Eighteen months ago, Applied Compute didn’t exist. Now Palo Alto Networks is spending $4.5 billion in stock on it, and that is why “applied compute” is blowing up on Google tonight. The Applied Compute acquisition, announced Wednesday alongside Palo Alto’s earnings, is one of the biggest AI deals of the year.
Here’s the short version. According to the company’s SEC filing, Palo Alto signed a definitive agreement to buy Applied Compute, a platform for training and running open-weight AI models. The deal is all-stock, valued at $4.5 billion “subject to adjustments,” and expected to close in Palo Alto’s fiscal third quarter of 2027. Even so, regulators still have to sign off.

Who is Applied Compute, and what does Palo Alto get?
Founded in May 2025 by ex-OpenAI researchers Yash Patil, Rhythm Garg and Linden Li, Applied Compute has about 50 employees. BankInfoSecurity reports it hit a $100 million annualized revenue run rate earlier this month.
It isn’t a chip company. It’s software. The platform takes an open-weight model, retrains it on a company’s own data, and runs it on that company’s own cloud. Reported customers include Microsoft, Shopify, DoorDash and Harvey.
Now do the math on the price. Applied Compute raised $350 million at a $3.25 billion valuation on Nov. 2, led by Elad Gil with Nvidia in the round.
Sixteen days later, it sold for about $1.25 billion more. That’s our arithmetic, and it’s a pretty good two weeks.
Why Palo Alto wants an AI lab: security is the pitch
Palo Alto’s pitch is simple. Companies won’t run just one AI model. They’ll run a mix of open and closed ones, and every one has to be protected. CEO Nikesh Arora put it this way: “Every enterprise is going to run a mix of open and closed models, and every one of those models needs to be secured.”
Palo Alto says no large provider delivers open-weight models securely today. So buying the team that builds them is how it plans to fill that gap. Patil becomes Palo Alto’s Co-Chief AI Officer once the deal closes, and Palo Alto expects the Applied Compute buyout to add more than $1 billion in next-gen security recurring revenue by 2029.
Security is a hot target for attackers, too. For a sense of how fast flaws can spread, see these two critical PHP flaws that let hackers run code without a login.
What the Applied Compute acquisition means for your wallet
Here’s the thing: will your company’s security bill go up? Nobody has said that. Palo Alto made no pricing announcement, and no source reports one.
What Arora does expect is more demand, because AI traffic lifts the need for network security, identity and monitoring products. Our read: when a vendor buys a $4.5 billion platform, it eventually has to earn that money back. If you buy enterprise security, watch your renewal quotes.
One more thing on the “cashed out” talk. The deal is paid in Palo Alto shares, not cash. That means Applied Compute’s founders and investors are betting on Palo Alto stock too.
The timing wasn’t an accident
The news broke after the market closed, riding a strong earnings report. CNBC reports revenue hit $3.44 billion, up 39% from a year ago and ahead of the $3.31 billion analysts expected. Palo Alto then raised its full-year revenue guidance to $14.52 billion to $14.57 billion.
Shares jumped about 7% after hours and are up 85% this year. (Nothing says “good time to announce a big purchase” like a good report card.)
Palo Alto isn’t alone in paying up. BankInfoSecurity ranks this Palo Alto AI acquisition the third-largest cybersecurity acquisition of 2026. The company has also bought CyberArk and Chronosphere this year. Big deals can reshape a company’s fortunes, as Nike’s stunning slide out of the S&P 100 shows.
Next up: regulatory review, a close sometime in the spring, and Patil’s promise of “more to come soon.”
Palo Alto just bought an AI lab. Now it has to convince customers to buy it from a security company.






