Palo Alto Networks announced on Wednesday that it will acquire Applied Compute, a platform for post-training and inference of open-weight AI models. The all-stock deal is valued at $4.5 billion. Wowtale reported earlier this month that the company had raised $350 million at a $3.25 billion valuation. It is being sold less than two weeks later.

Why a security company bought a model builder
Palo Alto Networks believes enterprise AI will be hybrid. A large share of companies will run open-source and open-weight models tuned to their own data, cost and control needs, it argues, and no large provider delivers those models securely today. Through Applied Compute, the company says, enterprises will be able to train, adapt and deploy open-weight models with security built in, in the public cloud or in sovereign environments.
Chairman and CEO Nikesh Arora said every enterprise will run a mix of open and closed models and that each one has to be secured. Palo Alto Networks saw the product and the research team firsthand as a prospective customer, he said, and its Unit 42 Frontier AI Assessment confirmed Applied Compute’s strong security foundations. Arora also argued that Palo Alto Networks cannot secure models it had no part in building.
Terms and leadership
The deal is subject to customary closing conditions, including regulatory clearance, and is expected to close in Palo Alto Networks’ fiscal third quarter of 2027, which runs from February to April. The value is subject to adjustments. The Applied Compute team will join Palo Alto Networks, and co-founder and CEO Yash Patil will become Co-Chief AI Officer after the close. Applied Compute will keep serving AI-native companies, digital natives and large enterprises, while Palo Alto Networks takes the platform to its global customers. The research team will also help embed AI across every Palo Alto Networks offering and train its own large and small models.
The price climbed quickly. The $4.5 billion is about 38% above the $3.25 billion valuation of this month’s round and about 3.5 times the $1.3 billion from an April round led by Kleiner Perkins. Both comparisons are Wowtale’s calculations.
What Applied Compute does
Applied Compute retrains open-weight models with reinforcement learning on a customer’s data and ways of working. It embeds engineers with customers and runs the finished models on its own cloud. All three founders came from OpenAI: Patil worked on Codex, Rhythm Garg on its first reasoning model and Linden Li on AI training infrastructure. The company said its annualized revenue run rate topped $100 million in under 18 months, a figure that is not audited. Microsoft, Shopify and Harvey are among its customers.
Palo Alto Networks’ position
The announcement came alongside fiscal first-quarter results on Wednesday. Revenue was $3.45 billion, up 39% from a year earlier, and the company raised its fiscal 2027 revenue forecast to $14.52 billion to $14.57 billion, from $14.10 billion to $14.20 billion. Shares rose about 9% in after-hours trading, and security peers CrowdStrike, Zscaler and Fortinet gained 2% to 5%.
It extends a pattern of buying its way into new categories. Palo Alto Networks has bought Chronosphere for $3.35 billion and CyberArk for $25 billion. This time it is reaching into training and serving models, which takes compute. The company is targeting more than 100 megawatts of compute within three to five years, mostly GPUs and servers.
Views are split. Some read the deal as extending the company from traditional security into enterprise AI. Others call it a risky move far from cybersecurity that raises the odds of competing with frontier AI labs. A supervisor trained by the party it oversees is not an independent check, critics note, and relying on one supplier for everything from firewalls to models raises vendor concentration risk.
Competition: one fewer independent platform
In the market for training and serving open-weight models for enterprises, Applied Compute will no longer be independent. Fireworks AI raised a $1.5 billion Series D in July at a $17.5 billion valuation, and Together AI raised $800 million in a Series C at an $8.3 billion valuation. Former OpenAI CTO Mira Murati’s Thinking Machines Lab offers a model customization platform called Tinker, and Reflection AI is pushing an open-weight model called Beam. For Applied Compute’s customers, the fact that their supplier will become part of a security vendor is now a variable. That is Wowtale’s reading.
MORE FROM THE POST
- CrowdStrike Signs Deal to Acquire AI Agent Observability Firm Fiddler AI
- Japanese Voice AI Startup Kotoba to Be Acquired by ElevenLabs
- Applied Compute Raises $350M at $3.25B as Run Rate Hits $100M in Under 18 Months
- 1Password Tops $500M ARR as It Extends Security to AI Agents
- Hugging Face Explores Sale at $13 Billion or More
Share
Most Read
- 1
- 2
- 3
- 4
- 5


Leave a Reply