Finn's Take· TL;DRPalo Alto Networks had a lot to celebrate this week. Shares jumped more than 7% after the cybersecurity firm beat Wall Street's fiscal first-quarter estimates, posting adjusted earnings per share of $1.05 against the 97 cents expected, with revenue of $3.44 billion versus the $3.31 billion forecast — a 39% jump from the year-ago quarter. But the earnings beat was almost secondary to the bombshell that accompanied it.
Alongside its earnings report, Palo Alto announced a $4.5 billion deal to buy Applied Compute, a startup that helps train customized open-weight AI models. It's one of several recent deals in CEO Nikesh Arora's aggressive acquisition spree and marks a significant shift beyond cybersecurity. The move sent a clear signal: Palo Alto isn't just defending the AI era — it wants to help build it.
To understand why investors are excited, it helps to understand what "open-weight" AI models actually are. Open-weight models are those that allow the user to adjust the model's biases, versus closed models — such as those from Anthropic and OpenAI — where the developer controls the behavior. Enterprises increasingly want that kind of control and ownership over their AI, and the numbers reflect it.
On the earnings call, Arora said, "We're already seeing the desire for sovereignty and control compel enterprises to adopt open-weight and open-source models, and early signs of that shift are clear, with open-weight models growing from under 5% of token traffic to more than 60% over the past year." That explosive growth, he argued, creates an enormous security vacuum — one Palo Alto intends to fill. The company expects "a future with a serving opportunity in the hundreds of billions of dollars of enterprise applications driven by open-weight and open-source adoption."
The acquisition brings a company that has already reached more than $100 million in annualized revenue just 18 months after its founding. Applied Compute develops technology for post-training and inference of open-weight AI models, allowing companies to adapt and deploy models in public and sovereign cloud environments. Arora said Palo Alto had evaluated nine providers before settling on Applied Compute.
With the acquisition, Palo Alto will add Applied Compute's platform and AI research team to help enterprises adopt AI securely, whether they choose open-weight models, closed frontier models, or both. The company also plans to develop over 35 AI-powered security applications and deliver sovereign open-weight models. Fears of AI-powered cyberattacks, heightened by incidents like the Hugging Face breach, are driving demand for cybersecurity vendors like Palo Alto Networks.
Palo Alto Networks sees its $4.5 billion acquisition of Applied Compute as the beginning of a business that could eventually surpass its cybersecurity operations — a bold claim from a CEO whose company was built on exactly that foundation. Arora didn't mince words on the earnings call. "If it works, I think this will dwarf the ability that we have in cybersecurity," he said. "This could be bigger than what we do."
Analysts at Bank of America touted the deal as a "bold move" to expand beyond securing AI into training open-weight models, allowing Palo Alto to lead the next phase of the AI revolution. The company's annual recurring revenue for its next-generation security business jumped 66% from a year ago to $9.73 billion, and Palo Alto bumped up guidance for fiscal year 2027, expecting adjusted earnings per share in the range of $4.26 to $4.30 and revenue between $14.52 billion and $14.57 billion. With AI threats escalating and enterprise adoption of open models accelerating, the cybersecurity giant appears to be positioning itself not just as a protector of the AI age — but as one of its architects.