Palo Alto Networks has agreed to acquire Applied Compute, a San Francisco startup that helps companies post-train and serve open-weight AI models, in an all-stock deal valued at $4.5 billion. A November 18 press release and SEC filing confirm the terms, announced with results showing revenue up 39% to $3.44 billion. It takes Palo Alto from securing AI into running model training and inference itself.
What Palo Alto Networks is paying and when it closes
The $4.5 billion price is subject to adjustments and will be paid entirely in Palo Alto Networks common stock. According to the company’s Form 8-K, the share count will be set using a reference price of $337.16, a 15-trading-day volume-weighted average up to November 17; the exact number of shares is not yet known.
Closing is expected in Palo Alto’s fiscal third quarter of 2027, which runs from February to April 2027, subject to customary conditions including regulatory clearance. After closing, Applied Compute co-founder and CEO Yash Patil is set to become Palo Alto’s Co-Chief AI Officer, and the startup’s research team will join the company. Applied Compute will keep serving its existing customers while Palo Alto also offers the platform to its own base.
What Applied Compute does
Applied Compute was founded in 2025 by three former OpenAI researchers, Yash Patil, Rhythm Garg and Linden Li. Its product, AC2 (the Applied Compute Agent Cloud), lets companies train, serve and continuously improve custom versions of open-weight models on their own data, with its engineers working alongside customer teams. The company says customers keep ownership of their models and data.
The startup has grown quickly. In April 2026 it raised $80 million led by Kleiner Perkins at a $1.3 billion post-money valuation, bringing its total funding to $160 million. On November 2, just 16 days before the deal, it announced $350 million at a $3.25 billion valuation, led by Elad Gil with NVIDIA, Blackstone and Shopify Ventures participating, and said it had passed $100 million in annualized revenue less than 18 months after founding. Named customers include Microsoft, DoorDash and Harvey. The $4.5 billion price is roughly 38% above that November valuation.
How the deal fits Palo Alto’s acquisition run
Applied Compute is the latest in a string of large Palo Alto purchases, all confirmed on the company’s press pages:
| Deal | Completed or expected | What it added |
|---|---|---|
| Protect AI | Completed July 2025 | AI model scanning and runtime protection for Prisma AIRS |
| Chronosphere | Completed January 2026 ($3.35 billion announced) | Observability and telemetry |
| CyberArk | Completed February 2026 ($45 cash plus 2.2005 PANW shares per CyberArk share) | Identity security, including AI agent identities |
| Applied Compute | Expected fiscal Q3 2027 ($4.5 billion in stock) | Open-weight model post-training and inference |
Unlike the others, Applied Compute is not a security product. Chief executive Nikesh Arora argued in the release that every enterprise will run a mix of open and closed models that all need securing, adding that Palo Alto first looked at Applied Compute as a prospective customer. On the earnings call, as reported by CNBC, management said it expects Applied Compute to add more than $1 billion of next-generation security annual recurring revenue by fiscal 2029, which would require about $3 billion of investment in more than 100 megawatts of compute over three to five years.
Who is affected
- Applied Compute customers will eventually buy from a security vendor rather than an independent lab.
- Palo Alto shareholders face dilution from the stock issuance and a new capital-heavy business line; CFO Dipak Golechha said the company intends to stay capital light, according to CNBC.
- Rival training and inference platforms such as Fireworks AI, which disclosed a breach exposing Notion AI chat transcripts this month, now compete with a vendor pitching security as the default.
- Enterprises adopting open-weight models such as Mistral Large 4 gain another hosting option in public or sovereign clouds.
What to watch next
- Regulatory clearance and the final share count.
- How Palo Alto prices and packages Applied Compute inside its platform.
- Announcements on the compute capacity build-out of more than 100 megawatts.
- Retention of Applied Compute’s researchers and customers.
Key takeaways
- Palo Alto Networks will buy Applied Compute for $4.5 billion in stock, expected to close in fiscal Q3 2027.
- Applied Compute, founded in 2025 by ex-OpenAI researchers, post-trains and serves open-weight models for enterprises.
- The price is about 38% above the $3.25 billion valuation set in a $350 million round on November 2.
- CEO Yash Patil will become Palo Alto’s Co-Chief AI Officer after closing.
- It follows the Protect AI, Chronosphere and CyberArk deals.
Frequently asked questions
How much is Palo Alto Networks paying for Applied Compute?
The deal is valued at $4.5 billion, subject to adjustments, and is paid entirely in Palo Alto Networks stock.
What does Applied Compute do?
It runs a platform for post-training and serving open-weight AI models on a company’s own data, founded in 2025 by three former OpenAI researchers.
When will the Applied Compute acquisition close?
Palo Alto Networks expects the deal to close in its fiscal third quarter of 2027, subject to customary conditions including regulatory clearance.




