Radware Adds Firewall for LLMs to Security Portfolio
None
<p>Radware has developed a firewall for large language models (LLMs) that ensures governance and security policies are enforced in real time.</p><p>Provided as an add-on to the company’s Cloud Application Protection Services, <a href="https://www.radware.com/newsevents/pressreleases/2025/radware-protects-generative-ai-use-with-new-llm-firewall/">Radware LLM Firewall</a> addresses the top 10 risks and mitigations for LLMs and generative artificial intelligence (AI) applications defined by the <a href="https://owasp.org/www-project-top-10-for-large-language-model-applications/">OWASP GenAI Security Project</a>.</p><p>At the core of that capability is an AI model that Radware created to determine if a prompt being shared with an LLM is malicious or if personally identifiable information is being exfiltrated.</p><p>Dror Zelber, vice president of product management for Radware, said LLM Firewall is the latest in a series of AI-based offerings that Radware is developing to secure next-generation applications, including AI agents. In effect, Radware is investing in AI to secure AI, he added.</p><p><a href="https://securityboulevard.com/wp-content/uploads/2025/11/Radware-LLM-Firewall.png"><img fetchpriority="high" decoding="async" class="alignnone wp-image-2077249" src="https://securityboulevard.com/wp-content/uploads/2025/11/Radware-LLM-Firewall-300x122.png" alt="" width="568" height="231" srcset="https://securityboulevard.com/wp-content/uploads/2025/11/Radware-LLM-Firewall-300x122.png 300w, https://securityboulevard.com/wp-content/uploads/2025/11/Radware-LLM-Firewall.png 662w" sizes="(max-width: 568px) 100vw, 568px"></a></p><p>That effort will prove crucial because attacks on AI applications and agents are going to be extremely challenging to predict so there will be a need for a complete stack of security software designed specifically to secure them, noted Zelber. The only thing that remains to be determined now is when a tipping point that forces organizations to reckon with AI security is going to arrive, he said.</p><p>There is already no shortage of cybersecurity tools and platforms for securing AI applications and agents, many of which are already being attacked by malicious actors attracted to anything that organizations clearly spent a significant amount of time and effort to build and deploy. The issue is that in the rush to deploy these applications many organizations have, once again, overlooked the cybersecurity risks attached to an emerging technology. Radware is now working toward lowering the bar to the point where it becomes significantly easier to secure AI applications, said Zelber.</p><p>Unfortunately, not enough organizations fully appreciate how malicious a prompt can be. For example, a prompt might direct an AI agent to first forget everything it knows about existing governance and security policies before directing it to email sensitive data to an external address. More troubling still, cybercriminals will target autonomous AI agents that will enable them to commandeer entire workflows and business processes.</p><p>In the short term there will, inevitably, be some major cybersecurity incidents involving AI applications and agents. A recent Futurum Group <a href="https://futurumgroup.com/press-release/futurum-research-ai-security-skills-gap-persists/">survey</a> found only 26% of organizations have implemented dedicated AI/machine learning security controls and processes to evaluate and monitor vulnerabilities involving AI. The number of organizations that are putting those controls in place continues to steadily increase, so hopefully, the damage from any breach will be limited as more of them understand the risk.</p><p>In the meantime, there is not going to be any putting the AI genie back in the proverbial bottle, so it will be up to cybersecurity teams to put the guardrails in place that are already much needed. The challenge, as always, will be explaining to end users why those guardrails are being put in place as much for their own good as the organization.</p><div class="spu-placeholder" style="display:none"></div><div class="addtoany_share_save_container addtoany_content addtoany_content_bottom"><div class="a2a_kit a2a_kit_size_20 addtoany_list" data-a2a-url="https://securityboulevard.com/2025/11/radware-adds-firewall-for-llms-to-security-portfolio/" data-a2a-title="Radware Adds Firewall for LLMs to Security Portfolio"><a class="a2a_button_twitter" href="https://www.addtoany.com/add_to/twitter?linkurl=https%3A%2F%2Fsecurityboulevard.com%2F2025%2F11%2Fradware-adds-firewall-for-llms-to-security-portfolio%2F&linkname=Radware%20Adds%20Firewall%20for%20LLMs%20to%20Security%20Portfolio" title="Twitter" rel="nofollow noopener" target="_blank"></a><a class="a2a_button_linkedin" href="https://www.addtoany.com/add_to/linkedin?linkurl=https%3A%2F%2Fsecurityboulevard.com%2F2025%2F11%2Fradware-adds-firewall-for-llms-to-security-portfolio%2F&linkname=Radware%20Adds%20Firewall%20for%20LLMs%20to%20Security%20Portfolio" title="LinkedIn" rel="nofollow noopener" target="_blank"></a><a class="a2a_button_facebook" href="https://www.addtoany.com/add_to/facebook?linkurl=https%3A%2F%2Fsecurityboulevard.com%2F2025%2F11%2Fradware-adds-firewall-for-llms-to-security-portfolio%2F&linkname=Radware%20Adds%20Firewall%20for%20LLMs%20to%20Security%20Portfolio" title="Facebook" rel="nofollow noopener" target="_blank"></a><a class="a2a_button_reddit" href="https://www.addtoany.com/add_to/reddit?linkurl=https%3A%2F%2Fsecurityboulevard.com%2F2025%2F11%2Fradware-adds-firewall-for-llms-to-security-portfolio%2F&linkname=Radware%20Adds%20Firewall%20for%20LLMs%20to%20Security%20Portfolio" title="Reddit" rel="nofollow noopener" target="_blank"></a><a class="a2a_button_email" href="https://www.addtoany.com/add_to/email?linkurl=https%3A%2F%2Fsecurityboulevard.com%2F2025%2F11%2Fradware-adds-firewall-for-llms-to-security-portfolio%2F&linkname=Radware%20Adds%20Firewall%20for%20LLMs%20to%20Security%20Portfolio" title="Email" rel="nofollow noopener" target="_blank"></a><a class="a2a_dd addtoany_share_save addtoany_share" href="https://www.addtoany.com/share"></a></div></div>