2025

Adopting on-demand computing for sensitive and private workloads is critical in today’s interconnected and data-driven world. We need simple, fast, and reliable security mechanisms to protect data, code, and runtimes to achieve this. The Confidential Computing Consortium’s new Messaging Guide is a comprehensive resource that explores how Confidential Computing (CC) addresses these challenges and supports organizations in securing their workloads.

Confidential Computing capabilities protect against unauthorized access and data leaks, enabling organizations to collaborate securely and comply with regulatory requirements. By encrypting data at rest, in transit, and during processing, CC technology allows sensitive workloads to move to the clud without requiring full trust in cloud providers, including administrators and hypervisors.

This white paper outlines the motivations, use cases, and solutions made possible by Confidential Computing, empowering organizations to:

Who Should Read This Guide?

This document is tailored to meet the needs of a diverse audience, including:

• Organization leaders to explore use cases and services that can be enabled by Confidential Computing.
• Organization leaders considering whether to use Confidential Computing for securing a new or existing
• product(s), projects, services, and capabilities
• Regulators, standards bodies and ecosystem members in Data Privacy and related fields.
• General public/Mainstream Media/Publication to raise general awareness of Confidential Computing and its benefits

Why This Matters

As the demand for secure data processing grows, Confidential Computing provides a critical solution to meet the challenges of modern cloud environments. Whether enabling secure AI applications, fostering inter-organizational data collaboration, or addressing compliance needs, CC empowers organizations to innovate without compromising security.

We invite you to explore the Messaging Guide and discover how Confidential Computing can transform your approach to secure computing. Together, we can build a future where privacy and security are foundational to all digital workloads.

Read the full report here.

The Confidential Computing Consortium proudly announces the first community release of ManaTEE, an open-source framework for private data analytics. Originally developed by TikTok as a privacy solution for secure data collaboration, ManaTEE is now open-sourced and part of the Linux Foundation’s Confidential Computing Consortium.

Highlights of the Community Release:

• Easy Deployment: Test ManaTEE locally with minikube, no cloud account needed.
• Comprehensive Demo Tutorial: Step-by-step guidance to get started.
• Extensible Framework: Refactored code with Bazel builds and a CI/CD pipeline, ready for contributions.

What’s Next?

ManaTEE is evolving with plans to:

• Expand backend support to multi-cloud and on-prem solutions.
• Integrate privacy-compliant data pipelines.
• Enhance output privacy protections.
• Support confidential GPUs for AI workloads.

A Technical Steering Committee will soon guide the project’s future.

Learn More

Explore the potential of ManaTEE and join the community effort. Read the full announcement: First Community Release of ManaTEE.

January 6, 2025

Author: Sal Kimmich

The rapid adoption of AI and data-driven technologies has revolutionized industries, but it has also exposed a critical tension: the need to balance robust security with explainability. Traditionally, these two priorities have been at odds. High-security systems often operate in opaque “black box” environments, while efforts to make AI systems transparent can expose vulnerabilities. 

Verified Computing bridges this gap that reconciles these conflicting needs. It enables organizations to achieve unparalleled data security while maintaining the transparency and accountability required for compliance and trust.

The Core Technologies That Make It Possible

1. Trusted Execution Environments (TEEs)

TEEs are hardware-based secure enclaves that isolate sensitive computations from the rest of the system. They protect data and processes even if the operating system or hypervisor is compromised. Examples include Intel® SGX, Intel® TDX and AMD SEV.

• How They Work: TEEs operate as secure zones within a processor, where data and code are encrypted and inaccessible to external actors. For example, during a financial transaction, a TEE ensures that sensitive computations like risk assessments are performed without exposure to the broader system.
• Why They Matter: They protect data “in use,” closing a crucial gap in the data lifecycle that encryption alone cannot address.

2. Remote Attestation

Remote attestation provides cryptographic proof that a TEE is genuine and operating as expected. This ensures trust in the environment, particularly in cloud or collaborative settings.

• How It Works: A TEE generates an attestation report, including a cryptographic signature tied to the hardware. This report confirms the integrity of the software and hardware running within the enclave (source).
• Why It Matters: Remote attestation reassures stakeholders that computations occur in a secure and uncompromised environment, a critical requirement in multi-tenant cloud infrastructures.

3. Confidential Virtual Machines (VMs)

Confidential VMs extend TEE principles to virtualized environments, making secure computing scalable for complex workloads. Technologies like Intel® TDX allow organizations to isolate entire virtual machines.

• How They Work: Confidential VMs use memory encryption to ensure that data remains secure during processing. Encryption keys are hardware-managed, inaccessible to the hypervisor or OS (source).
• Why They Matter: They enable secure data processing in public clouds, even in shared infrastructures.

4. Verified Compute Frameworks

Verified Compute frameworks build on TEEs by introducing mechanisms for generating immutable logs and cryptographic proofs of computations. An example is EQTY Lab’s Verifiable Compute.

• How They Work: These frameworks capture the details of computations (inputs, outputs, and environment integrity) in tamper-proof logs. These logs are cryptographically verifiable, ensuring transparency without compromising confidentiality.
• Why They Matter: They allow organizations to meet regulatory requirements and provide explainable AI outputs while safeguarding proprietary algorithms and sensitive data.

5. Homomorphic Encryption and Secure Multi-Party Computation (SMPC)

In cases where external collaboration or ultra-sensitive data handling is needed, additional cryptographic techniques enhance confidentiality.

• Homomorphic Encryption: Enables computations on encrypted data without decryption.
• SMPC: Distributes computations across multiple parties, ensuring that no single party has access to the complete dataset.
• Why They Matter: These techniques complement TEEs by enabling secure collaboration across untrusted parties.

How Verified Computing Bridges Security and Explainability

Achieving Transparency Without Sacrificing Security

Traditionally, efforts to make AI systems explainable required exposing internal processes or sharing sensitive data—practices that risked data breaches or model theft. Verified confidential computing changes the game by:

• Allowing computations to occur in TEEs or confidential VMs, ensuring data is secure at all times.
• Using verified compute frameworks to provide cryptographic evidence of computation integrity, allowing external parties to trust results without accessing sensitive details.

For example, a healthcare provider running an AI diagnostic tool can securely process patient data in a TEE. The AI’s decisions can be explained to regulators or patients using cryptographic proofs, without exposing proprietary algorithms or patient information.

Supporting Regulatory Compliance

Regulations like the General Data Protection Regulation (GDPR) and the Health Insurance Portability and Accountability Act (HIPAA) demand both robust security and transparent handling of sensitive data. Verified confidential computing offers a solution by generating immutable logs and proofs that demonstrate compliance. This reduces audit complexity and ensures adherence to privacy laws.

Building Trust in AI Systems

As AI plays a growing role in critical sectors, trust is paramount. Verified computing ensures that stakeholders can verify:

1. The Security of the Data: Through TEEs and confidential VMs.
2. The Integrity of Computations: Via cryptographic attestation and verifiable compute frameworks.
3. The Explainability of Results: Through transparent logging and auditable records.

For instance, financial institutions can use verified computing to process loan applications, providing regulators with evidence of fairness and transparency without compromising customer data security.

Verified Compute

Verified Computing is more than a technological advancement—it is a paradigm shift. By integrating technologies like TEEs, remote attestation, confidential VMs, and verifiable compute frameworks, it resolves the long-standing tension between security and explainability. Organizations can now protect sensitive data, ensure compliance, and provide transparent, trustworthy AI systems.

As industries adopt this approach, verified computing will become the gold standard for secure and accountable digital transformation. Bridging these historically conflicting priorities paves the way for a future where trust in AI is not just an aspiration, but a guarantee. For more insights and resources, visit the Confidential Computing Consortium.