Check Point Software Technologies has taken a further step in its “security-first” strategy for the AI era. The company has unveiled Quantum Firewall Software R82.10, a new version of its perimeter protection platform that incorporates 20 additional capabilities designed to help businesses adopt AI securely, protect distributed environments, and simplify the Zero Trust model in hybrid networks.
The solution is part of Check Point’s Infinity Platform and relies on the “open-garden architecture” approach, which allows network security to integrate with more than 250 third-party solutions, from endpoint tools to identity and orchestration platforms.
AI at the center of the network… and also in the threat model
Organizations are increasing their use of generative AI tools, LLMs, and intelligent agents, both for internal use cases and customer-facing products. This surge in adoption brings a completely new attack surface: identity abuse, sensitive data flows ending up in external models, unregulated API use, or inconsistent configurations across cloud, office, and remote users.
“As companies embrace AI, security teams are pressured to protect more data, more applications, and much more distributed environments,” explains Nataly Kremer, Chief Product Officer at Check Point. “R82.10 helps shift to a prevention-first model, unifying management, reinforcing Zero Trust, and adding specific protections for safe and responsible AI adoption.”
Visibility and control over AI tool usage
A key element of the new version is the segment “Supporting Safe AI Adoption”. Here, Check Point introduces functions to:
- Detect unauthorized use of GenAI tools, both within corporate networks and remote connections.
- Identify and classify popular AI applications like ChatGPT, Claude, or Gemini, enabling fine-grained policies: allow, block, or limit based on user type, department, or context.
- Monitor the Model Context Protocol (MCP), increasingly present in workflows where AI agents connect to internal tools, databases, or external services.
With these capabilities, cybersecurity teams can differentiate between legitimate use—for example, a developer consulting a corporate AI assistant—and risky behavior, such as exfiltrating sensitive info to unauthorized services.
Consistent security in hybrid networks and SASE
The second set of improvements targets the protection of hybrid networks under the banner of “Strengthening Hybrid Mesh Network Security”. R82.10 enables:
- Centralized management of Internet access for both traditional gateways and the SASE layer, preventing policy inconsistencies between physical sites, remote users, and cloud workloads.
- Simplified connectivity between firewalls and SASE, reducing deployment complexity in mesh environments and improving network team experience.
- Enhanced identity and device posture validation, reinforcing Zero Trust principles at scale.
Practically, this allows organizations to apply the same security policies to users connecting from an office, from home, or from a public cloud—regardless of entry point to the network.
More prevention, less noise: phishing without breaking HTTPS and adaptive IPS
The third area, “Taking a Prevention-First Approach to Modern Threats”, adds capabilities aimed at reducing reaction times to advanced attacks without increasing alert volumes.
Highlights include:
- Phishing protection without full HTTPS inspection, reducing privacy impact and performance hit while maintaining malicious page blocking.
- Adaptive IPS, adjusting signatures and priorities to minimize false positives and alert fatigue in SOC teams.
- Threat Prevention Insights, a module focusing on configuration errors and posture gaps before attackers can exploit them.
The goal is clear: shift from a reactive incident response approach to a proactive platform that detects and addresses vulnerabilities.
Unified platform with over 250 integrations
The fourth set of features in R82.10 centers on “Eliminating Silos with a Unified Security Platform”. Check Point expands its open-garden architecture with more than 250 integrations, allowing identity and posture data from other providers to be used directly in firewall policies.
For example, an organization can:
- Incorporate endpoint posture signals (antivirus, EDR, disk encryption, patching) into perimeter access decisions.
- Correlate identity data with risk information to enable more precise Zero Trust policies.
- Automate incident responses by integrating the firewall with SIEM, SOAR, or identity management solutions.
AI and AI security as core strategies in Check Point’s approach
R82.10 integrates with the full stack of Check Point AI security, strengthened by the recent acquisition of Lakera, a company specializing in protecting AI workloads from training to inference. The provider aims to deliver a end-to-end AI security stack that covers:
- Protection of the model itself (data poisoning, prompt injection, model extraction).
- API and agent security interacting with models.
- Data control and compliance in hybrid and multi-cloud environments.
Chris Konrad, Vice President of Global Cyber at World Wide Technology, summarizes the approach: “Check Point’s AI-powered security capabilities help defend against the latest threats while providing enterprise-level protection for AI workloads, from training to inference—without compromising performance.”
Availability and next steps
Check Point will showcase the new features of Quantum Firewall Software R82.10 in the webinar “Securing the AI Transformation in a Hyperconnected World”, featuring also representatives from NVIDIA and IDC. The software will be available for download later this month and can be integrated with the Infinity Platform and other company solutions.
With R82.10, Check Point reaffirms its message to a market racing to adopt AI without overlooking the risks: AI security cannot be an afterthought but must be embedded into the network itself.
Frequently Asked Questions about Check Point Quantum R82.10
1. What does Check Point R82.10 add compared to previous versions of the Quantum firewall?
R82.10 introduces 20 new capabilities focused on safe AI adoption, hybrid network protection, modern threat prevention, and a unified management platform with over 250 third-party integrations.
2. How does R82.10 help control the use of generative AI tools within the organization?
The software detects apps like ChatGPT, Claude, or Gemini, identifies unauthorized use, and monitors protocols like MCP, enabling policies based on user, department, or data type to reduce data exfiltration risks.
3. Is full HTTPS inspection necessary for phishing protection?
Not necessarily. One of R82.10’s new features is phishing protection without complete HTTPS inspection, improving performance and compliance where decrypting encrypted traffic isn’t desired.
4. Is R82.10 only suitable for large enterprises or also for mid-sized hybrid environments?
While Check Point primarily targets large organizations and service providers, the hybrid mesh and SASE approach enables medium-sized companies with distributed infrastructure to adopt a consistent Zero Trust model, integrating firewalls, cloud, and remote users within a single security platform.
via: checkpoint