Hardware-Defined Security: The Evolving Landscape of Confidential Computing

Confidential Computing (CC) is undergoing a pivotal transformation, anchored increasingly in hardware as the trusted foundation for securing sensitive workloads. This evolving landscape is shaped by the drive for greater data integrity, hardware-enforced trust, and the growing applicability of CC in high-throughput applications like Artificial Intelligence (AI). While the Confidential Computing Consortium (CCC) has yet to formalize a universal definition, prevailing industry understanding now centers on the role of hardware-based Trusted Execution Environments (TEEs) in protecting data in use.

Market Overview

Confidential Computing (CC) continues to gain traction, though adoption is still nascent. Revenue from CC implementations remains modest through 2026 due to high setup complexity, a lack of enterprise awareness, and the absence of out-of-the-box solutions. Specialized expertise is required to manage enclave attestation, cryptographic asset handling, and data privacy policy enforcement. Nonetheless, the CCC’s efforts, alongside initiatives from standardization bodies like the Internet Engineering Task Force (IETF), have primed the market for broader uptake.

A key growth catalyst will be the expanding use of CC within Graphics Processing Units (GPUs), enabling larger-scale AI and Machine Learning (ML) workloads. GPU-based CC, exemplified by NVIDIA’s solutions, is poised to trigger renewed market interest, particularly as high-performance compute applications demand higher throughput.

(Source: ABI Research Charts & Data)

“The CCC defines Confidential Computing as the protection of data in use by performing computations in a hardware-based, attested Trusted Execution Environment (TEE), justified based on the enhanced performance and down-to-the-silicon security proffered by hardware TEEs.”

Diverging Definitions Pose Strategic Risk

Despite the CCC’s influence, a lack of formal standardization creates ambiguity. Disagreements among consortium members may hinder the scale of CC. For example, Amazon Web Services (AWS) is promoting Nitro Enclaves, which diverges from hardware-only interpretations. Vendors must navigate these discrepancies carefully to avoid mismatched expectations about CC’s security guarantees.

Further, firmware associated with hardware TEEs plays a critical role in CC assurance. As vulnerabilities emerge, firmware updates at the edge level are vital to maintaining enclave integrity, necessitating their inclusion within the scope of CC definitions and protections.

The GPU Turn: A New Performance Frontier

The shift from CPU-centric to GPU-centric CC architectures marks a significant inflection point. GPUs offer increased throughput and are better suited for resource-intensive applications such as Large Language Models (LLMs). NVIDIA’s hardware roadmap, transitioning from the H100 to the GB200, exemplifies the performance leap expected in CC environments.

This architectural evolution aligns with innovation around heterogeneous collaborative TEEs that integrate Field-Programmable Gate Arrays (FPGAs), Network Processing Units (NPUs), and Network Interface Cards (NICs). The future of CC points to system-wide trust architectures, rather than isolated CPU-based security islands.

Post-Quantum Considerations Enter the Equation

As the global cryptography landscape prepares for the era of quantum computing, CC’s future-proofing depends on early integration of Post-Quantum Cryptographic (PQC) mechanisms. Although symmetric encryption used in TEEs is quantum-resistant by design, attestation protocols reliant on asymmetric encryption are at risk. Consequently, integrating post-quantum encryption at the silicon level and incorporating crypto-agility into enabling software will become essential—particularly for vendors operating in regulatory-heavy environments.

Forecasts and Revenue Trajectory

From 2024 to 2026, CC revenue is projected to grow slowly, restrained by cost, complexity, and a general lack of market awareness. Software will dominate revenue in this phase, serving as the key enabler that lowers the entry barrier for CC adoption. By 2027, growth will accelerate as standards mature and legal mandates reinforce CC’s role in cybersecurity compliance.

The transition from confidential CPUs to confidential systems—extending trust protections to storage, AI processors, and system peripherals—will drive new use cases, particularly in edge and serverless computing. GPU-based CC will gain momentum as workloads scale and as CC protection becomes integral to broader Information Technology (IT) architectures.

From 2030 onward, CC will benefit from widespread industry education, decreasing setup friction and expanding cross-vertical penetration. Revenue is forecast to grow at a 44% Compound Annual Growth Rate (CAGR) from 2024 to 2032, with software and services outpacing hardware due to their critical role in enablement, customization, and compliance management.

Strategic Recommendations for Security Vendors

• Support Research and Acknowledge Hardware Limitations: Vendors should proactively support research into mitigating physical and side-channel attacks, recognizing CC’s limitations in preventing direct hardware manipulation. Transparency on such limitations, combined with collaborative development of Privacy-Enhancing Technologies (PETs), will help accelerate CC maturity.
• Strengthen GPU-Centric Partnerships and Trust Pipelines: As GPU-based CC gains prominence, partnerships with GPU leaders like NVIDIA—as well as initiatives to bridge trust gaps between the CPU and GPU—will be key. Vendors like Intel are already exploring bounce-back methods to secure this trust pipeline.
• Prepare for PQC Compliance: Ahead of looming regulatory deadlines, including the EU Cyber Certification (EUCC) scheme in 2030 and the Commercial National Security Algorithm (CNSA) suite from the U.S. government, early experimentation with post-quantum implementations is essential. This is particularly true for use cases in the Internet of Things (IoT), smart cards, and national identity documents, where symmetric and asymmetric cryptography both play roles.
• Prioritize System-Level Confidential Computing Architectures: Finally, to remain competitive, vendors must adopt a system-level view of CC. This means architecting for hardware-to-hardware trust across devices, especially in AI-dominated environments where performance and security must go hand-in-hand.

Key Companies

• Amazon Web Services (AWS)
• AMD
• Anjuna
• Arm
• Google Cloud
• Intel
• Microsoft Azure
• NVIDIA
• PQShield
• PsiQuantum
• SEALSQ

Get the full report

For a deeper look at the evolving architecture, standards, and competitive dynamics of Confidential Computing, download the full ABI Research report. Discover how hardware, software, and services providers can capitalize on the next wave of secure computing innovation.