This newsletter covers SpaceX’s $6.3 billion AI compute contract, a new exploit targeting Cisco devices, and Microsoft’s push for AI-driven cloud observability agents. It also highlights ongoing Linux network‐share headaches, the role of LLMs as software front ends, and the link between AI adoption and security incidents.

• SpaceX signed a $6.3 billion deal for custom AI compute hardware to power in-house model training and inference.
• Attackers are actively exploiting a newly disclosed Cisco IOS vulnerability to deploy ransomware on enterprise routers.
• Microsoft and others are moving toward "agentic observability," where AI agents triage anomalies and suggest fixes instead of relying on static dashboards and alerts.
• A survey found heavier enterprise AI use correlates with more security incidents, pointing to a need for access controls and governance before deploying AI in critical workflows.