Palo Alto Networks agrees to buy Applied Compute for 4.5 billion dollars in stock
US cybersecurity company Palo Alto Networks signed an all-stock agreement to buy Applied Compute, a US startup that helps businesses adapt and run existing AI models. If the deal clears regulatory review, the startup’s researchers would join the buyer. Palo Alto Networks plans to make Applied Compute’s tools available to its corporate customers.
Artificial Intelligence··Midday
Palo Alto Networks would pay in shares
US cybersecurity company Palo Alto Networks signed an agreement on November 18 to buy Applied Compute, a US startup that makes software for adapting and running artificial intelligence models. The deal is valued at 4.5 billion dollars, subject to closing adjustments, and would be paid entirely in shares. Palo Alto Networks expects it to close in its third fiscal quarter of 2027, subject to regulatory clearance and other closing conditions.[1], [2]
Businesses could adapt models to their own data
Applied Compute makes tools for companies to start with an open-weight model, whose learned settings are available for others to use, and adjust it for their own data or tasks. Its software also runs the adjusted model and can vary the computing resources it uses as demand changes. Palo Alto Networks plans to offer the tools for use in public clouds and in environments that keep data under local controls.[1]
Applied Compute’s researchers would join the buyer
Former OpenAI researchers Yash Patil, Rhythm Garg and Linden Li founded Applied Compute in 2025. Patil, its chief executive, would become one of Palo Alto Networks’ two chief AI officers after closing. The buyer says the incoming research team would also work on models it develops in-house. Applied Compute is expected to continue serving its existing customers.[1]