Palo Alto Networks has signed a definitive agreement to buy Applied Compute, a platform specialising in post-training and inference for open-weight AI models, in a move designed to give enterprises a secure path to adopting AI regardless of which models they rely on.
The deal will see Applied Compute’s technology and its AI research team move under the Palo Alto Networks umbrella. The cybersecurity group intends to use the platform to let organisations train, customise and deploy open-weight models with protection embedded from the outset, whether those workloads sit in the public cloud or in sovereign environments where compliance demands are stricter.
The acquisition rests on a clear thesis. Palo Alto Networks expects enterprise AI to become a hybrid landscape, with many organisations turning to open-source and open-weight models that they can shape around their own data, budgets and governance requirements. According to the company, no major provider currently offers these models with security at their core, a gap it now aims to fill.
Applied Compute will keep serving its existing base of AI-native firms, digital natives and large corporates, while Palo Alto Networks will introduce the platform to customers worldwide.
Applied Compute builds tools that help organisations refine and run open-weight models after their initial training, covering both the post-training stage and live inference.
Beyond the platform itself, Palo Alto Networks is gaining a research group it regards as among the strongest in the sector. These specialists will work on embedding AI throughout the company’s product portfolio and on training its in-house models of every size.
Source link







