French telecom group Orange is partnering with Mistral AI to integrate a cybersecurity-focused version of the startup’s latest artificial intelligence model into its digital security business, as European companies seek to strengthen their defenses against sophisticated AI-assisted attacks.
The agreement gives Orange Cyberdefense, one of Europe’s largest cybersecurity providers, access to a European-developed AI model for selected security products and services. The companies are positioning the partnership around two priorities: identifying weaknesses in clients’ systems before hackers exploit them and accelerating investigations after security breaches occur.
Orange said the models would run on computing infrastructure located in Europe, allowing the company to keep its dedicated cybersecurity operations within the region. The initial rollout will be gradual and restricted to selected services for corporate and public-sector customers.
Register for the next Tekedia Mini-MBA.
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Financial terms of the agreement were not disclosed.
The partnership comes as cybersecurity providers face a changing threat environment in which attackers can use AI tools to automate parts of their operations, develop malicious content more quickly and increase the scale of attempted intrusions. Security companies are also turning to AI to process large volumes of alerts, examine suspicious activity and help analysts identify threats that might otherwise take longer to investigate.
For Orange Cyberdefense, the arrangement provides a way to incorporate a European AI model into its security operations while retaining control over where the associated computing takes place. Orange reported €1.3 billion ($1.45 billion) in revenue for the cybersecurity business in 2025.
The use of European computing infrastructure is a notable feature of the agreement as businesses and public institutions pay greater attention to data residency, operational control and dependence on technology providers outside the region.
Cybersecurity operations can involve sensitive information about network configurations, software vulnerabilities, employee accounts and suspected intrusions. Using AI to analyze this material raises questions about where data is processed, which systems can access it, and how information is handled by model providers.
Running dedicated models on infrastructure located in Europe may help Orange address some of these concerns for customers with regional data-handling requirements. However, the location of computing alone does not establish the full extent of data protection. Contractual safeguards, access controls, retention policies and the way models are deployed will also determine how sensitive information is managed.
The choice of Mistral AI also gives Orange access to a European-developed alternative in a market where many organizations rely on AI systems from major US technology companies. For European businesses and public-sector institutions, the ability to choose among providers can be relevant to procurement decisions and efforts to limit dependence on a narrow group of technology suppliers.
That does not necessarily mean a European model will outperform competing systems on every cybersecurity task.
The partnership’s initial focus on selected services suggests Orange intends to test the technology before expanding its use. That approach is important in cybersecurity, where an AI system that generates incorrect findings or misses a genuine threat can impose costs on customers and security teams.
AI’s Dual Role in Cybersecurity
The deal comes as the AI industry is grappling with a major security challenge: the same advances that help defenders analyze threats can give attackers additional capabilities.
Security teams must process growing volumes of system logs, alerts, and incident reports, often under pressure to determine whether suspicious activity represents a genuine breach. AI models could help summarize evidence, identify patterns across large datasets, and accelerate the investigation of incidents. They may also help security specialists test systems for weaknesses before those vulnerabilities are exploited.
Orange’s stated priorities cover both sides of this defensive process. Testing customer systems can help identify vulnerabilities before an attack, while faster breach investigations can reduce the time analysts spend sorting through evidence after an incident.
Yet deploying AI does not remove the need for human oversight. Models can produce false positives, overlook important details, or recommend actions that are inappropriate for a particular network. Their effectiveness, thus, is determined by the quality of the underlying data, the security tools they can access and the procedures used to validate their findings.
The companies have not disclosed the financial terms or detailed performance targets for the partnership. Its initial limited deployment also means the scale of any commercial benefits remains uncertain.