloud demand. Cloud edit figures: AWS 10, Azure 9.5, GCP 10 - Referenced as quality-of-growth metrics or shorthand ratings. S&P 500 operating margins: ~11.8% (Q1 2024) to 13% (current year) - Brad cites margin expansion to argue efficiency gains are real. Public market valuations: Meta 17x, NVIDIA 19x, Microsoft 20x, Google 24x - Used to argue markets are not in obvious bubble territory. Index performance: S&P 500 up 8% YTD - Cited to argue the market is not euphoric despite strong AI. AI contribution to GDP growth: 75% of Q1 GDP growth - David Sachs attributes much of Q1 GDP expansion to AI. Labor market: US unemployment ~4.2% - Used to argue AI is not yet causing broad job losses. Labor force participation: 63.3% vs 61.9% pre-COVID - Mentioned to contextualize labor-market conditions. AI salience in polling: 29th out of 30 - AI is unpopular but not a top voter issue. Cost of living / economy: Top 2 voter concerns - Used to argue AI’s deflationary effects could matter politically. Construction worker wage increases: 25% to 30% - Cited as a blue-collar boom tied to infrastructure buildout. Polymarket odds of Trump order on federal AI review: 21% chance - Mentioned during the FDA-for-AI discussion.

Pivotal Quotes: "The problem here is not lack of housing. It's an addiction issue and it's a mental illness issue." — Speaker discussing Spencer Pratt: Used to summarize Pratt’s argument on homelessness in Los Angeles. "This deal fixes that problem." — David Sachs: Sachs explains how Elon leasing compute to Anthropic solves xAI’s capex/revenue mismatch. "I call it the biggest monopoly in human history." — David Sachs: Sachs argues Anthropic’s trajectory could create unprecedented market power.

Implications: Listeners should expect the AI race to hinge on power, compute, and distribution—not just model quality. Policy will likely focus on cyber safeguards, KYC, and coordination, while markets will reward firms that can prove real ROI and societal benefits.