Palo Alto Acquires Applied Compute for $4.5 Billion

Palo Alto Networks announced on Wednesday an agreement to acquire Applied Compute for $4.5 billion, payable entirely in stock. The deal, still subject to regulatory approvals, is expected to close in the third fiscal quarter of 2027. The company plans to offer its clients training and operation of open-weight AI models.
Planned Investments and Quarterly Results
Palo Alto Networks aims to build an inference capacity exceeding 100 megawatts, with capital expenditures expected to remain below $3 billion until fiscal year 2030. The company anticipates that the acquisition will generate over $1 billion in annual recurring revenue by fiscal year 2029.
The financial results for the first fiscal quarter were released at the time of the announcement. For the three months ending October 31, revenue increased by 39% to $3.44 billion, and net income rose from $334 million a year earlier to $482 million. Annual recurring revenue from next-generation security grew by 66% to reach $9.73 billion.
The company has revised its annual forecasts and now expects revenue between $14.52 billion and $14.57 billion, with adjusted earnings per share between $4.26 and $4.30. According to Nikesh Arora, Chairman and CEO of Palo Alto Networks, recent outages explain part of the demand. He partially attributes these incidents to the eagerness of IT teams to fix flaws identified by AI.
An Offer to Adapt Open Models
The weights of a model are the numerical values that determine its behavior: when they are open, users can download and modify them. Applied Compute is already assisting companies in training and executing their own versions of these models. Palo Alto Networks plans to offer these operations, as well as their adaptation, in a public cloud or in a client-specific sovereign environment. Applied Compute will retain its clientele, while the acquirer will market its platform to its own customers.
Le brief IA que les pros lisent chaque soir
Les 7 actus IA du jour, décryptées en 5 min. Gratuit.
Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.
Choisissez votre rythme
Gratuit · Pas de spam · Désabonnement en 1 clic
Nikesh Arora believes that companies will jointly employ both open and closed models, all requiring security. Palo Alto Networks claims that no major provider currently offers open models securely. According to its leader, their share of token traffic now exceeds 60%, up from less than 5% a year ago.
For Nikesh Arora, this acquisition goes beyond the realm of cybersecurity. If successful, he believes this business could become larger than its security segment. He had previously advocated for a reduction in AI prices as a condition for its growth and deemed the idea of slowing its development unrealistic.
Yash Patil to Co-Lead the Group's AI
Founded last year by three researchers who left OpenAI, Applied Compute raised $350 million earlier this month at a valuation of $3.25 billion, according to the Wall Street Journal. Its annual recurring revenue is said to have reached $100 million. The announcement of the acquisition came just minutes after journalist Kate Clark reported it in the Wall Street Journal.
Yash Patil, co-founder and CEO of Applied Compute, is set to take on a co-leadership role in AI at Palo Alto Networks once the deal is completed. Regarding employee retention, Nikesh Arora wrote on X that they have "millions of reasons" to stay for the next three to four years. The team is expected to retain its premises and a significant degree of autonomy.
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.