
AI-generated editorial illustration of secure enterprise model training and inference.
Palo Alto Networks has agreed to buy Applied Compute for $4.5 billion in stock. The exact sum can change. The plan brings tools to train and run open-weight AI models.
The November 18 deal release sets out the terms. The buyer expects to close in its fiscal third quarter of 2027. It needs the go-ahead from regulators. Other closing terms also apply.
The plan for Applied Compute
Applied Compute helps firms adapt a model for a task. It also runs models to answer users' requests. The buyer plans to offer the tools to its global clients.
The business will keep serving the same groups of clients. They include AI firms, digital businesses and large enterprises.
CEO and co-founder Yash Patil will become co-chief AI officer. That change would take place after the deal closes. His research team will also help train the buyer's own models.
Train, test and run a model
The Applied Compute platform offers more than one way to work. Teams can run their own jobs. They can work with the firm's researchers. A managed service is another option.
The firm says clients keep the model weights. These files hold the values that a model learns.
Its AC2 tools link training, tests and the live model. Teams can bring their own agent tools and task setup. This lets them train in a way that matches real use. Data from that use can feed the next training run.
Our Local AI guide explains where open models can run. They can use a PC, private servers or cloud tools.
The price includes prepaid compute
The buyer's earnings slides add the figures. The deal includes about $500 million in cash and prepaid compute commitments.
They list about 70 staff, all in the US. Applied Compute reached a sales pace of $100 million a year. That came within 18 months of its start. The figure projects the current pace across a full year.
Our local and cloud AI guide compares cost, control and upkeep.
Read also
Popular news

Anthropic opens at $111.50 on Nasdaq, 34.34% above IPO price

GAO tests found safety gaps in four major AI chatbots

Nvidia revenue doubles to $115.8 billion as Rubin ships

Your AI Model Hasn’t Changed. The Company Running It Might Have.

Nvidia RTX Spark mini PCs bring up to 128GB memory to local AI

