Palo Alto Networks plans to release a proprietary large language model for security within a year as it embeds generative AI into its products and workflows, CEO Nikesh Arora said last week during the company’s earnings call.
The cybersecurity vendor, which uses other forms of AI across its services, is eyeing generative AI for its ability to improve detection and prevention, customer interactions with datasets and telemetry, and internal processes and operations.
The immediacy of the moment was abundant as Arora mentioned generative AI within the first two minutes of his prepared remarks during the company’s earnings call. He subsequently mentioned “AI” 32 times during the call.
While Palo Alto Networks is taking a more restrained approach to generative AI — it hasn’t released new products based on the technology as quickly as Microsoft or Google, for example — the company’s leaders are no less convinced generative AI will herald a significant shift in capabilities for the industry.
“I think we are going to undergo a transformation both at Palo Alto Networks as well as generally in the enterprise software industry over the next 12 months to 24 months as we embrace generative AI,” Arora said.
“That’s the real opportunity and challenge in front of us, and I think half of the people out there will get it wrong. And hopefully, we’re on the right side of history,” Arora said.
Palo Alto Networks plans to introduce a proprietary AI model, which the company will build and use specific security use cases and tasks, rather than self-training an LLM.
The vendor is working with every public and open-source model to understand how it can build these capabilities using proprietary data.
Generative AI benefits for security
The benefits of generative AI, which most enterprises have yet to fully realize, lie in its ability to quickly access and summarize data. But the size of that dataset from which it generates intelligence and responses is critical, according to Arora.
Generative AI favors companies that have large data lakes and the financial leverage to make significant investments, both of which Palo Alto Networks has and intends to put to use, he said.
The company has almost $2 billion in current cash assets.
Palo Alto Networks banked almost $108 million in net income on $1.7 billion in revenue during its fiscal third quarter of 2023, for the period ending April 30.
While the company is taking generative AI product releases slower than some peers in the cybersecurity market, Arora’s comments suggest the window of opportunity will close by mid-2025, which puts some pressure on the company to deliver new capabilities soon.
“As you’ve been hearing in all the conversations in the industry, we’re still working on it, we’re understanding it, we’re re-looking at processes,” Arora said, “but will believe there is a there there.”