- Breakthrough provides the first-ever certificates of authenticity and compliance for independent verification of AI training, inference, and benchmarks at runtime.
- The solution delivers on-silicon, real-time governance — an industry-first created from two years of intensive research joined by advisors at Stanford and MIT.
- Delivered this month to first clients with transformative applications across the life sciences, public sector, finance, and media to certify smarter, safer AI systems and agents.
LOS ANGELES — (BUSINESS WIRE) — December 18, 2024 — EQTY Lab, in collaboration with Intel (NASDAQ: INTC) and NVIDIA (NASDAQ: NVDA), announced today the release of the Verifiable Compute AI framework, the first hardware-based solution to govern and audit AI workflows. Verifiable Compute represents a significant leap forward in ensuring that AI is explainable, accountable, and secure at runtime. It gives consumers and businesses new confidence to accelerate AI adoption and development.
“As a new era of autonomous AI agents emerges, we must evolve our trust in AI systems,” said Jonathan Dotan, Founder of EQTY Lab. “Verifiable Compute protects and controls AI data, models, and agents with the industry’s most advanced cryptography. It transforms how organizations enforce AI governance, automate auditing, and collaborate to build safer and more valuable AI.”
Verifiable Compute introduces a patent-pending hardware-based cryptographic AI notary and certificate system to isolate sensitive AI operations and notarize them with a tamperproof record of every data object and code computed in AI training and inference. It also provides real-time compliance checks and enforcement of AI business policies and new sovereign AI regulations such as the EU AI Act. Verifiable Compute’s new layer of trust is rooted right in the silicon of next-generation hardware from NVIDIA and Intel, setting the pace for a new standard for AI safety and innovation. A copy of the Verifiable Compute whitepaper is available for download at eqtylab.io/verifiablecompute.
“Intel is pushing the boundaries on delivering Confidential AI from edge to cloud, and EQTY Lab provides another level of trust to the confidential computing ecosystem,” said Anand Pashupathy, VP & General Manager, Security Software & Services Division, Intel Corporation. “Adding Verifiable Compute to Confidential AI deployments helps companies enhance the security, privacy, and accountability of their AI solutions.”
“The true potential of AI won’t be fully realized until we can provide confidential computing to verify every component in the stack,” said Michael O'Connor, NVIDIA Chief Architect for Confidential Computing, NVIDIA. “Securing the trust boundary in the processor sets a standard for next-generation AI workloads to be cryptographically secure and verifiable.”
The Verifiable Compute framework and notary system unlocks a powerful new capability in Trusted Execution Environments (TEEs) available on the 5th Gen Intel® Xeon® Processors with Intel® Trust Domain Extensions (Intel® TDX), extending the trust zone through confidential VMs to the NVIDIA H100/H200 GPUs and NVIDIA’s forthcoming Blackwell GPU architecture. The demand for confidential computing has surged this year owing to requirements for compliance with data sovereignty laws and new AI regulations. The market is projected to reach global sales of $184.5 billion by 2032.
Verifiable Compute addresses the unique and escalating risks to AI supply chains, from AI poisoning and information extraction to privacy backdoors and denial-of-service attacks. According to recent studies, 91% of organizations have experienced supply chain attacks on traditional software systems — an issue that becomes even more pronounced in the context of AI agents that automate tasks with less supervision.
By providing a cryptographically secure record of every stage of the AI lifecycle, Verifiable Compute demonstrates how innovation can thwart attacks with provable authentication, security, and assurance rooted in silicon. Verifiable Compute also allows for provable records of conformity with regulatory frameworks that can preserve AI artifacts years after a model has delivered results. If mandatory controls are not satisfied, a verifiable governance gate halts an AI system and can notify or integrate into an enterprise’s remediation tooling, with native connectors to ServiceNow, Databricks, and Palantir. If the system is compliant, it can issue an AI audit and lineage certificate that is verified instantly in a browser or can be independently audited at any point in the future. Together, these advanced capabilities eliminate a major trust gap for enterprises, allowing them to innovate responsibly with AI and prepare to meet the new promise of autonomous AI agent systems.
About EQTY Lab
EQTY Lab pioneers solutions to accelerate innovation and trust in AI. Their flagship product, the AI Integrity Suite, applies cryptographic technology to ensure that the governance of AI data, models, and agents is accountable to all stakeholders. With applications spanning the public sector, life sciences, and media, EQTY Lab is at the forefront of enabling trusted and responsible AI. To learn more about Verifiable Compute go to eqtylab.io/verifiablecompute and eqtylab.io.
View source version on businesswire.com: https://www.businesswire.com/news/home/20241218897420/en/
Contact:
EQTY Lab — Zenobia Godschalk
zenobia@zagcommunications.com
Intel — Jennifer Foss
jennifer.foss@intel.com
NVIDIA — Laura Martinez
lauram@nvidia.com
