Organizations across the globe are rapidly deploying new AI applications and building new AI-ready data centers to automate and achieve operational excellence in their organizations. This requires ultra-highperformance for AI and large language model (LLM) inference environments to deliver real-time response, as well as new cybersecurity solutions to secure them.
To help prepare and protect these new AI environments, A10 Networks (NYSE: ATEN) is demonstrating new AI firewall and predictive performance capabilities at the upcoming Interop Tokyo conference, “AI Steps into Reality,” on June 11-13, 2025.
Preventing, Detecting and Mitigating AI and LLM-level Cyber Threats
A10 is announcing new AI firewall capabilities that can be deployed in front of APIs or URLs that expose large language models – either as a custom LLM or developed on top of a commercial solution like OpenAI or Anthropic. Built on edge-optimized architecture with GPU-enabled hardware, these capabilities protect AILLMs at high performance and can be deployed in any infrastructure as an incremental security capability.
These capabilities can help prevent, detect and mitigate AI-level threats by enabling customers to test theirAI inference models against known vulnerabilities and to help remove them using A10’s proprietary LLM safeguarding techniques. The capability detects AI-level threats like prompt injections and sensitiveinformation disclosure by inspecting request and response traffic at the prompt level and enforcing security policies required for mitigating these threats.
Delivering Real-time Experience for AI and LLM Inference Environments
A10 continues to deliver high performance and resilience for AI and LLM-enabled applications. This is done by offloading processor-intensive tasks like TLS/SSL decryption, caching, optimizing traffic routingand by providing actionable insights to maximize network availability and performance.
New capabilities allow early detection of network performance issues, much like an early warning system.This helps identify near-term congestion or capacity deficiencies, helping customers to take proactive action before an issue becomes critical. The capabilities help prevent unscheduled downtime and plan for optimal network performance. Predictive performance will run on A10 appliances that are powered by GPUs, allowing faster processing speed with the ability to quickly analyze vast amounts of data and provide insights into anomalies ahead of time.
Together, these AI security and infrastructure capabilities allow for ease of management, broaderintelligence to accurately detect threats, and help deliver an optimal customer experience.
“Enterprises are deploying and training AI and LLM inference models on-premises or in the cloud at a rapid pace. New capabilities must be developed to address three key challenges of these new environments: latency, security and operational complexity.
With over 20 years of experience in securing and delivering applications, we are expanding our capabilities to deliver on these needs to provide resilience, high performance and security for AI and LLMinfrastructures,” said Dhrupad Trivedi, president and CEO, A10 Networks.
