Descriptions:
Episode 255 of Moonshots with Peter Diamandis brings together hosts Peter Diamandis, Dave, Ismail, and Alex for a wide-ranging discussion of the week’s most consequential AI industry moves. The lead story is Anthropic’s takeover of SpaceX’s Colossus 1 data center in Memphis — a development notable not just for the infrastructure it unlocks (Claude Code rate limits were immediately doubled) but for what it signals politically: Elon Musk, a longtime Anthropic critic, effectively backing the company in what the panel describes as an “enemy of my enemy” dynamic against OpenAI.
The episode digs into Anthropic’s extraordinary growth figures, citing CEO Dario Amadei’s disclosure at the company’s recent developer conference that Anthropic experienced 80x growth in Q1 2026 — well beyond the projected 10x. Annualized revenue reportedly climbed from $9 billion at end of 2025 to $30 billion in April, with estimates already pushing past $40 billion in May and projections of $100 billion ARR by year-end. The panel also covers Leopold Aschenbrenner — famously fired from OpenAI’s alignment team — now running a $5.5 billion AI investment fund just two years later.
The group examines the rise of Hermes, an open-source agent written in Python that has surpassed OpenClaw as the top model by token volume on Open Router. Discussion centers on Hermes’s native recursive self-improvement architecture, its easier hackability compared to OpenClaw’s TypeScript base, and what that distinction means for the future of open-source agentic systems. The episode closes with topics ranging from AI alignment frameworks to UAP disclosures from the White House.
📺 Source: Peter H. Diamandis · Published May 16, 2026
🏷️ Format: Podcast
