Categories
Corporate

Palo Alto Networks makes $4.5 billion AI bet

Acquisition aims to secure enterprise AI models as businesses expand open-weight deployments

Palo Alto Networks has agreed to acquire artificial intelligence startup Applied Compute in an all-stock deal valued at $4.5 billion, marking a major push by the cybersecurity company into enterprise AI development and security. The acquisition is aimed at helping businesses train, customise and deploy AI models securely as organisations increasingly adopt artificial intelligence for their operations.

Announced on November 18, the deal will bring Applied Compute’s technology platform and AI research team into Palo Alto Networks. The company plans to combine its cybersecurity capabilities with Applied Compute’s expertise in training and running open-weight AI models, allowing enterprises to build and operate customised AI systems with security features built into the process.

The acquisition comes at a time when businesses are moving beyond simply using ready-made AI tools. Companies are increasingly looking to develop models using their own data, business information and operational requirements. However, this shift also raises concerns around data privacy, unauthorised access, model manipulation and the security of AI-generated outputs.

Palo Alto Networks believes that the future of enterprise AI will involve a combination of open-weight and closed AI models. Its proposed acquisition is intended to address the growing need to secure both types of systems, particularly as businesses seek greater control over how their AI applications are developed and deployed.

Applied Compute specialises in post-training and inference of open-weight AI models. Post-training involves refining an existing AI model to make it more suitable for specific tasks, while inference refers to the process through which a trained model generates responses or predictions when given new information.

Open-weight models allow developers and businesses to access and modify the underlying model parameters, making it possible to customise AI systems for specific applications. Unlike closed models, which are generally accessed through controlled platforms, open-weight models offer organisations greater flexibility over deployment, costs and data management. However, this flexibility also creates additional cybersecurity challenges.

Through the acquisition, Palo Alto Networks wants to provide enterprises with the tools to train, adapt and deploy these models securely across public cloud infrastructure and sovereign cloud environments. Sovereign cloud deployments allow organisations to retain greater control over where their data is stored and processed, an important consideration for businesses operating under strict regulatory and data protection requirements.

The company also plans to use Applied Compute’s research capabilities to strengthen its own artificial intelligence development. Its researchers and engineers will help integrate AI across Palo Alto Networks’ products, including the development and training of its proprietary large and small AI models.

Applied Compute co-founder and chief executive Yash Patil will become Palo Alto Networks’ Co-Chief AI Officer after the transaction is completed. The startup’s research and engineering team will also join the cybersecurity company, adding specialist talent to its expanding AI operations.

Palo Alto Networks chairman and chief executive Nikesh Arora said enterprises would increasingly use a mix of open and closed AI models, making security essential across both categories. He said Applied Compute had developed a strong platform for post-training and inference and that its research team would help the company support secure AI deployment in public cloud and sovereign environments.

Founded by Yash Patil, Rhythm Garg and Linden Li, Applied Compute has grown rapidly since its launch. The startup reportedly achieved annualised revenue of around $100 million within approximately 18 months of its founding, highlighting the growing commercial demand for enterprise AI infrastructure.

The company has also attracted significant investor interest. Earlier in November, it raised $350 million in funding at a valuation of approximately $3.25 billion. Investors included Nvidia and other prominent technology and venture capital backers. Its customer base includes businesses looking to train and operate proprietary AI models using their own data.

Palo Alto Networks‘ investor presentation indicated that Applied Compute has around 70 employees, all based in the United States. The company expects the acquisition to contribute approximately $1 billion to its next-generation security annual recurring revenue by fiscal 2029, underlining the potential business opportunity it sees in enterprise AI.

The $4.5 billion transaction will be completed through shares of Palo Alto Networks, subject to adjustments and customary closing conditions, including regulatory approval. The company expects the acquisition to close in the third quarter of its fiscal 2027. Its regulatory and integration requirements will determine the final completion timeline.

The acquisition also reflects a broader shift in the cybersecurity industry. As artificial intelligence becomes part of everyday business operations, cyber threats are becoming more sophisticated. AI tools can help organisations identify vulnerabilities and strengthen their defences, but they can also be exploited to develop attacks, expose sensitive information and identify weaknesses in existing systems.

The company has been pursuing a broader expansion strategy through acquisitions and investments in cloud security and AI-related capabilities. Bringing Applied Compute into the business could help it deepen its presence in enterprise AI infrastructure while strengthening its position in the increasingly competitive cybersecurity market.