Palo Alto Networks has agreed to acquire Applied Compute in an all-stock transaction valued at $4.5 billion, subject to adjustments. Applied Compute provides post-training and inference infrastructure for open-weight models. The buyer says the combination will support enterprise deployment across public and sovereign clouds while bringing Applied Compute's research team into its wider AI portfolio.
The transaction is subject to customary conditions, including regulatory clearance, and is expected to close in Palo Alto Networks' fiscal third quarter of 2027. Until then, it is an agreement rather than a completed acquisition. The filing says Applied Compute co-founder and chief executive Yash Patil would become a co-chief AI officer at Palo Alto Networks after closing.
Strategically, the deal links security with the model operations layer. Enterprises choosing open-weight systems often want greater control over data, tuning, cost and deployment location, but those choices also transfer more responsibility for model provenance, access, post-training integrity and runtime monitoring. Palo Alto Networks is arguing that these controls can become part of a security platform rather than a separate engineering stack. The valuation and product benefits are company claims, and integration risk remains. Even so, the scale of the proposed purchase shows that secure model adaptation and serving are becoming a material competitive market in their own right.
