White Paper | AI Data Center & AI Factory Security Blueprint

How to secure corporate AI infrastructure and Private LLMs leveraging Check Point AI security technologies stack, securing access to the AI Data Center and Public Cloud, protection of AI workloads and applications, data security and management.

/live_board/v2/downloads?uid=86bb66ab-c466-40cf-aeb3-e1dd63c19343&redirect=true

AI Data Center & AI Factory Security Blueprint

How to secure corporate AI infrastructure and Private LLMs with Check Point's AI security technologies stack. Secure access to

the AI Data Center, AI Factory, and Neocloud, protecting AI workloads, applications, data, and management.

2AI DATA CENTER & AI FACTORY SECURITY BLUEPRINT

External Users / Internet / Web Apps

Enterprise Network / IdP / SSO / SIEM / SOAR / SOC

Check Point Maestro Hyperscale Firewall

- Virtual Systems (FW) - User Access / ZTNA - AI infra protection - External DC VPN

Check Point WAF / AI Agent Security

- Prompt Injection - DLP - LLM Guardrails - API Abuse Protection - Red Teaming Prompts

External Access Zone / AI-DC Perimeter

Ingress LB / API Gateway MGMT VLAN / DevOps / SecOps / GitHub / Update Servers Internet

Management Plane Data Plane

Inference Cluster Training Models Cluster

Storage / Data Lake

GPU Server / NVIDIA DPU BlueField + Check Point Firewall

- Inline FW/IPS line-rate (North-South) - Segmentation - Runtime K8s protection (AppShield) - Telemetry - Control Plane secure access / DevOps / Threat Prevention

Kubernetes Security - Intra-cluster traffic mapping - Lateral movement protection - Micro-Segmentation for EW traffic (labels, namespace)

1. Introduction The adoption of private AI and LLM (Large Language Model) infrastructures by enterprises introduces a new class of risks. Unlike traditional IT workloads, AI data centers manage sensitive training data, powerful GPU clusters, distributed inference services, and high-throughput pipelines that can easily become attack vectors.

Another segment building their own AI factories are Neocloud providers, who deliver GPU-as-a-Service, building hyperscale AI factories powered by NVIDIA and other leading GPU platforms to give enterprises on-demand, high-performance compute for training and inference.

Organizations face threats to data, intellectual property, AI models, and end-users. Building AI capabilities without embedding security increases exposure to poisoning, data leakage, and governance failures.

To ensure resilience, AI data centers must be secured end-to-end — from the fabric and GPU clusters to Kubernetes workloads, and API-driven inference workloads and services.

2. Value Check Point provides for AI Factory Security Check Point enables organizations to confidently adopt and scale AI by addressing both traditional IT threats and the new, unique risks of AI-driven environments. Check Point solutions based on the modern security technologies and integration with advanced 3rd party products, part of general open-garden strategy; - provides embedded cyber security by design to cover all sensitive blocks of AI-Data Center and Private LLMs.

www.checkpoint.com

3AI DATA CENTER & AI FACTORY SECURITY BLUEPRINT

The solution provides layered protection approach from the access towards AI workloads as shown on the figure above:

• Perimeter Layer Protection – external access control zone, entry point to the AI Data Center fabric, secured by Zero-Trust (ZTNA) enabled Hyper Scale NGFW (Maestro)

• Application Layer Protection – API based security of AI application and Agentic / LLM layer protection which is different from “traditional” api security

• AI-Server Layer Protection – HW embedded NGFW to take a security close to the AI workloads, segment servers (DGX) and protect Management traffic.

• Kubernetes Layer Protection – ensure East-West traffic inside K8s cluster visibility and containers micro-segmentation policy enforcement

The value delivered goes far beyond technology, translating directly into measurable business outcomes:

Business Value Description

Protection of Intellectual Property and Data Assets Safeguards proprietary AI models, training datasets, and inference results — preventing theft or manipulation of high-value R&D assets.

Business Continuity and Service Reliability Ensures AI services remain resilient and available even under attack, minimizing downtime and avoiding costly disruptions.

Regulatory Compliance and Trust Provides governance, traceability, and auditability to meet emerging AI regulations and maintain customer and regulator trust.

Risk Reduction and Cost Avoidance Reduces the likelihood of breaches, data leakage, and compliance fines, protecting both finances and brand reputation.

Operational Efficiency Simplifies security operations across training and inference with centralized policy management and automation, lowering overhead for DevOps and SecOps.

Secure Innovation and Faster AI Adoption Embeds protection from the ground up, enabling organizations to adopt and scale AI with confidence while safeguarding users and customers.

Customer and Partner Confidence Demonstrates strong security commitment, becoming a differentiator in markets where trust and reliability drive adoption and partnerships.

View Executive Guide Speak with an Expert

www.checkpoint.com https://engage.checkpoint.com/executive-guide-securing-the-ai-data-center-ai-factory https://engage.checkpoint.com/executive-guide-securing-the-ai-data-center-ai-factory https://pages.checkpoint.com/2026-mar-ww-ai-factory-and-data-center-security.html https://pages.checkpoint.com/2026-mar-ww-ai-factory-and-data-center-security.html https://pages.checkpoint.com/2026-mar-ww-ai-factory-and-data-center-security.html

4AI DATA CENTER & AI FACTORY SECURITY BLUEPRINT

3. High-Level Overview of AI Data Center Architecture An AI data center is based on model training and inference domains at scale, combining high-performance GPU clusters (e.g NVIDIA), secure connectivity, and orchestration layers.

At the edge, a frontend application layer with API gateways, load balancers, firewalls, and WAFs manages and protects user and application traffic, while a dedicated management layer hosts DevOps, SecOps, and control functions over isolated VLANs.

Inference Cluster hosts deployed AI models that process user queries and application requests in real time. It consists typically of GPU-powered DGX servers orchestrated by Kubernetes with Cilium or other K8s CNI technologies. This cluster handles model inference requests, ensuring high- performance execution and efficient resource utilization across distributed nodes.

Training clusters, typically based on DGX servers, use InfiniBand interconnects for ultra-fast GPU communication and frameworks like Slurm or Ray to coordinate distributed workloads, whereas Inference clusters rely on Kubernetes with Cilium for real-time model serving and policy enforcement.

* Slurm = Simple Linux Utility for Resource Management

Training Models Cluster (GPU) Workloads Slurm / Ray

Inference Clusters (GPU) Workloads Kubernetes / Cilium

Storage

Frontend Application Layer API GW , Load Balance , Perimeter FW , WAF

InfiniBand DGX Intra traffic

NVMe over RDMA/Ethernet

North/South Data Plane Traffic API: Ethernet

Management Layer GitLab, GitHub, DevOps, Nvidia MGMT , Slurm Controller

InfiniBand DGX Intra traffic

MGMT Control Plane : Ethernet

Internet

*Optional – customers can use pre-trained models

NVIDIA DGX / Open Server

RAG

LLM/SLM

* Slurm = Simple Linux Utility for Resource Management

www.checkpoint.com

5AI DATA CENTER & AI FACTORY SECURITY BLUEPRINT

4. AI Infrastructure Security Risks AI infrastructure introduces unique risks that extend beyond traditional IT systems. These risks directly affect the confidentiality, integrity, and availability of sensitive data, models, and applications. Unlike standard data centers, AI environments combine high-performance computing, large-scale data pipelines, and distributed training clusters — all of which create new attack surfaces, regulatory challenges, and risks of misuse.

Category Key Risks

Infrastructure & Platform Risks

• Compromise of servers, workloads, or system integrity • Lateral movement through East-West traffic within AI clusters • Compromise of containers or workloads via malicious libraries from GitHub • Exploitation of misconfigured FWs, API GW, or exposed MGMT interfaces / DevOps misuse • Container escape or privilege escalation attacks targeting Kubernetes runtimes • API gateway bypass or misuse exposing model endpoints directly • Resource exhaustion or GPU/memory DoS attacks impacting AI availability

AI Supply Chain Risks • Tampering in third-party frameworks, APIs, or pre-trained models • Dependency poisoning through open-source libraries • Unverified model weights or firmware updates introducing hidden risks

Data & Training Risks

• Training data poisoning creating hidden backdoors • PII leakage from prompts, responses, or logs • Cross-tenant data bleeding in multi-tenant GPU or inference environments • Unauthorized access or exfiltration of proprietary datasets

Model Risks

• Model poisoning attacks compromising training integrity • Model theft or extraction via API abuse or query enumeration • Adversarial attacks causing targeted misclassification • Model drift and performance degradation over time

Application Risks

• Prompt injection and jailbreak attempts bypassing model controls • Output manipulation or generation of harmful/unethical content • RAG poisoning affecting context accuracy • Agent hijacking or manipulation of autonomous behavior • Vulnerabilities in third-party integrations or plug-ins

AI Governance & Operational Risks

• Lack of AI system accountability or auditability • Insufficient access control and monitoring across AI pipelines • Shadow AI deployments bypassing enterprise governance • Misalignment between security, compliance, and DevOps ownership

Compliance & Regulatory Risks

• Violations of AI-specific regulations (EU AI Act, U.S. Executive Order 14110) • Failure to meet model explainability and “right to explanation” (GDPR) • Breach of data residency and cross-border AI processing laws • Non-compliance with industry frameworks (HIPAA, PCI-DSS, ISO 42001)

www.checkpoint.com

6AI DATA CENTER & AI FACTORY SECURITY BLUEPRINT

5. AI Security by Design “AI must be Secure by Design. This means that manufacturers of AI systems must consider the security of the customers as a core business requirement, not just a technical feature, and prioritize security throughout the whole lifecycle of the product, from inception of the idea to planning for the system’s end- of-life. It also means that AI systems must be secure to use out of the box, with little to no configuration changes or additional cost.” - CISA, Software Must Be Secure by Design, and Artificial Intelligence Is No Exception

AI must be Secure by Design, not an afterthought bolted on top of existing systems.

• Zero Trust everywhere: every user, API, and service interaction must be authenticated, authorized, and continuously validated.

- Microsegmented service mesh with mTLS between training and inference zones.

- Strong NHI (Non-Human Identity) security: service accounts, API key management, and secret rotation.

- Context-aware access control based on model risk and data sensitivity.

• AI-Native Security Controls:

- LLM gateways with prompt injection detection and rate limiting.

- Input sanitization and output filtering using policy-based guardrails.

- Model access control via RBAC/ABAC and runtime protection.

- Agent authorization boundaries to isolate agentic AI actions.

• Data & model integrity: enforce signed models, encrypted data pipelines, and isolated training zones.

• Continuous validation: runtime monitoring, adversarial red-teaming, and anomaly detection to expose blind spots.

• Governance & accountability: embed policy controls that regulate what AI can access, generate, or share.

- Maintain audit trails across training, deployment, and inference workflows.

- Ensure alignment with EU AI Act, GDPR, and sector mandates (HIPAA, PCI-DSS, ISO 42001).

Value: Secure by Design transforms AI from “functional but fragile” into resilient, governed, and business-ready.

www.checkpoint.com https://www.cisa.gov/news-events/news/software-must-be-secure-design-and-artificial-intelligence-no-exception

7AI DATA CENTER & AI FACTORY SECURITY BLUEPRINT

6. AI Data Center Generic Security

This design illustrates a typical secure AI data center architecture that separates training and inference clusters while enforcing strict controls accross dedicated network zones (Training, Inference, Storage and Management VLANs and segments).

The Training clusters in an isolated environment with no direct Internet access, leverages high-speed interconnects for model development. The Inference clusters handle real-time workloads through Kubernetes orchestration, with users connecting to AI workloads by generating prompts via API gateway located at the front end of the AI fabric.

Security is layered across the stack: the management plane is isolated and protected, ensuring DevOps and administrative access is tightly controlled; north-south traffic from the internet is secured with perimeter firewalls, WAF API / AI-aware gateways that enforce prompt injection detection, rate limiting, and guardrails. The access control is enforced for east-west traffic between clusters and storage, for segmentation and monitoring.

Key principles such as AI-aware security, Zero Trust access, and continuous inspection of both API and DevOps traffic ensure that sensitive models, data, and workloads remain protected against tampering, leakage, and misuse.

Training Models Cluster (DGX clusters / K8s Slurm, Ray)

Inference Clusters (DGX clusters / K8s Cilium)

Ethernet Ethernet

DGX Server

DPU BlueField Card (Ethernet & Infiniband)

GPU GPU GPU GPU GPU GPU

Load Balancer

API GW

Perimeter FW

Internet

WAF API Security

NS API data traffic:Ethernet

NGFWDevOps / Ops

Internet (GitLab ; GitHub)

K8s K8s K8s K8s K8s K8s

K8s : Kubernetes Cilium Nodes with Apps and

LLMs

Storage MGMT VLAN MGMT VLAN

MGMT VLAN

K8s / DGX operations traffic from Devs and Admins

North/South Security: K8s / DGX updates and LIBs download from the Internet

North/South user traffic from AI apps / web

GPU Server

* Slurm = Simple Linux Utility for Resource Management

www.checkpoint.com

8AI DATA CENTER & AI FACTORY SECURITY BLUEPRINT

7. Check Point AI Data Center Security Architecture

The Check Point AI Data Center Security architecture delivers end-to-end protection for both training and inference domains by tightly integrating network, application, and AI-specific security controls.

• At the core, Check Point Maestro Hyperscale Firewall provides elastic security with the ability to carve out separate Security Groups (SGs), functioning as independent virtual firewalls. Each SG can enforce distinct security policies—for example, one segment dedicated to DevOps/ administrative access, another to private API traffic, and another for public-facing inference services. This segmentation not only simplifies policy management but also ensures operational resilience by isolating policy domains, updates, and maintenance windows.

• On the application side, Check Point WAF integrates with AI Security, adding advanced protection tailored for AI environments.

• Check Point secures AI APIs and LLM endpoints by detecting prompt injection attacks, data exfiltration attempts, adversarial queries, and misuse of generative AI outputs, enabling safe consumption of AI services without exposing sensitive models or data.

• The design also clearly separates the management plane (MGMT VLAN), which handles Kubernetes, DevOps, and DGX operational traffic, from the data plane (Ethernet/InfiniBand) that carries training and inference workloads. This segregation enforces strict Zero Trust principles, ensuring that user/API traffic cannot bypass management protections.

• At the node level, NVIDIA BlueField DPUs integrated with Check Point DOCA FW and AppShield provide inline inspection and micro-segmentation for ingress/egress Ethernet traffic before it reaches the InfiniBand fabric.

SG2: NGFW Private Access

SG1: NGFW Management

DevOps / Ops / Admin / NVIDIA Ops

K8s / DGX operations traffic from Devs and Admins

North/South Security: K8s / DGX updates and LIBs download from the Internet

SG3: NGFW Public Access

Zero-Trust Network Access (ZTNA)

Check Point Maestro Hyperscale FW Cluster

Slurm Controller

InternetInternet (GitLab ;GitHub)

Check Point WAF / API / AI Security

MGMT VLAN MGMT VLANMGMT VLAN

GPU K8s

DPU Check Point Firewall

GPU K8s

DPU Check Point Firewall

GPU K8s

DPU Check Point Firewall

Training Models Cluster (DGX clusters)

Inference Clusters (DGX clusters)

Load Balancer

AI / API GW

DPU Check Point Firewall

GPU Slurm

DPU Check Point Firewall

GPU Slurm

DPU Check Point Firewall

GPU Slurm

East/West : Namespaces Traffic Control / K8s Runtime protection

CNI

North/South traffic inspection app data plane into the K8s Cluster

North/South API data traffic: Ethernet

Kubernetes / Container Security

Policies

DGX Server (ConnectX Dual

SuperNIC)

K8s : Kubernetes Cilium Nodes with Apps and

LLMs

NVIDIA BlueField DPU Check Point

Firewall

Check Point

East/West Lateral Movement / Segmentation Control / Access Control Zero trust per DevOps ID / IPS libs inspections / Command Execution analysis

Check Point embeds AI runtime security and prompt defense in all Check Point firewalls, Check Point WAF, and runs natively in AI Factory Firewall on NVIDIA BlueField DPUs

* Slurm = Simple Linux Utility for Resource Management

www.checkpoint.com

9AI DATA CENTER & AI FACTORY SECURITY BLUEPRINT

• In parallel, Check Point integrates with Illumio solution and enforces east-west traffic segmentation within Kubernetes clusters, controlling traffic between namespaces, pods, and services, as well as quarantining malicious containers.

• Together, this architecture unifies traditional firewalling, AI-specific application security, and in-fabric DPU enforcement into a Zero Trust AI Data Center blueprint, capable of protecting sensitive training data, inference APIs, and operational environments against both traditional and emerging AI-driven threats.

8. Use Cases and Security Components The following use cases highlight how Check Point secures AI Factories and data centers across training, inference, management, and application layers, while directly delivering a value:

Use Case Check Point Component Value

Segregating Training and Inference Domains and Servers Protection

Check Point NGFW running on DGX BlueField, enforcing inter-zone traffic policies and segregation of data and control planes.

Reduces risk of lateral movement, ensuring sensitive training data and production inference remain isolated and protected.

Zero Trust User and DevOps Access

Check Point Maestro NGFW Security Groups (SGs) create dedicated access policies for DevOps, admins, and external users.

Ensures least-privilege access, minimizes insider threats, and improves compliance posture.

AI Application Security Check Point WAF integrated with AI security validates API calls, protects against prompt injection, and secures AI-specific threats, DLP and more.

Protects AI applications from novel attacks, preserving customer trust and regulatory compliance.

Container and Namespace Segmentation and Protection

Check Point integration with Illumio enforces micro- segmentation across K8s namespaces and services. As well as integrating with Maestro NGFW via Logs API to block and quarantine infected workloads.

Prevents unauthorized east–west traffic inside clusters, reducing attack surface, protect from lateral movement inside K8s nodes.

Out-of-Band Threat Detection

BlueField DPUs mirror traffic to Check Point IDS/ Maestro for anomaly and behavior analysis without impacting performance.

Identifies threats early with minimal impact on GPU workloads, maintaining high system throughput.

Training Data Protection and Poisoning Detection

Check Point AI Agent Security + Check Point WAF (API validation) + Data FLow DLP Controls

Detects and blocks poisoned or manipulated training data, prevents ingestion of harmful or corrupted datasets, and ensures clean, trusted data flows into model training pipelines.

Adversarial Attack and Inference Abuse Detection

Check Point AI Agent Security + Check Point WAF behavioral analysis + Sig-based & ML-based anomaly detection

Identifies adversarial prompts, inference-time model manipulation attempts, and abnormal query patterns, protecting AI outputs from hijacking, jailbreaks, and harmful content generation.

Model Theft and Exfiltration Prevention

Check Point NGFW egress filtering + DLP + Bluefield DPU + Model Protection Policies

Prevents unauthorized export of model parameters, embeddings, or sensitive weights, blocks covert exfiltration channels, and detects abnormal

www.checkpoint.com

10AI DATA CENTER & AI FACTORY SECURITY BLUEPRINT

Use Case Check Point Component Value

Secure API Traffic Management

API traffic from AI apps is inspected by perimeter firewalls, WAF, and DOCA-based controls before entering inference clusters.

Protects against API abuse, supply chain vulnerabilities, and DDoS attacks.

Data Security and Compliance Monitoring

Check it Centralized telemetry from DGX, K8s, BlueField, integrated into Check Point management

Enables visibility, auditing, and reporting for compliance frameworks (AI Act, GDPR, HIPAA).

9. Runtime security for your GenAI Check Point AI Agent Security: AI-Native Runtime Guard for LLM Security

Capabilities

• Check Point AI Agent Security provides real-time visibility and control over GenAI applications by intercepting both inputs (prompts) and outputs (generated content).

• It supports customizable “guardrails” (predefined or custom) including prompt defenses, data leakage prevention, content moderation, malicious link detection, and policy-driven filters.

• It continuously evolves threat models via Check Point’s intelligence platform, updating detections (e.g. adversarial queries, injections) with low latency and minimal false positives.

• Centralized policy and control: You can define policies across multiple GenAI apps without touching application code, enabling consistent enforcement and monitoring.

• Low overhead, high performance: The solution is designed to add minimal latency to inference pipelines, making it usable in production-scale GenAI.

Organization

Internal Models

CISO AI Policies

Internet

Users

Applications

Integrations

Check Point Security Intelligence

Check Point AI Agent Security

Security Prompt injection

Data loss Infiltration

Threat blocking

Safety Harmful content

Data privacy Misinformation

Observability & Governance Alerts

Usage insights Policy violations

GenAI Applications

External Models

Use Cases in AI / Private LLM Environments

• Conversational Agents / Chatbots: Protects against prompt injection, exposure of system prompts or internal instructions, and out-of-bound responses.

• RAG / Document Agents: Secures retrieval-augmented generation pipelines, guarding against poisoned documents, tampered references, and unauthorized data leakages.

www.checkpoint.com

11AI DATA CENTER & AI FACTORY SECURITY BLUEPRINT

• GenAI Gateways / API Frontends: Used at the boundary of AI systems to centralize protection of API traffic, apply policies across services, and monitor AI service behavior.

• Connected Agents / Integrations: When AI agents call external services (APIs, third-party tools), Check Point ensures security over those agent workflows (blocking malicious API calls, ensuring flow integrity).

• Model Context Protocol (MCP) Architectures: Check Point can wrap MCP tool calls (prompts, resources, tools) to secure interactions and prevent injection or misuse in emerging AI frameworks.

Value & Impact for an AI Factory / Private LLM Security

• AI-specific defense layer: Check Point fills a gap that traditional security tools can’t cover — attacks crafted in natural language (prompt injections, subtle manipulations) that could lead to data leaks or model abuse.

• Separation of concerns / clean interface: Because Check Point functions through APIs at the application layer, it doesn’t require deep instrumentation in GPU nodes or container runtimes — simplifies deployment.

• Consistent policy across environments: You can apply the same security governing logic across multiple inference clusters, AI apps, and gateways, ensuring uniform protection and lowering policy fragmentation.

• Support for compliance & auditing: Logging, threat detection, and response features provide traceability for AI interactions. For audits, AI regulation compliance, or investigation.

• Low friction adoption: Minimal latency impact and code-agnostic integration make it viable in production AI environments — security without sacrificing performance.

• Adaptive security posture: Since threat models evolve, continuous intelligence updates help maintain protection against novel AI threats (jailbreaks, chained attacks) even after deployment.

www.checkpoint.com

12AI DATA CENTER & AI FACTORY SECURITY BLUEPRINT

10. Securing AI GPU Infrastructure Check Point Firewalls & NVIDIA BlueField DPU Integration

Check Point security is embedded directly into NVIDIA BlueField DPUs through the DOCA framework, allowing enforcement of inline firewall and microsegmentation policies at the hardware-accelerated NIC level. This provides high-performance inspection of north–south Ethernet traffic (ingress/egress) before it reaches the GPU/CPU workloads inside the DGX server. The integration extends Check Point’s advanced firewall and IPS protections without consuming host resources, preserving GPU cycles for AI workloads.

Key Features

• Hardware-accelerated FW/IPS: Runs Check Point’s next-gen firewall engine on the DPU, enabling low-latency policy enforcement. Zero negative impact on AI system performance.

• Microsegmentation: Isolates applications, VLANs, and tenants at the NIC level, limiting blast radius of compromise.

• Traffic Offload: Optionally mirrors traffic to Check Point Maestro IDS for anomaly detection and advanced behavioral analytics.

• Policy Orchestration: Managed centrally via Check Point SmartCenter, ensuring consistent rules across DPUs, DGX servers, and Kubernetes clusters.

Use Cases • Secure DevOps access to AI fabric

DGX nodes • Protection for MGMT VLAN • Block malicious code to be uploaded /

downloaded (GitHub / GitLab)

• Secure Application Data Plane (API) user access (NW)

• Runtime Threat Detection with DOCA • EW K8s access control and

micro-segmentation (Ilumio)

www.checkpoint.com

13AI DATA CENTER & AI FACTORY SECURITY BLUEPRINT

AppShield on BlueField

DOCA Argus is a DOCA service running on NVIDIA® BlueField® networking platforms, designed to immediately detect and enable response to attacks, minimizing their potential impact and risk.

The DOCA Argus framework provides real-time situational awareness and runtime threat detection by inspecting host memory using advanced memory forensics. Live machine introspection is performed at the hardware level, analyzing specific snippets of volatile host memory to monitor threats in real time without impacting system performance. DOCA Argus does not violate privacy, as information is extracted only from kernel structures.

Unlike conventional tools, Argus runs independently of the host, requiring no agents, integration, or reliance on host-based resources. This agentless, zero-overhead design enhances system efficiency and ensures resilient security in any compute environment, including bare-metal, virtualized, containerized, and multi-tenant infrastructures. By operating outside the host, isolated in its own trust domain, DOCA Argus remains invisible to attackers—even if the system is compromised.

Use Cases:

• Threat Intelligence Correlation (Hash Reputation) Extracts MD5/SHA hashes of executed files and checks them against Check Point ThreatCloud for malware detection.

• SBOM Integrity Enforcement Detects deviations from expected SBOM baselines, identifying tampered libraries, supply-chain attacks, or unauthorized components.

• Reverse Shell Detection Identifies suspicious process and syscall patterns that indicate reverse shells or post-exploitation C2 activity.

• Automated Remediation via Playblocks / SOAR When malicious behavior is detected, Check Point invokes and triggers playblocks or any SOAR system for remediation e.g. —blocking traffic and quarantining the compromised node / container, alerting SOC, enforcing security policies accordingly.

Illumio Containers Security

Illumio delivers deep visibility and micro-segmentation within Kubernetes clusters, operating via its C-VEN or agentless architecture to block lateral movement and isolate compromised workloads.

On detection of threats at the perimeter (for example, when the Check Point NGFW identifies bot traffic or access to a compromised website), NGFW logs can be forwarded to Illumio, enabling it to instantly quarantine affected containers and enforce enforcement policies at the node or pod level.

This combined approach ensures that an intrusion detected at the edge triggers immediate containment inside the cluster, preventing attacker propagation across containers or services. By linking perimeter threat prevention with runtime workload isolation, enterprises gain a robust, layered defense in containerized environments.

www.checkpoint.com

14AI DATA CENTER & AI FACTORY SECURITY BLUEPRINT

11. Protecting Public Cloud AI

In public cloud environments, AI workloads and private LLMs are deployed across managed Kubernetes platforms such as Amazon EKS, Azure AKS, and Google GKE. The same layered security approach used in AI data centers is applied here, ensuring protection across network, application, and workload levels.

Check Point Cloud Firewall delivers ingress/egress firewalling, segmentation, and Zero Trust access control, while Check Point AI Security provides AI-native WAF protection to defend against prompt injection, data exfiltration, and schema abuse.

To secure cloud-native workloads, Wiz CNAPP integration strengthens posture management, vulnerability detection, and runtime protection across containers and Kubernetes clusters, creating a unified and compliant security fabric for public cloud AI infrastructures.

Cloud AI vNET/VPC

Landing Zone

Internet

ABL Check Point WAF / AI Security

Public Cloud AWS / Azure / GCP

Check Point Cloud Firewall (IaaS) - NGFW/IPS/AV/DLP, TLS inspect,

egress control - Security Policies: User/API, DevOps/MGMT, Inter-VPC

- LLM Guardrails - Prompt injection, data exfil - Schema/abuse protection

Managed K8s (EKS/AKS/ GKE) Agent/Orchestrators LLM Server (vLLM/Triton/TensorRT) RAG Service + Embeds Vector DB (pgvector/Milvus/…)

CNAPP • CSPM • CWPP • CI/CD security • CIEM

Firewall (Ingress Zero Trust Access / Egress Hubs)

www.checkpoint.com

15AI DATA CENTER & AI FACTORY SECURITY BLUEPRINT

12. Security Technologies Full Stack for AI security

Check Point Management

Check Point WAF with AI Security for API security ingress / Prompt LLM protection

Threat Cloud AI

Check Point Maestro Hyperscale

Firewall

North / South API data plane protection / Out of the band / mirroring

/ Sampling Pattern network device

analysis (FW + IPS aka NDR)

NVIDIA BlueField DPU with CP DOCA FW (access control + IPS ) for ingress /egress traffic and DGX management

control plane

Kubernetes / Cilium eBPF /CNI access control (namespaces and pods

segmentation) secure by CP container FW / managed by CP MGMT rulebase

over API

NGFW Physical Appliances – secure communication between DC and

external DC + IPSEC

Check Point WAF

Check Point AI Security

CPRA. National Al Initiative Act, Al Bill of Rights

Mexico Al Strategy, Cuba's ChatGPT Ban

VARIOUS Al REGULATIONS AND LAWS, BY REGION (AS OF 2024)

India's IT Act, National Al Strategy

Australia Data Availability and Transparency Act

China's ChatGPT Ban, Chinese Federal Al Regulation. Shanghai Regulations for Al

EU Al Act. France National Al strategy, Swiss Federal Council's AlGuidelines, Italy's ChatGPT Ban

13. Alignment of Check Point AI Factory Security with AI Governance Frameworks Through 2026, at least 80% of unauthorized AI transactions will be caused by internal violations of enterprise / governance policies concerning information oversharing, unacceptable use or misguided AI behavior rather than malicious attacks (Gartner)

www.checkpoint.com

16AI DATA CENTER & AI FACTORY SECURITY BLUEPRINT

Check Point’s AI Factory security blueprint protects infrastructure and workloads and also enables governance, trust, and compliance per NIST AI RMF and Gartner AI TRiSM:

Framework Key Requirement Check Point AI DC Alignment

NIST AI RMF Governance & Transparency SmartCenter delivers centralized policy management, audit trails, and full visibility across training and inference domains.

Risk Mitigation BlueField DPU integration with Check Point Firewall and Kubernetes segmentation prevents oversharing, misuse, and lateral movement.

Lifecycle Security End-to-end protection from data ingestion and model training to inference APIs exposed to end users (Check Point AI Agent Security)

Gartner AI TRiSM Trustworthiness & Integrity Check Point and BF AppShield provide runtime protection, detect anomalies, and stop model/output manipulation.

Explainability & Accountability Unified logging and monitoring allow traceability of API calls, user activity, and container behavior to meet audit/compliance needs.

Policy Enforcement & Zero Trust NGFWs, container security - enforce Zero Trust at management and data planes, reducing insider and external risks.

Value: By embedding AI-aware controls into the network, container, and application layers, Check Point extends beyond perimeter security to address policy-driven AI risks, ensuring organizations can meet governance standards, pass audits, and deploy AI responsibly.

www.checkpoint.com

Worldwide Headquarters 5 Shlomo Kaplan Street, Tel Aviv 6789159, Israel | Tel: +972-3-753-4599 U.S. Headquarters 100 Oracle Parkway, Suite 800, Redwood City, CA 94065 | Tel: 1-800-429-4391 www.checkpoint.com

17AI DATA CENTER & AI FACTORY SECURITY BLUEPRINT

About Check Point Check Point Software Technologies Ltd. is a global cyber security leader protecting more than 100,000 organizations worldwide. Its mission is to secure enterprises’ AI transformation. Built on a prevention first approach and an open ecosystem architecture, Check Point helps organizations reduce risk, simplify operations, and innovate with confidence. The unified security architecture continuously adapts to evolving threats and expanding AI attack surfaces, protecting hybrid networks, cloud environments, digital workspaces, and AI systems. Structured around four strategic pillars, Hybrid Mesh Network Security, Workspace Security, Exposure Management, and AI Security, Check Point delivers consistent protection and visibility across complex multivendor environments.

14. Summary Check Point’s AI Factory Security Blueprint enables organizations to adopt AI with confidence by protecting sensitive models, data, and applications from misuse and regulatory risk. Leveraging Maestro, integrated AI Security, and NVIDIA DPU integration, the blueprint delivers business continuity, regulatory compliance, and customer trust while reducing the risk of data leakage, insider threats, and unauthorized AI use. This approach safeguards multi-million-dollar AI investments, accelerates secure innovation, and ensures AI services remain reliable and resilient at scale.

Learn More

View Executive Guide Speak with an Expert Check Point AI Factory Firewall

http://www.checkpoint.com https://engage.checkpoint.com/executive-guide-securing-the-ai-data-center-ai-factory https://engage.checkpoint.com/executive-guide-securing-the-ai-data-center-ai-factory https://pages.checkpoint.com/2026-mar-ww-ai-factory-and-data-center-security.html https://pages.checkpoint.com/2026-mar-ww-ai-factory-and-data-center-security.html https://pages.checkpoint.com/2026-mar-ww-ai-factory-and-data-center-security.html https://www.checkpoint.com/ai/cloudprotect/ https://www.checkpoint.com/ai/cloudprotect/ https://www.checkpoint.com/ai/cloudprotect/

White Paper | AI Data Center & AI Factory Security Blueprint

Item Type: pdf