Cybersecurity firm Palo Alto Networks will acquire AI model training and inference startup Applied Compute for $4.5 billion, SiliconANGLE reported on Tuesday. The deal will be all stock.
Applied Compute was founded in 2025 by former OpenAI researchers and provides a platform that helps companies further train and run open-weight models.
Earlier this month, it raised $350 million after being valued at $3.25 billion. At the time, its annual recurring revenue exceeded $100 million, and it had Microsoft and DoorDash among its customers, SiliconANGLE said.
Palo Alto expects corporate AI environments to evolve into a mix-and-match approach using multiple models as companies increasingly use open-weight models tuned with their own data.
Palo Alto CEO Nikesh Arora (니케시 아로라) said, "Every company will use a mix of open-weight and closed models, and every model needs security." He added, "Applied Compute built the best platform."
Palo Alto also reported fiscal first-quarter results for the period ended Oct. 31 that far exceeded market expectations. Quarterly revenue was $3.44 billion, up 39 percent from a year earlier. Adjusted earnings per share rose to $1.05 from 93 cents a year earlier.
Palo Alto raised its annual revenue forecast to $14.52 billion to $14.57 billion from $14.1 billion to $14.2 billion in September. It forecast adjusted earnings per share of $4.26 to $4.30, above analysts' estimate of $4.18. Palo Alto shares rose nearly 10 percent in regular trading and gained more than 8 percent in after-hours trade.