On November 18, 2026, Palo Alto Networks announced a definitive agreement to acquire open-weight model platform Applied Compute in an all-stock deal valued at $4.5 billion. It is the largest AI acquisition the cybersecurity industry has seen — and the first time a security vendor has stated plainly that the future of enterprise AI is hybrid: open-weight models and closed frontier models will coexist for a long time, and security companies must move from "adding guardrails" to "building the foundation."
Deal terms: $4.5 billion in stock, founder becomes co-chief AI officer
Under the agreement announced that day, the all-stock transaction is valued at $4.5 billion, subject to adjustments, and is expected to close in Palo Alto Networks' fiscal third quarter of 2027, pending regulatory clearance. Applied Compute co-founder and CEO Yash Patil will become Palo Alto Networks' co-chief AI officer after the close, with the whole research team joining. Patil and colleagues founded Applied Compute in 2025 after leaving OpenAI; the company runs a platform for post-training and inference of open-weight models, helping enterprises fine-tune models on their own data and put them into production. On November 2 it disclosed a $100 million annualized revenue run rate less than 18 months after founding, after recently closing $350 million in funding at a $3.25 billion valuation.
Why a security vendor wants a model platform
Palo Alto Networks put it plainly in its release: many enterprises will fine-tune open-source and open-weight models to their own data, cost and control needs while continuing to use closed frontier models. Buying Applied Compute means acquiring both a ready-made training and serving platform and a model research team — the missing piece for a security vendor: letting enterprises train, adapt and deploy their own models, with security controls built in, in the public cloud or sovereign environments. The fiscal first-quarter 2027 results released the same day underscored the deal: quarterly revenue of $3.44 billion, up 39% year over year. The deal is not done yet, though — regulatory review comes first. An 8-K filing shows the share count will be based on $337.16 per share, the volume-weighted average price over the 15 consecutive trading days ending November 17.
For enterprise customers, the change will not land today but when the platform goes live: deploying open-weight models may shift from "build it yourself" to "buy it as a service," and security moves from post-deployment checks into training and inference. Two things are worth watching: when regulators clear the deal, and whether Palo Alto Networks opens the platform to smaller businesses or serves large customers first. AI is now finding vulnerabilities faster than teams can patch them — two remote-code-execution flaws in NGINX that hid for eleven years are the latest example — and the security industry's battleground is shifting from blocking bugs to governing models.