Palo Alto Networks has acquired the American startup Applied Compute for $4.5 billion, paid entirely in stock. The move was announced alongside quarterly results that beat expectations and is intended to accelerate the group’s enterprise AI strategy within cybersecurity.
Applied Compute has built an AI infrastructure platform that enables companies to train, deploy, and continuously improve their own AI models and agents, with scalable hosting and inference capabilities.
Why Applied Compute?
Palo Alto Networks observes that enterprise AI demand is rapidly shifting toward open models that can be deployed within customer environments. This shift creates a need for security, governance, and dedicated infrastructure.
The acquisition aligns with Prisma AIRS, the company’s platform dedicated to securing AI applications, models, and agents. It posted roughly $170 million in ARR in the first quarter, with more than 1,100 customers.
Palo Alto notes that Applied Compute, founded less than two years ago, generates approximately $100 million in annualized revenue and employs about 70 people, all based in the United States.
Beyond software, the acquisition also opens an industrial agenda. Palo Alto plans to build a global AI inference capacity of more than 100 MW over a three-to-five-year period. The objective is to offer clients an infrastructure that enables them to deploy open AI models with a level of security and sovereignty tailored to their constraints.