Palo Alto Networks has signed a definitive agreement to acquire Applied Compute in an all-stock transaction valued at $4.5 billion, subject to adjustments. The cybersecurity company says it wants to use the startup’s platform to help enterprise customers train, adapt and run open-weight AI models with security built into the process. The deal is still pending, and the planned combined offering is not a generally available product today. Palo Alto Networks’ announcement, Reuters’ report and The Wall Street Journal’s coverage confirm the agreement and its broad strategic aim.

What does Applied Compute do?

Applied Compute focuses on post-training and inference for open-weight AI models. Post-training is the work of adapting a model after its initial training; inference is running the model to produce outputs. The Wall Street Journal reports that the San Francisco startup helps businesses train and run custom models using their own data and was founded by former OpenAI researchers. WSJ’s report describes that product focus. Palo Alto says Applied Compute will continue serving its existing customers, while it plans to bring the platform to its wider enterprise base, according to the company announcement.

In practice, the acquisition is a bet that businesses will want more control over how models are tuned and where they run. Palo Alto says the platform could support public-cloud and sovereign-cloud deployments. Those are future plans, not proof that the acquisition has already delivered a new security product or measurable risk reduction.

Why are open-weight models part of a security deal?

Open-weight models make trained parameters available so users can download, adapt or run a model themselves. That can give organizations more flexibility over cost, customization and deployment location. But “open-weight” is not automatically the same as fully open-source AI. The Open Source Initiative’s definition says an open-source AI system needs more than weights alone: its preferred form for modification also includes information about training data and the code used to build and run the system, alongside the parameters.

Running models inside an organization can also create security questions around model tampering, data leakage and misuse. Reuters identifies these as risks associated with open-weight deployments. Palo Alto’s stated aim is to help protect the training, adaptation and deployment process, but the announcement does not set out independent test results, specific security guarantees or detailed product controls.

What are the deal terms?

The announced $4.5 billion consideration is in stock and remains subject to adjustments. The companies still need to meet customary closing conditions, including regulatory clearance. Palo Alto expects the transaction to close in its fiscal third quarter of 2027. After closing, Applied Compute co-founder and CEO Yash Patil is expected to become Palo Alto Networks’ Co-Chief AI Officer, according to the company announcement.

The announcement arrived alongside Palo Alto Networks’ fiscal first-quarter results. The company reported revenue of $3.44 billion, up 39% year over year, and raised its full-year fiscal 2027 revenue outlook to $14.52 billion–$14.57 billion. These figures provide business context for the acquisition; they do not show that the proposed Applied Compute integration has already contributed to revenue. See the official Q1 results release and Reuters’ independent coverage.

What should enterprise buyers watch next?

The practical test will be how the combined platform handles model provenance, access controls, data boundaries, software updates and independent security testing. Palo Alto has not yet published detailed integration milestones, product availability dates or customer pricing for the planned offering. Until the deal closes and those details emerge, the acquisition should be read as a strategic move—not as evidence that open-weight AI is now secure by default.

Frequently asked questions

How much is Palo Alto Networks paying for Applied Compute?

The announced transaction is an all-stock deal valued at $4.5 billion, subject to adjustments. It has not closed. See the company announcement.

What is Applied Compute’s platform for?

It provides technology for post-training and inference of open-weight AI models. Palo Alto plans to use it to help enterprise customers adapt and deploy those models with security features.

When is the acquisition expected to close?

Palo Alto Networks expects a close in its fiscal third quarter of 2027, subject to customary conditions including regulatory clearance.

Are open-weight models the same as open-source AI?

No. The availability of model weights alone does not establish that an AI system meets the Open Source Initiative’s definition, which also addresses training-data information and code.