· ai-devtools
Palo Alto Networks acquires Applied Compute to secure open-weight AI for enterprises
The desire among businesses to maintain data sovereignty and control is driving enterprises toward adopting open-weight and open-source models, a trend that is already evident as open-weight models…
The desire among businesses to maintain data sovereignty and control is driving enterprises toward adopting open-weight and open-source models, a trend that is already evident as open-weight models have increased from less than 5 percent of token traffic to over 60 percent in the last year. This shift aligns with Palo Alto Networks' strategic move to acquire Applied Compute, an AI model training platform, positioning the security vendor to better address the challenges posed by AI in the industry.
Nikesh Arora, the chairman and CEO of PAN, noted during the company's fiscal Q1 2027 earnings call on Wednesday that "We believe many of the major disruptions of the last few weeks, including outages across major enterprises and governments, were partially due to unintended consequences of patching and configuration changes, as IT teams simply cannot patch these vulnerabilities fast enough without compromising their production environments." Furthermore, he pointed out that "The second effect is the rise of AI-assisted cyberattacks." Consequently, in a move that can be seen as addressing a problem created by the technology itself, PAN is integrating Applied Compute to enhance the security of open-weight AI for its corporate clientele.
PAN has publicly stated that "the future of enterprise AI is hybrid, and that a large share of enterprises will run on open-source and open-weight models tuned to their own data, cost, and control needs." This commitment comes despite some recent turbulence in the open-weight sphere, notably Jensen Huang, co-founder of Nvidia, whose departure from the CEO role has caused some apprehension regarding the ecosystem's influence, particularly as the US remains cautious about models originating from China. Huang had previously advocated for open-weight models as a countermeasure to the dominance of closed-weight commercial offerings from entities such as Google, Microsoft, OpenAI, and Anthropic.
Despite these uncertainties, PAN maintains a clear vision for the role of open-weight models. Arora asserted, "Every enterprise is going to run a mix of open and closed models, and every one of those models needs to be secured." He added praise for the acquired firm, stating, "Applied Compute has built the best post-training and inference platform we've seen, and its research team is exceptional."
The acquisition, valued at $4.5 billion in all-stock, is projected to contribute over $1 billion to PAN's Next Generation Security business by the end of fiscal year 2029. Arora elaborated on the significant investment during the earnings call, framing it as an acknowledgment of the corporate drive to avoid being locked into a single vendor and to maintain command over proprietary information. He confirmed, "We're already seeing the desire for sovereignty and control compel enterprises to adopt open-weight and open-source models."
In addition to the financial aspect of the deal, PAN is also scaling up its underlying infrastructure capabilities. Arora mentioned that the business plans to erect infrastructure capable of supporting more than 100 megawatts of inference capacity, a project budgeted to require "a modest CapEx commitment of less than $3 billion through fiscal 2030."
Why it matters
This acquisition signals a major industry shift toward hybrid AI deployments, where enterprises prioritize control and avoiding vendor lock-in by using open-weight models. Developers should watch how PAN integrates Applied Compute's platform to secure these increasingly prevalent open-source models.





