Palo Alto Buys Applied Compute in $4.5B AI Security Bet

Published (Updated )

Palo Alto Networks is making its biggest bet yet on the future of enterprise AI. The cybersecurity giant announced on November 18, 2026 that it will acquire Applied Compute, a fast-growing AI infrastructure startup, in an all-stock deal valued at $4.5 billion. The move puts Palo Alto at the centre of one of the hottest questions in business technology: how do companies actually run powerful AI models on their own data, safely?
The quick answer:
- The deal: Palo Alto Networks will acquire Applied Compute for $4.5 billion, paid entirely in stock
- Announced: November 18, 2026, alongside Palo Alto's fiscal first-quarter 2027 results
- What Applied Compute does: a platform for post-training and inference of open-weight AI models
- Expected close: Palo Alto's fiscal third quarter of 2027, subject to regulatory clearance
- Key person: Applied Compute co-founder and CEO Yash Patil will become Palo Alto's Co-Chief AI Officer
- The context: Palo Alto also reported record quarterly revenue of $3.44 billion, up 39% year over year, and raised its full-year forecast
What Applied Compute actually is
Applied Compute is not a household name, but in AI circles it has become one of the most talked-about startups of 2026. Founded in 2025 by Yash Patil, Rhythm Garg and Linden Li, the San Francisco company builds a platform that takes open-weight AI models, trains them on a customer's own data, and runs them in production. That description needs unpacking, because it sits at the heart of why Palo Alto is paying $4.5 billion. Open-weight models are AI models whose weights β the trained parameters that make them work β are publicly available. Companies can download them, adapt them and run them on their own systems, instead of sending their data to someone else's cloud. The best-known examples come from labs like Meta, Alibaba and DeepSeek. For a bank, a hospital or a government agency, that control matters: sensitive data never has to leave their own infrastructure. But running these models well is hard. They need to be fine-tuned on company data (that is post-training), served efficiently at scale (that is inference), and kept secure throughout. Applied Compute's platform does all three, and it does it fast enough that the company reported a $100 million annualised revenue run rate less than 18 months after founding, according to its November 2, 2026 announcement. Just sixteen days before the acquisition, Applied Compute raised $350 million from investors including Nvidia, Blackstone, Kleiner Perkins and Lux Capital, at a $3.25 billion valuation. Palo Alto's $4.5 billion offer values the startup about 38% higher than that round β a premium that says a lot about how badly Palo Alto wanted it.
Why Palo Alto wants an AI company
Palo Alto Networks is the world's largest pure-play cybersecurity company, and its chairman and CEO Nikesh Arora has been steering it towards AI for years. The logic of this deal is straightforward: every enterprise customer is now trying to adopt AI, and every one of them is worried about doing it safely. Arora said Palo Alto had evaluated Applied Compute firsthand as a prospective customer before deciding to buy it. That is a telling detail. Palo Alto was not shopping for a trophy asset β it was using the product, liked what it saw, and decided to own it. The strategic pitch has three parts:
- Secure AI adoption: Palo Alto wants to help enterprises adopt AI securely, and owning the platform that trains and runs their models gives it control over the whole stack
- A hybrid AI future: the combined company is pitching a "secure hybrid AI" model, where companies run some AI in the cloud and some on their own systems, with Palo Alto securing both
- An in-house research team: Applied Compute brings a world-class AI research team, not just software. Palo Alto plans to use that team to build models of its own
The last point matters. Palo Alto is not just buying a product β it is buying the people who can build the next generation of AI security products. Yash Patil becoming Co-Chief AI Officer signals that the Applied Compute team will have real influence inside Palo Alto, not just a quiet integration.
The numbers behind the deal
Palo Alto announced the acquisition alongside strong quarterly results, and the market liked both pieces of news. Shares rose sharply after the announcement.
- Deal value: $4.5 billion, all stock, subject to adjustments β no cash changes hands
- Applied Compute's last valuation: $3.25 billion, set on November 2, 2026
- Applied Compute's funding: $350 million raised in that round, from Nvidia, Blackstone, Kleiner Perkins and Lux Capital
- Applied Compute's revenue run rate: $100 million annualised, reached in under 18 months
- Palo Alto's quarterly revenue: $3.44 billion in fiscal Q1 2027, up 39% year over year
- Next-generation security ARR: $9.73 billion, up 66% year over year
- Palo Alto's full-year forecast: raised to $14.52β$14.57 billion, up from $14.10β$14.20 billion

The all-stock structure is worth noting. Applied Compute's backers are not getting a cheque β they are getting shares in Palo Alto. That means they are betting that the combined company will be worth more than the sum of its parts. It also means the final value of the deal will move with Palo Alto's share price until it closes, which is not expected until fiscal Q3 2027. For Palo Alto's shareholders, the question is whether $4.5 billion is a fair price for a company founded less than two years ago. The $100 million revenue run rate implies a multiple of roughly 45 times annualised revenue β expensive by any normal standard. But in the current AI market, where the best infrastructure teams are scarce and every large company is racing to own them, Palo Alto is paying for speed and scarcity as much as for revenue.
What open-weight AI means for businesses
To understand why this deal matters beyond the two companies, it helps to understand the shift towards open-weight AI. For the first few years of the AI boom, most companies used AI through someone else's API. You sent your data to OpenAI, Google or Anthropic, and their model sent back an answer. That was simple, but it created three problems: your data left your building, you had no control over the model, and you were locked into one vendor's pricing. Open-weight models flip that. A company downloads a model, fine-tunes it on its own documents, contracts or patient records, and runs it on its own servers or in a private cloud. The data stays home. The model can be customised. And the company is not dependent on a single provider. The catch is that running these models well is a specialist job. You need to:
- Post-train the model: adapt a general model to your specific data and tasks, without breaking what it already knows
- Serve it efficiently: run it fast enough and cheaply enough for real users, which means serious engineering around inference
- Keep it secure: make sure the model cannot be tricked into leaking data, and that the whole pipeline meets compliance rules
- Keep it current: update the model as new data arrives, without starting from scratch each time

Applied Compute's platform is built to do exactly that. Palo Alto's bet is that most large enterprises will want this capability, and that they will want it from a vendor they already trust with their security.
The risks and the open questions
A $4.5 billion acquisition of a company founded in 2025 is not without risk, and several questions will decide whether it pays off.
- The deal is not done yet: it still needs regulatory clearance and other customary conditions, and it is not expected to close until fiscal Q3 2027. That is a long window, and the value of an all-stock deal can shift with the market.
- Integration is the hard part: buying a fast-moving startup and folding it into a company of Palo Alto's size is genuinely difficult. The best people can leave, the product roadmap can stall, and the culture clash can slow everything down. Putting Yash Patil in a co-chief AI officer role is a good start, but it is not a guarantee.
- The AI market moves fast: open-weight models are improving at a furious pace. What looks like a leading platform today can look ordinary in a year. Palo Alto is betting that Applied Compute's team will keep it ahead.
- Competition is fierce: every major cloud provider, chipmaker and cybersecurity vendor is chasing the same enterprise AI opportunity. Palo Alto is not the only company that wants to own this space.
- The price is high: at roughly 45 times annualised revenue, Palo Alto needs Applied Compute to keep growing fast for the deal to look good in hindsight.
None of these risks means the deal is a mistake. But they do mean the next year will be watched closely β by Palo Alto's customers, its shareholders, and its competitors.
What this means for the AI security market
This acquisition is part of a bigger pattern. Cybersecurity companies are realising that AI is not just a new thing to protect β it is a new thing to sell. The companies that secure AI will be the ones that help enterprises actually run it. Palo Alto's move follows a string of AI deals across the industry in 2026, and it sends a clear message: the security vendors that win the next decade will be the ones that own the AI infrastructure, not just the firewalls around it. For enterprise buyers, the practical takeaway is simpler. If your company is planning to run open-weight models on your own data β and many are β the tools for doing that securely are consolidating into the hands of a few large vendors. Palo Alto now wants to be one of them. For the wider market, the deal is another sign that the AI infrastructure boom is not slowing down. A startup founded in 2025, with $100 million in annualised revenue, just sold for $4.5 billion. That is the kind of number that tells you where the money β and the momentum β is going.
The bottom line
Palo Alto Networks is paying $4.5 billion in stock for a company that did not exist two years ago, because it believes the future of enterprise AI is open-weight models run securely inside the enterprise β and it wants to own the platform that makes that possible. The deal pairs strong quarterly results with a bold strategic bet, and the market's first reaction was positive. Whether it works will depend on integration, execution and a regulatory process that runs into 2027. But the direction of travel is clear: cybersecurity and AI infrastructure are becoming the same business, and Palo Alto intends to be at the centre of it.
FAQs
How much is Palo Alto Networks paying for Applied Compute?
Palo Alto Networks will acquire Applied Compute in an all-stock transaction valued at $4.5 billion, subject to adjustments. No cash changes hands β Applied Compute's investors will receive Palo Alto shares.
When will the Palo AltoβApplied Compute deal close?
Palo Alto expects the deal to close in its fiscal third quarter of 2027, subject to customary closing conditions including regulatory clearance.
What does Applied Compute do?
Applied Compute builds a platform for post-training and inference of open-weight AI models β it helps enterprises adapt public AI models to their own data and run them in production, on their own infrastructure.
Who is Yash Patil?
Yash Patil is the co-founder and CEO of Applied Compute. After the acquisition closes, he will become Palo Alto Networks' Co-Chief AI Officer, and the Applied Compute team will join Palo Alto.
How much revenue does Applied Compute make?
Applied Compute reported a $100 million annualised revenue run rate on November 2, 2026, less than 18 months after the company was founded. It had just raised $350 million at a $3.25 billion valuation.
Why is Palo Alto buying an AI company?
Palo Alto wants to help enterprises adopt AI securely. Owning Applied Compute's platform β and its research team β gives Palo Alto control over how companies train and run open-weight models, which it sees as the future of enterprise AI.
How did Palo Alto perform in its latest quarter?
Palo Alto reported fiscal first-quarter 2027 revenue of $3.44 billion, up 39% year over year, with next-generation security annual recurring revenue of $9.73 billion, up 66%. It raised its full-year revenue forecast to $14.52β$14.57 billion.
Was this article helpful?

Written By
TechorCoffeeContributing Writer
TechorCoffee is a technology reporter and software creator writing reviews and tutorials for TechOrCoffee.
Comments (0)
Leave a Comment
Read Next

Tue Apr 21 2026
Best AI Tools for Students & Creators in 2026:Complete Guide
AI tools are transforming how students learn and creators produce content. From writing assistants to design and research tools, these platforms save time, enhance creativity, and improve efficiency. Choosing the right AI stack can significantly boost productivity and output quality.

Tue Jun 16 2026
Apple Intelligence 2.0: Siri AI & iOS 27 at WWDC 2026
Apple Intelligence 2.0 is here. I cover Siri AI, iOS 27 release dates, supported devices, and how Apple's new AI really compares to ChatGPT and Gemini.

Tue Oct 13 2026
Gemini 4 Argon: Everything You Need to Know
Gemini 4 Argon is Googleβs new flagship AI model. See its benchmarks vs GPT-6 Astra and Claude, pricing, 1M-token output, safety features and how to get access.