Key Takeaways from the 2025 State of AI in the Cloud Report

**State of AI in the Cloud (Executive Summary):** Wiz Research reports accelerating AI adoption—notably a jump in self-hosted models—and identifies significant security risks, including a discovered exposure of a DeepSeek database that leaked chat history and logs, a critical NVIDIA vulnerability (CVE-2024-0132) affecting over 35% of cloud environments, an RCE vulnerability in Ollama (Probllama), and other AI PaaS flaws across vendors like Hugging Face, Replicate, and SAP; the report urges improved AI security posture management to mitigate data exposure and governance risks.