Aqua Security, the pioneer and leading pure-play cloud-native security company, is making waves at Black Hat 2025 with groundbreaking announcements that address the most pressing cybersecurity challenges facing organizations today. As enterprises rapidly integrate generative AI and large language models into their applications, Aqua has positioned itself at the forefront of this new battleground with its comprehensive platform that secures the entire software development lifecycle from code to cloud to prompt. The company’s deep understanding of how modern applications are built, deployed, and attacked�especially in cloud and containerized AI environments�has enabled them to develop solutions that actively stop cloud-native and AI attacks in real time.
This year’s Black Hat presence showcases Aqua’s latest innovations, including live demonstrations of their AI workload security capabilities and the unveiling of their strategic partnership with Akamai, which delivers the industry’s first joint solution for AI workload and prompt protection across the stack. With more than 70% of AI applications built and deployed in containers according to Gartner, Aqua’s unique position as a container security pioneer makes them uniquely suited to solve the emerging challenges of AI security. Attendees should prioritize visiting Aqua’s booth to witness firsthand how they’re closing the critical gap in protecting AI workloads with no SDKs or code changes needed�a capability that no other vendor can match in today’s rapidly evolving threat landscape.
Ahead of Black Hat, don’t miss this exclusive VMblog Q&A with Matt Richards, CMO, Aqua Security. Make sure you get Aqua Security on your MUST SEE list at the show.
++
VMblog: For readers who may not be familiar, can you give us the elevator pitch on your company and what makes you stand out in today’s cybersecurity landscape?
Matt Richards: Aqua Security is the pioneer and leading pure-play cloud-native security company. We provide the most comprehensive platform to secure applications, protecting the entire software development lifecycle. What sets us apart is our deep understanding of how modern applications are built, deployed and attacked-especially in cloud and containerized AI environments. Our solution actively stops cloud native and AI attacks from day zero in real time, from code to cloud to prompt.
The solution integrates directly into the CI/CD pipeline and works across all cloud native and AI workloads and environments. It scans container images, VM images and functions for known vulnerabilities, embedded secrets, malware and misconfigurations. Aqua then takes this to the next level with runtime protection, applying and enforcing security policies informed by real-world cybersecurity research and in-the-wild threat analysis of the cloud native stack beyond previously learned behaviors.
VMblog: Black Hat attendees expect cutting-edge demonstrations and hands-on experiences. What innovative booth activities or live demos are you planning that will give visitors a real taste of your technology in action?
Richards: At Black Hat this year, we will have two big announcements. The first is our AI workload security story and live demo. This will highlight our unique AI security capabilities to help customers protect their AI workloads in runtime, including the all-new AI Advisory Program. All of this will be shown in real time at our kiosk. Related to this, but separate, will be our new strategic partnership with Akamai. We will show our AI workload security capabilities integrated with Akamai’s Firewall for AI Applications, providing full edge and workload-level security. Both will be available at the Black Hat AI Summit and in Akamai’s booth at Black Hat.
VMblog: If a CISO visits your booth and walks away remembering just one thing about your company, what should that key message be?
Richards: If you’re a security leader responsible for protecting containers, Kubernetes environments or generative AI models at scale, the key message is this: Aqua closes the gap in protecting the workloads where you run AI with no SDKs or code changes needed. Today there is a gap between existing solutions and the real world of AI workload security. With Aqua, you can discover and protect your AI estate from initial model visibility to policy creation and enforcement so you can not only see threats, but also stop threats that bypass traditional controls – and do it in real time.
Aqua is the only company that delivers unified, real-time protection for every stage of the cloud native and AI application lifecycle from code to cloud to prompt. While many vendors focus on static analysis or CSPM, Aqua protects running workloads and active AI applications in production, where the most critical threats unfold in real time.
VMblog: With AI attacks, supply chain vulnerabilities, and cloud security challenges dominating headlines in 2025, which of these threat vectors does your solution specifically address or focus on?
Richards: Aqua specializes in all three of these areas, but our recent investments focus on threats to AI applications. We recently launched Secure AI, which delivers full lifecycle security from code to cloud to prompt.
All AI models have one thing in common: a workload must interact with them. Aqua resides in the workload, and Aqua can see all model interactions inside and outside the organization. Building on years of experience securing cloud-native environments, Aqua has applied those learnings to the new and urgent challenge of AI-powered application security. Traditional tools were never designed to understand prompt-level behavior or track the use of LLMs within applications. Aqua bridges this gap by innovating at the application layer with runtime visibility and protection explicitly built for AI.
Development of this solution was driven by real-world customer demand. As enterprises began integrating LLMs and AI copilots into their environments, they struggled to understand new threat vectors and enforce safe usage. Through market research, customer feedback and internal threat modeling, Aqua designed a solution that doesn’t just patch over AI risks, it understands and mitigates them. Secure AI is a fully integrated capability of the Aqua Platform and part of our broader CNAPP offering. It remains the only solution that delivers real-time, application-layer protection for AI workloads across hybrid and multi-cloud environments.
VMblog: The “AI security arms race” is a hot topic this year. How is your company leveraging AI defensively, and what’s your take on the AI-powered threats we’re seeing emerge?
Richards: We’re already in the midst of the AI security arms race, and the threats are evolving faster than traditional tools can keep up. Our Nautilus research team recently uncovered a sophisticated Linux malware campaign that used generative AI to produce part of the malicious code and concealed it inside a seemingly innocent panda image. This was a persistent threat that auto-executed payloads and maintained stealth via legitimate-looking cron jobs and hidden backdoors. The attackers likely used LLMs to craft obfuscated payloads and vary their techniques dynamically, hallmarks of AI-enabled malware. That kind of polymorphism and speed is exactly what makes AI-powered threats dangerous.
At Aqua, we’re taking a two-pronged approach to this rising challenge: defending AI workloads with runtime protection and using AI offensively and defensively to uncover and stop novel threats. Aqua Secure AI provides real-time detection at the application and workload layer. It can observe model behavior, track prompt activity and detect anomalies like the misuse of AI-generated binaries or command execution attempts that originate from within containerized environments. Our use of eBPF gives us visibility deep into runtime, even when attackers attempt to hide behind AI-generated scripts or masked payloads. This is the one place in the stack where AI use can be identified and secured because it is the one place AI, and attacks on AI, cannot hide.
The takeaway: AI is accelerating both sides of the arms race. Attackers are already using generative models to build evasive threats. Defenders must move just as fast with solutions built to observe, contextualize and stop malicious activity at the point of execution. Aqua’s platform is purpose-built for this fight.
VMblog: Are you unveiling any major product announcements, partnerships, or technological breakthroughs at Black Hat 2025?
Richards: Yes, we’re highlighting our strategic partnership with Akamai, which delivers the industry’s first joint solution for AI workload and prompt protection across the stack. We’re also showcasing Trivy’s new partnership with Root for automated, agentic AI-powered remediation, reinforcing our leadership in open source and cloud-native developer security. Root joins the newly announced Trivy Partner Connect program, which ensures continued investment in reliability and cutting-edge capabilities the community has come to expect from the world’s most popular security scanner. Our first two partners to join the program were Echo and Minimus, both delivering secure-by-design image solutions that align with Trivy’s mission of empowering developers and security teams through open, trusted tools.
We’ll also share more about our recently unveiled Secure AI solution and the Secure AI Advisory Program for this new solution. We’ll be talking a lot about this during the show, offering demos and seeking new advisory program members.
VMblog: What’s the biggest cybersecurity blind spot you’re seeing organizations struggle with in 2025, and how does your solution illuminate that gap?
Richards: As organizations rapidly integrate generative AI and LLMs into applications, security teams are struggling to keep pace with new risks. AI capabilities are often embedded into services without security oversight, making it difficult to track where models run, what data they process or whether usage complies with internal policy. Traditional infrastructure and network tools fall short because they cannot observe prompt activity, model behavior or data flows within the application itself. As a result, AI risk often enters production environments undetected.
Here’s our take: The key to securing AI is securing containers. As organizations deploy AI capabilities across modern application stacks, most use containers and Kubernetes. According to Gartner, more than 70 percent of AI applications are built and deployed in containers, running on cloud native infrastructure. This means that containers are the single point of control for AI security. As a pioneer in container security – one of the first and continuously most effective solutions on the market – Aqua is uniquely suited to solve the challenges of AI security.
VMblog: Beyond your product pitch, what’s one piece of actionable cybersecurity wisdom you’d share with every Black Hat attendee who stops by your booth?
Richards: AI may be new, but AI security is only as strong as the more traditional security base on which it is built.
VMblog: Does your team have any speaking engagements, workshop sessions, or research presentations at Black Hat that attendees should prioritize?
Richards: This is one of my favorites: “The Hidden Path to Root: Shadow Resources, Roles and AWS Service Exploits.”
Yakir Kadkoda, from the Aqua Nautilus research team, will talk about:
- Exploit predictable S3 buckets used by AWS services (Shadow Resources)
- Abuse default IAM roles with overly permissive policies (Shadow Roles)
- Escalate privileges inside the AWS Glue managed environment
We will explore how overlooked design flaws in cloud environments can create unexpected paths to serious security risks.
DEF CON 33 – Cloud Village from 10:50 to 11:30 am PT on Aug. 8.
VMblog: What’s the most creative or memorable giveaway item you’re bringing to your booth this year?
Richards: My favorite? A new luggage tag, which can be tracked as a device in “Find My” on your iPhone. It helps me find my suitcase at the airport all the time.
VMblog: As someone who’s invested in the Black Hat community, what’s your best advice for first-time attendees to maximize their conference ROI?
Richards: Don’t just listen – engage. Figure out what problem(s) you want to solve and go solve it. Ask vendors to prove their claims live. Push for real demos. Network as much as possible with your peers. And attend talks outside your immediate domain. You’ll often find your next strategy in someone else’s problem space.
##
