Moves by the security industry are accelerating to offer commercial services that use large language models to help public and private organisations detect software vulnerabilities. IBM and OpenAI unveiled an enterprise AI security service, while SoftBank Group launched a product based on OpenAI models. OpenAI updated GPT-5.5-Cyber and expanded partner programmes. Intelligence agencies warned of destructive AI cyber attacks. Companies worldwide also announced new tools, acquisitions, funding and partnerships, including PQC standardisation efforts in South Korea.

Cybersecurity startup Ent has raised $100 million in seed funding. The company has developed an “intent-aware workspace security” platform that analyzes the behavior of employees, AI agents and applications. Hosted in a customer-managed cloud, a lightweight endpoint agent observes activity across applications, browsers, AI workflows, data movement and local software execution. Ent says it can intervene before incidents by analyzing behavioral context.