Skip to main content
All-In with Chamath, Jason, Sacks & Friedberg

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

82 min episode · 3 min read
·

Episode

82 min

Read time

3 min

Topics

Investing, Artificial Intelligence

AI-Generated Summary

Key Takeaways

  • Elon Web Services Revenue Model: XAI's lease of Colossus 1 to Anthropic generates an estimated $4–5B in incremental annual revenue, offsetting massive CapEx commitments and removing pressure on Grok to produce immediate returns. This positions SpaceX as a fourth hyperscaler competing directly against AWS ($150B run rate), Azure ($108B), and GCP ($80B), with Elon retaining newer Blackwell GPU clusters for proprietary model training.
  • Anthropic Growth Trajectory: Anthropic expanded ARR from $10B to $30B in Q1 2025, then accelerated to $44B in April alone — roughly 10x annualized growth. The sole constraint identified is compute and power supply, not demand. If the trajectory holds for 18 more months, the panel projects Anthropic could reach $1T ARR, potentially exceeding the combined revenue of Apple, Microsoft, Alphabet, Amazon, Nvidia, and Meta.
  • AI Regulatory Capture Risk: Sacks frames safety-focused AI regulation as a potential Rockefeller "Safe Oil" strategy — where dominant incumbents advocate oversight frameworks that entrench their market position. Proposed pre-release model approval regimes would require Washington sign-off before deployment, functionally blocking smaller competitors. The panel argues specific, targeted cybersecurity coordination is legitimate; blanket approval regimes are not.
  • Cyber-Capable Models and KYC: Anthropic's Mythos and an equivalent OpenAI model now possess advanced cyber capabilities. Within three to six months, all major frontier labs — including Chinese models — will have comparable tools. The panel recommends mandatory Know Your Customer identity verification for API access to frontier models during preview periods, with government-industry coordination to harden systems before adversarial actors exploit the same capabilities.
  • Hyperscaler Revenue Signals Market Direction: AWS grew 28%, Azure 39%, and Google Cloud 63% year-over-year in Q1 2025, collectively adding $30B in quarterly revenue. Brad Gerstner shifted his portfolio from medium to large AI exposure in December when private market data confirmed enterprise revenue was materializing. Memory stocks — SK Hynix at 5x earnings, Samsung at 6x, Micron at 7x — remain undervalued relative to the infrastructure buildout they support.

What It Covers

XAI leases Colossus 1 data center to Anthropic, adding 220,000 NVIDIA GPUs as Anthropic grows from $10B to $44B ARR in four months. The panel debates whether Anthropic is becoming a historic monopoly, analyzes a proposed FDA-style AI oversight regime, and assesses cloud hyperscaler growth driving broader market gains.

Key Questions Answered

  • Elon Web Services Revenue Model: XAI's lease of Colossus 1 to Anthropic generates an estimated $4–5B in incremental annual revenue, offsetting massive CapEx commitments and removing pressure on Grok to produce immediate returns. This positions SpaceX as a fourth hyperscaler competing directly against AWS ($150B run rate), Azure ($108B), and GCP ($80B), with Elon retaining newer Blackwell GPU clusters for proprietary model training.
  • Anthropic Growth Trajectory: Anthropic expanded ARR from $10B to $30B in Q1 2025, then accelerated to $44B in April alone — roughly 10x annualized growth. The sole constraint identified is compute and power supply, not demand. If the trajectory holds for 18 more months, the panel projects Anthropic could reach $1T ARR, potentially exceeding the combined revenue of Apple, Microsoft, Alphabet, Amazon, Nvidia, and Meta.
  • AI Regulatory Capture Risk: Sacks frames safety-focused AI regulation as a potential Rockefeller "Safe Oil" strategy — where dominant incumbents advocate oversight frameworks that entrench their market position. Proposed pre-release model approval regimes would require Washington sign-off before deployment, functionally blocking smaller competitors. The panel argues specific, targeted cybersecurity coordination is legitimate; blanket approval regimes are not.
  • Cyber-Capable Models and KYC: Anthropic's Mythos and an equivalent OpenAI model now possess advanced cyber capabilities. Within three to six months, all major frontier labs — including Chinese models — will have comparable tools. The panel recommends mandatory Know Your Customer identity verification for API access to frontier models during preview periods, with government-industry coordination to harden systems before adversarial actors exploit the same capabilities.
  • Hyperscaler Revenue Signals Market Direction: AWS grew 28%, Azure 39%, and Google Cloud 63% year-over-year in Q1 2025, collectively adding $30B in quarterly revenue. Brad Gerstner shifted his portfolio from medium to large AI exposure in December when private market data confirmed enterprise revenue was materializing. Memory stocks — SK Hynix at 5x earnings, Samsung at 6x, Micron at 7x — remain undervalued relative to the infrastructure buildout they support.
  • AI ROI Fork in the Road: S&P 500 operating margins expanded from 11% in 2023 to 13% in 2025, but the panel debates whether AI or post-COVID headcount reductions drove the gains. Chamath identifies a roughly 500-day window before enterprises must demonstrate traceable ROI — spending X on tokens and recovering Y in margin expansion or revenue growth. Until that proof point appears in GDP, productivity statistics, or S&P earnings, the AI investment cycle remains unvalidated at scale.

Notable Moment

Sacks constructed an extended analogy comparing Anthropic to Standard Oil — arguing that if Rockefeller had rebranded kerosene as "Safe Oil" and lobbied for a safety regulator, the public would have missed the monopoly being built. He applied the same logic directly to AI safety rhetoric, suggesting it functions as cover for regulatory capture.

Know someone who'd find this useful?

You just read a 3-minute summary of a 79-minute episode.

Get All-In with Chamath, Jason, Sacks & Friedberg summarized like this every Monday — plus up to 2 more podcasts, free.

Pick Your Podcasts — Free

Keep Reading

More from All-In with Chamath, Jason, Sacks & Friedberg

We summarize every new episode. Want them in your inbox?

Similar Episodes

Related episodes from other podcasts

Explore Related Topics

This podcast is featured in Best Tech Podcasts (2026) — ranked and reviewed with AI summaries.

Read this week's Investing & Markets Podcast Insights — cross-podcast analysis updated weekly.

You're clearly into All-In with Chamath, Jason, Sacks & Friedberg.

Every Monday, we deliver AI summaries of the latest episodes from All-In with Chamath, Jason, Sacks & Friedberg and 192+ other podcasts. Free for up to 3 shows.

Start My Monday Digest

No credit card · Unsubscribe anytime