Securing the AI Frontier: Luigi Caramico's Vision for Encrypted Innovation at DataKrypto
Luigi Caramico, a veteran in the data protection industry, has been at the forefront of cybersecurity innovation for over two decades. As the founder and CTO of DataKrypto, Caramico is pioneering a new era of data security with fully homomorphic encryption (FHE) technology that promises to revolutionize how organizations protect their most sensitive information in the age of AI.
With a career spanning multiple successful ventures in data analytics and protection, Caramico’s journey from ethical hacker to encryption innovator has been driven by a singular vision: to create a world where data remains secure from creation to use, even during computation.
In an exclusive interview with AI World Today, Caramico delves into the groundbreaking FHEnom for AI™ solution, which combines FHE with Trusted Execution Environments to deliver unprecedented security for AI models and datasets. He explains how this zero-knowledge framework is addressing critical vulnerabilities in AI systems and enabling organizations to harness the full potential of AI without compromising on data privacy or regulatory compliance.
Can you tell us about your background and what led you to found DataKrypto?
My career has always been defined by advancing data protection and privacy-enhancing technologies. I started out hacking, quickly shifting into ethical hacking which then led me to work with law enforcement to help secure critical systems. Over the years, I’ve founded multiple companies focused on data protection, including DRSLab Inc. back in 2008 and Cloak Inc. in 2016. Earlier on, I led Intelligentias Inc. and served as CEO of Klaustech Group, co-founded Xalted Inc. (which later merged with Kranem), and served as board member and Senior Vice President at Kranem Corporation, where we focused on analytics for law enforcement and intelligence agencies. In 2021, I founded DataKrypto to tackle what I saw as a major gap in cybersecurity: traditional methods protect systems, but not the data or the AI models themselves. My vision was to create a world where data stays encrypted - from the moment it’s created to the moment it’s used. We’ve built continuous encryption solutions that make that possible, even for complex AI workloads like images, audio, and text. And by using advanced symmetric encryption, we can do it with near-plaintext performance - so security doesn’t come at the expense of speed.
DataKrypto is known for its fully homomorphic encryption (FHE) technology. Can you explain in simple terms what FHE is and why it’s important for data security?
FHE holds the key to true continuous encryption and end-to-end data security, but not all solutions are the same. Traditional FHE and hardware accelerators have fallen short, plagued by high costs, performance lags, and integration headaches. DataKrypto’s cutting-edge FHE technology breaks through these limitations, making real-time, seamless encryption a reality without costly add-ons. In summary, DataKrypto is the first and only encryption solution that protects all types of data throughout its lifecycle, including data-in-use, allowing modern users to work the way they are used to - at the speed of digital - supporting parallel processing needs of AI and LLM data, without compromising performance, security, or utility. Unlike competitive solutions, DataKrypto’s FHEnom™ does not require additional infrastructure investments. With FHEnom, encrypted data retains its original size, maximizing efficiency and reducing compute and storage costs, enabling massive scalability. Our innovative FHE solutions make continuous encryption at scale a reality.
Your company recently announced FHEnom for AI™. What are the key features of this solution and how does it address the security challenges in AI?
FHEnom for AI™ is a zero-knowledge AI framework that integrates FHEnom™ Fully Homomorphic Encryption (FHE) from DataKrypto with Trusted Execution Environments (TEEs) to deliver end-to-end privacy for AI workflows. FHEnom for AI™ serves as a critical layer in a holistic AI security strategy by protecting data and models during every stage of the AI lifecycle. While traditional security measures guard networks and endpoints, FHEnom for AI™ uniquely secures data and computations themselves, even in untrusted environments. This approach complements existing controls, closes key privacy and compliance gaps, and enables secure AI adoption across cloud, hybrid, and collaborative scenarios - making it an essential component for organizations seeking robust, end-to-end AI protection.
How does the combination of FHE and Trusted Execution Environments (TEEs) in FHEnom for AI create a zero-knowledge framework?
FHE enables computation on encrypted data (embeddings, weights, parameters, etc.), keeping model and user data protected, while TEEs provide hardware-enforced isolation for secure tokenization and output within a secure enclave. This approach allows entire AI pipelines to remain encrypted, enabling data to flow through standard AI models - including those for agents, retrieval-augmented generation (RAG), and multi-modal workflows - with minimal or no architectural changes.
DataKrypto recently partnered with Tumeryk to deliver encrypted guardrails for Operational AI. Can you elaborate on this collaboration and its significance?
Our partnership with Tumeryk is all about making AI safer and more trustworthy from the ground up. Together, we’ve built encrypted guardrails that bring privacy and security right into the way AI systems work. By combining technologies like FHE, TEEs, and policy enforcement, we can help both large and small language models stay compliant and tamper-resistant in real time. It’s also built to support emerging regulations, like the EU AI Act, so organizations can stay ahead of what’s coming. At the end of the day, it’s about ensuring AI runs with privacy, performance, and trust already built in - not added on later.
What are some of the industries or sectors that you believe can benefit the most from DataKrypto’s solutions?
Practically every industry that handles sensitive data can benefit from what we’re building at DataKrypto, but a few stand out. Sectors like healthcare, finance, and government have an urgent need for continuous encryption because they deal with highly regulated and confidential information. We’re also seeing growing demand in AI-driven industries - from defense and critical infrastructure to cloud providers and large enterprises that need to train or deploy AI models securely. Anywhere data privacy, compliance, and intellectual property protection intersect, our technology helps organizations keep their data encrypted end-to-end, without sacrificing performance or usability.
As the founder and CTO of DataKrypto, what’s your vision for the company’s future? Are there any exciting developments on the horizon?
As AI transforms every industry, data and model security have become the gatekeepers for adoption, trust, and innovation. Our vision is to make continuous encryption not just a feature, but the new baseline for enterprise AI. With FHEnom for AI™, we believe every piece of data, every inference, every agentic decision, and every model parameter should remain encrypted at all times - not just at rest and in transit, but also in use without compromising on the needs of the product & Security teams. We see a future where AI can be deployed anywhere - on-prem, cloud, hybrid, or edge - without ever exposing sensitive data or intellectual property. AI models and workflows are protected by default against prompt injection, data exfiltration, model inversion, and regulatory non-compliance. Organizations can now harness the full power of generative and agentic AI - knowing that privacy, compliance, and business continuity are guaranteed by cryptographic security, not just policy or perimeter defenses.
Can you share an example of how a company has successfully implemented DataKrypto’s technology to enhance their data security?
Our largest single deployment is with CoopVoce, a telecom customer with approximately 2.2 million users generating around 500 million Call Detail Records (CDRs) and Internet Data Records (IDRs) daily. They use DataKrypto’s FHEnom solution to encrypt all sensitive telecom data fields via Fully Homomorphic Encryption, seamlessly integrated into their existing ingestion pipelines. This deployment is measured by daily record volume and user base size—specifically, 2.2 million subscribers and half a billion records ingested and encrypted every day. Our solution operates within their current infrastructure without requiring additional compute or storage resources from their side, maintaining throughput and operational efficiency while ensuring end-to-end data privacy compliance and enabling secure downstream analytics on encrypted data. This scale demonstrates our ability to support high-volume, privacy-critical telecom environments with advanced encryption technologies.
How does DataKrypto’s technology help organizations comply with data protection regulations like GDPR and HIPAA?
The widespread adoption of open-source and proprietary AI models has introduced critical vulnerabilities. Companies share sensitive corporate data with these models to improve decision-making, automate workflows, and unlock operational gains - but current AI tools lack robust security capabilities to protect against malicious access, data breaches, and adversarial attacks like model poisoning. These vulnerabilities allow hackers to manipulate AI outputs, compromise decision-making, and expose organizations to regulatory penalties for violating data privacy mandates such as GDPR and HIPAA. FHEnom for AI safeguards both customized open-source AI models (adapted for specific enterprise needs) and proprietary models, while ensuring the privacy and protection of sensitive data. It defends against malicious activities, misuse, and adversarial manipulation, securing intellectual property and maintaining the confidentiality of critical datasets. This capability establishes trust and reliability in AI-generated outputs, giving enterprises confidence in AI to make accurate, safe, and fair decisions. Our solution removes barriers that prevent companies from fully embracing AI, given data privacy and model security concerns. While AI promises to drive significant advancements across industries, many companies have strictly limited its use for fear of their data falling into the wrong hands. Their concerns are warranted. Until now, AI has not been secure. Sensitive data, intellectual property, and the AI models companies build and train using their data have been vulnerable. DataKrypto is changing that by enabling fine-tuning of AI models on encrypted data, at near-plaintext speeds, we are making it safe for companies to leverage AI to innovate new solutions and solve the most complex challenges.
Your background includes founding several companies in the data protection space. How have these experiences shaped your approach with DataKrypto?
Our flagship solution FHEnom for AI™, integrates FHE with trusted execution environments (TEEs) to enable computation on encrypted data without decryption. This innovation secures every stage of AI processing - from queries and embeddings to model weights and responses - ensuring confidentiality even if systems are compromised. By eliminating the need to decrypt, FHEnom for AI also enforces a zero-knowledge framework, meaning neither the data owner nor the AI provider gains visibility into the underlying content of inputs or outputs.
DataKrypto claims to have the fastest FHE solutions on the market. Can you elaborate on the performance aspects of your technology?
DataKrypto delivers fully homomorphic encryption (FHE) that operates at plaintext speeds - something no other FHE solution has achieved. This breakthrough is especially significant now, as AI adoption is driving organizations to process and analyze massive volumes of sensitive data at unprecedented scale and velocity. Traditional FHE solutions are far too slow for practical AI use, with performance gaps ranging from 1,000 to 100,000 times slower than plaintext computation, effectively disqualifying them from real-world AI applications. In contrast, DataKrypto’s technology allows data and AI models to remain encrypted even during computation, enabling privacy-preserving AI without sacrificing performance. This makes it possible, for the first time, to protect data in use at the exact moment it is most vulnerable with real time inferencing, aligning security and product teams needs with the demands of today’s data-driven, AI-powered enterprises.
As AI continues to advance, what do you see as the most critical security challenges that need to be addressed?
When I look at how AI is evolving, I see a historical pattern in that innovation often moves faster than security. Every major technological leap creates new opportunities, but it also opens new gaps that defenders have to scramble to close. AI is no different, and in many ways, it’s amplifying that challenge. One of the biggest ongoing issues will be protecting intellectual property and personal data as AI becomes more deeply embedded in everything we do. That’s why I believe encryption that protects data in use and in process - not just at rest or in transit - is going to be critical for keeping pace with AI’s rapid growth.
As AI continues to reshape industries and drive innovation, Caramico’s work at DataKrypto stands at the intersection of cutting-edge technology and critical security needs. With a vision for a future where continuous encryption becomes the new baseline for enterprise AI, Caramico and his team are not just solving today’s security challenges – they’re anticipating and preparing for the complex data protection landscape of tomorrow. As organizations grapple with the dual imperatives of AI adoption and data security, DataKrypto’s innovations offer a promising path forward, ensuring that the benefits of AI can be realized without compromising on privacy or trust.