Palo Alto Networks opens wallet for open-weight training biz
Security shop buys the hand that feeds it
Palo Alto Networks (PAN) has agreed to buy Applied Compute, an AI model training platform, in order to bolster security for corporate customers in the face of AI-related threats. PAN CEO Nikesh Arora stated that many recent disruptions, such as outages and cyberattacks, were partially caused by inadequate patching and configuration changes due to the inability to quickly address vulnerabilities without risking production environments.
He also noted the rise of AI-assisted cyberattacks as a factor in the company's decision to acquire Applied Compute. PAN believes that the future of enterprise AI involves a hybrid approach, with a significant portion of enterprises running on open-source and open-weight models tailored to their specific needs for data, cost, and control.
Despite concerns about the loss of lobbying power for open-weight models following Nvidia co-founder Jensen Huang's resignation, PAN remains optimistic about the future of open-weight models. The company expects the $4.5 billion all-stock transaction to add over $1 billion to its Next Generation Security business by FY 2029. Arora believes that the corporate desire for sovereignty and control is driving enterprises toward open-weight and open-source models, with open-weight models growing from under 5% of token traffic to over 60% in the past year.
PAN also plans to build infrastructure capable of supporting more than 100 megawatts of inference capacity, a project with a modest capital expenditure commitment of less than $3 billion through fiscal 2030.
Written by urgent.news from The Register's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
This story
This is one outlet's version. Read the fullest account.
- Palo Alto Networks opens wallet for open-weight training biz theregister.com