Palo Alto Networks buys Applied Compute for $4.5 billion
The chief executive tells that the new business can be larger than cybersecurity, if it works.
Palo Alto Networks, a US cybersecurity firm, agreed to buy Applied Compute, a startup that helps companies operate and customize their AI models. The deal is $4.5 billion, all in stock. The firm announced it with results for its first fiscal quarter: revenue of $3.44 billion, up 39%, and shares up more than 7%, after it beat the estimates of analysts.
What it buys
The chief executive, Nikesh Arora, tells that the deal helps businesses make open weight models safe as more firms use them. Open weight models are AI models that firms can operate by themselves. The startup marks a large investment in a field other than cybersecurity.
The numbers
Annual recurring revenue (income that comes again each year) of its next generation business is $9.73 billion, 66% more. Shares are up 85% this year. Customers want more cyber tools, because they think AI cyberattacks are a threat, as in the Hugging Face breach.
What is not known
The firm gave no financial data and no test results for Applied Compute. Arora puts a condition on the claim: if it works.
Insight: A cybersecurity firm now wants to supply AI
In October 2026 Palo Alto Networks told that AI agents in a company are an insider risk. In November 2026 it announced a sensor for Claude Cowork, a desktop agent. It will pay $4.5 billion for a firm that operates models. This reads as a move from protection of AI to supply of AI, and it seems to want a position where models operate.
Sources
Posted