Hello. Securing AI agents, or using AI agents to secure other systems, has captured the attention of investors who are pouring money into such startups. 

Noma Security, a New York-based startup that polices artificial intelligence agents inside corporate networks, raised $100 million in investment capital, the company is set to announce Thursday. That's less than a year after snagging $32 million. Read our story. 

Scroll down for a handful of other new raises for AI-related security firms.  

Also today: We talked to the head of cybersecurity at Columbia University about how the institution defends itself amid intense activity from hackers. Read our Q&A.

More news below.

• Microsoft has attributed early attacks to three China-linked groups. 
• The nonprofit Center for Internet Security, an information-sharing group for governments, said about 90 state and local governments have been targeted but none successfully hacked. (Reuters)

Voice fakes get hackers inside. The ShinyHunters hacking group is using deepfake audio in exploits against Salesforce customer relationship management systems, according to Google's threat intelligence group. ShinyHunter members are posing as tech staff to trick employees at companies that use the Salesforce product to turn over credentials. 

• Salesforce noted that its software hasn't been hacked; rather, Salesforce accounts were breached through social-engineering tactics. (Bleeping Computer)

German companies would have 24 hours to report cyberattacks to the country's Federal Office for Information Security under new guidelines proposed Wednesday. An update would be due within 72 hours, with a more detailed report required after one month. (DPA)

• Safe Security, which focuses on quantifying cyber risk, plans to announce Thursday that it has raised $70 million in a Series C round led by Avataar Ventures. The company's prior round, in April 2023, was $50 million.
• San Francisco-based BlinkOps said Wednesday it closed a $50 million Series B round led by O.G. Venture Partners. BlinkOps provides tools for building AI agents to perform security tasks such as managing identity access.  
• Legion, a startup formed last year, has raised $38 million in seed and Series A funding led by Coatue. Legion builds AI agents for improving cybersecurity investigations and managing alerts. (SecurityWeek)

The WSJ Pro Cybersecurity team is Deputy Bureau Chief Kim S. Nash and reporters Angus Loten, James Rundle and Catherine Stupp. Follow us on X @WSJCyber. Reach the team by replying to any newsletter you receive or by emailing Kim at kim.nash@wsj.com.