Fortanix has been recognized as a Sample Vendor in five 2026 Gartner Hype Cycle reports, highlighting confidential computing and Key Management as a Service (KMaaS) as enterprises rethink security for AI infrastructure. The recognition spans data security, digital sovereignty, privacy, compute and telco cloud services, underscoring a broader shift toward protecting sensitive data and AI workloads across increasingly distributed infrastructure.
AI infrastructure is creating a security problem that traditional perimeter controls and encryption-at-rest strategies were not designed to solve.
Enterprises are increasingly moving sensitive information, proprietary models and AI workloads across public cloud, private infrastructure, accelerated computing environments and sovereign AI deployments. As those systems become more distributed, security teams need to control not only where data is stored and how it moves, but also what happens to data while it is being processed.
That is where confidential computing is becoming increasingly relevant.
Fortanix says it has again been recognized as a Sample Vendor in five 2026 Gartner Hype Cycle reports: Hype Cycle for Data Security Technologies, Hype Cycle for Digital Sovereignty, Hype Cycle for Privacy, Hype Cycle for Compute and Hype Cycle for Telco Cloud Services.
The company was recognized in the Confidential Computing and Key Management as a Service (KMaaS) categories.
The distinction matters because the two technologies address different but complementary parts of the AI security problem.
Key management controls the cryptographic material that protects data and applications. Confidential computing is designed to protect data while it is actively being processed inside trusted execution environments.
Gartner’s June 2026 research describes confidential computing as a technology that protects data in use within secure hardware enclaves and says it can be important for regulated industries and confidential AI. Gartner also cautions that its complexity and specific focus mean it should be deployed as part of a broader defense-in-depth architecture rather than treated as a universal security solution.
AI Is Changing the Cryptographic Problem
Traditional enterprise encryption strategies generally divide protection into three states: data at rest, data in transit and data in use.
The third category is becoming more important as AI workloads become more sophisticated.
An AI model may need access to customer records, proprietary intellectual property, financial information or other sensitive datasets during training and inference. Encrypting the data while it is stored does not by itself prevent exposure while authorized workloads are processing it.
Confidential computing attempts to close that gap through hardware-enforced isolation.
For enterprises, the appeal is straightforward: sensitive information can be processed within protected execution environments while reducing the amount of trust placed in the underlying infrastructure.
That becomes particularly relevant when workloads run on infrastructure operated by cloud providers or other third parties.
Fortanix Connects Confidential Computing With Key Control
Fortanix is positioning its technology around combining confidential computing with cryptographic management.
Its Fortanix Data Security Manager (DSM) brings together hardware security module capabilities, key management, certificate management and secrets management. The company says the platform is designed to provide centralized control across on-premises and cloud environments while supporting crypto-agility and automation.
That last point is becoming increasingly important.
AI infrastructure is changing rapidly, and organizations increasingly need to move between cloud environments, processors, workloads and cryptographic technologies without rebuilding their security architecture every time.
Crypto-agility refers to the ability to change cryptographic algorithms, keys and associated controls without extensive disruption.
For enterprises planning long-lived AI infrastructure, that flexibility could become a practical requirement rather than a theoretical security feature.
Confidential AI Extends Protection to AI Workloads
Fortanix extends its confidential-computing approach into AI through what it calls Confidential AI.
The company says the technology is designed to protect sensitive data and proprietary AI models during training, fine-tuning and inference.
The architecture combines hardware-enforced isolation with CPU and GPU attestation and secure key release.
Attestation is important because it provides a mechanism for verifying that a workload is running within an expected, trusted environment before sensitive cryptographic keys are released.
That creates a different security model from simply encrypting a database or network connection.
The question becomes not only whether the data is encrypted, but also whether the system processing that data can prove that it is running in an approved environment.
For AI workloads involving proprietary models or highly sensitive datasets, that distinction could become increasingly important.
Sovereign AI Adds Another Layer
Digital sovereignty is another reason confidential computing and centralized cryptographic control are attracting attention.
Governments and enterprises are increasingly concerned about where data is stored, which jurisdictions govern it, who controls the infrastructure and whether sensitive workloads can be exposed to external operators.
Sovereign AI initiatives intensify those concerns because AI systems can involve sensitive government, industrial and citizen data.
Gartner’s 2026 Hype Cycle research also reflects the growing relationship between AI, sovereignty and security. Its Data Management Hype Cycle notes that AI-related ambitions are driving innovation around governance and sovereignty, while its China cybersecurity research highlights technological sovereignty alongside the need to secure generative and autonomous AI.
Confidential computing does not automatically solve every sovereignty requirement. Data residency, jurisdiction, identity, infrastructure ownership and regulatory controls remain separate considerations.
But it can provide another technical control for reducing exposure while workloads are being processed.
Why Hardware-Backed Security Matters for AI
The move toward AI factories and accelerated computing is also changing the underlying hardware stack.
GPUs and specialized accelerators increasingly sit alongside CPUs, networking infrastructure and high-speed storage. Sensitive AI workloads can therefore span several layers of infrastructure.
Protecting only the CPU environment may not be enough if data moves into accelerators during inference or training.
Fortanix’s emphasis on CPU and GPU attestation reflects this emerging requirement.
The broader market is moving in the same direction, with cloud and semiconductor providers developing confidential-computing capabilities around trusted execution environments and accelerator infrastructure.
The strategic goal is to establish a chain of trust extending from hardware through the workload and into the cryptographic controls governing access to sensitive information.
Gartner Recognition Is Not a Product Endorsement
The Gartner recognition should also be viewed in the correct context.
Being named a Sample Vendor in a Hype Cycle report does not mean Gartner recommends Fortanix or validates its products as the best solution.
Gartner itself explains that its Hype Cycle framework is intended to help organizations evaluate emerging technologies and understand their maturity, risks and potential business value.
That distinction is especially important for emerging technologies such as confidential computing.
The technology is gaining relevance, but enterprise adoption still involves questions around application compatibility, hardware availability, performance overhead, cloud support, operational complexity and integration with existing security systems.
Fortanix’s recognition across five Hype Cycle reports nevertheless illustrates how broadly these technologies are beginning to intersect with enterprise infrastructure.
AI Security Is Becoming an Infrastructure Discipline
The larger shift is that AI security is moving below the application layer.
Enterprises initially focused heavily on model security, prompt injection, data leakage and AI application governance. Those remain important, but increasingly powerful AI systems require security controls throughout the infrastructure stack.
That includes:
Hardware → trusted execution → cryptographic keys → data → AI model → inference workload → cloud infrastructure
Each layer introduces a different trust boundary.
The challenge for security teams is making those controls work together without creating an infrastructure environment that is too complex to operate.
This is where Fortanix’s strategy is aimed: combining hardware-backed cryptography, enterprise key management and confidential computing into a more unified architecture.
Competition Will Extend Beyond Traditional HSMs
Fortanix is operating in a market that includes established hardware security module providers, cloud key-management services, confidential-computing technologies and specialized data-security vendors.
Cloud providers such as Microsoft Azure, Google Cloud and AWS already offer combinations of encryption, key management and confidential-computing capabilities.
That creates a difficult competitive environment for independent security platforms.
Fortanix’s opportunity is to serve enterprises that need a common cryptographic and data-security control plane across multiple clouds, on-premises systems and specialized AI infrastructure.
The value proposition is therefore less about replacing a single HSM or encryption service and more about centralizing control across increasingly fragmented infrastructure.
The Next AI Security Battle Is About Trust
The most important implication of Fortanix’s announcement is that AI security is becoming increasingly tied to infrastructure trust.
As organizations move from experimentation to production AI, they will need to demonstrate that sensitive data and proprietary models remain protected throughout the AI lifecycle.
Confidential computing offers one piece of that architecture.
Key management provides another.
Neither is sufficient on its own.
But together, they point toward a security model in which enterprises can verify the environment processing sensitive data, control the cryptographic keys that unlock that data and maintain visibility across cloud and on-premises infrastructure.
That could become increasingly important as AI factories, sovereign AI environments and accelerated computing become standard components of enterprise technology strategies.
Fortanix’s presence across five Gartner Hype Cycle reports is therefore less significant as a standalone recognition than as a marker of a larger market transition: AI infrastructure is becoming too valuable and too sensitive to secure with conventional perimeter controls alone.
Market Landscape
The AI security market is increasingly converging across data protection, cryptography, confidential computing and infrastructure security.
- AWS, Microsoft Azure and Google Cloud provide confidential-computing, encryption and key-management capabilities directly within their cloud ecosystems.
- Hardware and semiconductor vendors are adding trusted execution and confidential-computing capabilities to CPUs and accelerators.
- Traditional HSM providers remain central to enterprise cryptographic key protection.
- Data security platforms are expanding into AI governance and protection as organizations give AI systems access to sensitive information.
- Fortanix is positioning itself across these layers by combining HSM technology, key management, secrets management and confidential computing.
Gartner’s 2026 research reinforces the direction of travel: confidential computing is being considered in the context of AI, privacy and regulated workloads, while the broader compute market is being reshaped by AI infrastructure investment and agentic workloads.
Top Insights
- Fortanix was recognized as a Sample Vendor across five 2026 Gartner Hype Cycle reports covering data security, sovereignty, privacy, compute and telco cloud.
- The company was recognized in Confidential Computing and Key Management as a Service, two technologies increasingly relevant to AI infrastructure.
- Confidential computing protects data while it is processed, addressing a security gap left by encryption at rest and in transit.
- AI factories and sovereign AI deployments are increasing demand for centralized cryptographic control and verifiable workload environments.
- Enterprise adoption will depend on integrating confidential computing with existing identity, cloud, key-management and security architectures.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
