Palo Alto Networks to Acquire Applied Compute for $4.5B
First reported by Investors.paloaltonetworks ·
The cost to securely train and deploy custom AI models drops significantly.
Palo Alto Networks announced its definitive agreement to acquire Applied Compute for $4.5 billion in an all-stock transaction. The acquisition aims to integrate Applied Compute's platform and research team into Palo Alto Networks, enhancing its capabilities for enterprises to securely train, adapt, and deploy open-source and open-weight AI models. The company believes the future of enterprise AI is hybrid, with a significant portion of workloads running on open-weight models tuned to specific enterprise needs. Applied Compute's platform focuses on the post-training and inference stages of these models, offering security features for public and sovereign cloud environments. The deal is expected to close in Palo Alto Networks' fiscal third quarter of 2027, subject to customary closing conditions and regulatory clearance. Yash Patil, Applied Compute's co-founder and CEO, will assume the role of Co-Chief AI Officer at Palo Alto Networks.
The acquisition signifies a strategic move by Palo Alto Networks to solidify its position in the burgeoning enterprise AI security market, specifically addressing the growing demand for secure open-weight AI model management. By integrating Applied Compute's specialized technology and talent, Palo Alto Networks aims to offer a comprehensive solution that bridges the gap between powerful, customizable open-weight models and enterprise-grade security requirements, catering to a hybrid AI future.
This integration will likely accelerate the adoption of AI by enterprises that prioritize data privacy and control, enabling them to leverage open-source AI without compromising their security posture. The move signals increased competition in the AI security space, prompting other cybersecurity and cloud providers to enhance their offerings for secure AI model development and deployment.
AI-written summary. May contain errors.