Import AI 436: Another 2GW datacenter; why regulation is scary; how to fight a superintelligence
A curated research newsletter surfaces RAND's rogue-AI containment analysis and a regulation-as-vetocracy critique - both relevant to APS AI safety and governance thinking.
Key points
- Import AI #436 covers four distinct topics: AI agent training infrastructure, a 2GW datacenter in Saudi Arabia, regulation critique, and a RAND paper on countering rogue AI.
- The RAND analysis on countering a rogue superintelligence offers sobering conclusions relevant to AI safety policy discussions.
- The regulation critique reflects a pro-innovation perspective on regulatory burden - a counterpoint worth noting in AI governance debates.
Implications for Australian agencies
- Monitor AI safety and strategy teams may want to monitor the RAND rogue-AI paper for its implications on frontier risk framing in Australian AI policy discussions.
- Consider APS AI governance practitioners could consider the regulation-burden critique as a useful counterpoint when assessing proportionality in any proposed AI regulatory interventions.
Implications are AI-generated. Starting points, not advice — see methodology for how they're framed.
View original source
Copied.
"Import AI 436: Another 2GW datacenter; why regulation is scary; how to fight a superintelligence"
Source: Import AI – Substack (Jack Clark)
Published: 24 November 2025
URL: https://importai.substack.com/p/import-ai-436-another-2gw-datacenter
Jack Clark's Import AI #436 covers four threads. First, OSGym - a low-cost academic platform for training AI agents to operate across general computer environments at scale. Second, Luma AI's $900m raise and plans for a 2GW compute supercluster in Saudi Arabia (Project Halo), built with Saudi PIF-backed Humain. Third, a critique by entrepreneur Peter Reinhardt arguing that over-regulation doubles costs and stifles innovation - framed by Clark as a cautionary tale for AI governance advocates. Fourth, a RAND paper examining technical options for countering a rogue AI system, concluding that current options - including nuclear EMP, internet shutdown, and counter-AI tools - are unlikely to be effective, and that prevention is therefore essential.
Implications for Australian agencies:
- [Monitor] AI safety and strategy teams may want to monitor the RAND rogue-AI paper for its implications on frontier risk framing in Australian AI policy discussions.
- [Consider] APS AI governance practitioners could consider the regulation-burden critique as a useful counterpoint when assessing proportionality in any proposed AI regulatory interventions.
Retrieved from SIMS, 18 July 2026.