Biggest LBO Ever, SPAC 2.0, Open Source AI Models, State AI Regulation Frenzy

Chamath and Friedberg argue that AI will enhance gaming far more than social or traditional media because games allow dynamic, interactive feedback loops. Examples include AI-driven bots in Fortnite that are tuned to player skill to reduce churn and increase engagement. With ~3 billion daily active users playing games globally, they see gaming as the primary ‘time sink’ of the future and a logical place for sovereign capital like Saudi PIF to bet on AI-driven entertainment growth.

Chamath’s bull case: as an 800-pound gorilla in gaming, EA can use private ownership to restructure OpEx, adopt next-gen AI tools, and reduce dependency on distribution gatekeepers like Xbox and PlayStation—especially as those platforms hike prices. If EA can build direct distribution and deeply integrate AI (for personalization, asset creation, and smarter game loops), he sees it as a potential multi-hundred-billion-dollar asset. The bear case is a world where cheap AI tools commoditize game creation and distribution via social platforms, undermining incumbent IP.

Chamath traces the explosion of private equity to the 60/40 portfolio era, zero interest rates, and leveraged buyouts that initially generated strong returns. As laggard funds flooded in, overpaying for assets and under-managing them, the industry’s AUM hockey-sticked to ~$5T while DPI (cash back to LPs) stalled. He expects capital to concentrate in a few elite PE firms (like Silver Lake) and leak into the next bubble, private credit, while warning LPs to focus on distributions, not IRR.

Chamath’s new SPAC structure strips out founder shares and warrants entirely and only pays the sponsor if the stock rises at least 50%, 75%, and 100% over the issue price (Elon-style milestone comp). He expects to focus on resilient, later-stage businesses and sees a future ‘Raptor 3.0’ where a sponsor pre-wires large common-stock PIPE capital (e.g., $1–3B) to deliver a fully-baked, fair-priced IPO alternative. He explicitly tells retail investors to avoid his SPAC (or cap at <1% of their portfolio), emphasizing that SPACs behave like venture bets with high loss rates.

DeepSeek 3.2 EXP and Moonshot’s Kimi offer dramatically cheaper token pricing (on the order of 10–35x cheaper than leading Western frontier APIs), partly via innovations like DeepSeek Sparse Attention. Chamath explains that EightyNinety now routes many workloads through Groq, which runs forked Chinese open-source models in US data centers—no data going back to China. While this slashes costs, it creates a coordination problem: every time a new model leapfrogs others, engineering teams must decide whether to refactor their stacks, because hot-swapping models is non-trivial.