AI second-wave reshapes business models and execution

AI commoditization and reliability challenges are forcing a shift from model size to system design. Key signals: Anthropic profitability, DeepSeek pricing pressure, 95% AI pilot failure rate, and new frameworks for system-centric AI (separation of reasoning from execution). The move from training to deployment (DigitalOcean 800% inference growth) raises commoditization vs. differentiation questions. New today: DeepSeek's open-source thesis with 545% profit margins and Tencent's AI infrastructure prioritization add concrete cases for business model and capital allocation decisions. Also: IBM IBV EASY framework (Eliminate, Automate, Standardize, Yield) warns against efficiency-only AI and advocates reinvesting gains into innovation. New: TeraWulf Q2 shows 71% HPC leases, signaling business model shift from spot mining to recurring revenue in AI/crypto.