The AI industry is grappling with a stark reality check as Anthropic reveals widespread misuse of its Claude models by both state-sponsored hackers and rival AI labs. Russian cyber actors used Claude to rebuild malware after detection, while seven China-based AI labs, including Alibaba and DeepSeek, executed industrial-scale distillation attacks to siphon the model's capabilities. The Anthropic report on automated exploitation underscores that frontier models are being weaponized faster than defenses can adapt, forcing the industry to confront the dual-use nature of its creations. Regulators are beginning to respond, though the approach remains fragmented. OpenAI quietly floated the idea of a shared industry-wide AI slowdown to Congress, asking whether coordinating on safety would violate antitrust law. Meanwhile, the EU's AI Act kicked in with Article 50, requiring AI agents to identify themselves to customers across Europe or face penalties. One of AI's fiercest critics argues that the constant focus on existential "doom" scenarios is a distraction from the concrete harms happening right now, such as automated cybercrime and undisclosed AI sales calls. Despite the alarms, investment in AI infrastructure continues at a blistering pace. Mecka AI is nearing a $500 million valuation from Sequoia to collect human motion data for robot training, while Nscale added former OpenAI executive Fidji Simo to its board ahead of a potential IPO. On the model side, Together AI expanded its fine-tuning service with live metrics and lower prices, and Baseten made DeepSeek-V4.1-Flash available with a million-token context. Enterprise adoption is also maturing, with OpenEvidence verifying Canadian physicians and NVIDIA using Palantir's Foundry for supply chain decisions. The week's events paint a clear picture: the AI race is not slowing down, and the tension between innovation, security, and regulation is reaching a critical inflection point. The next few months will test whether voluntary coordination, government mandates, or market forces can shape a safer trajectory.