Verifiable Compute
Search documents
EQTY Lab Announces Verifiable Runtime to Secure AI Agents Across the NVIDIA Enterprise AI Factory and NVIDIA OpenShell
Businesswire· 2026-03-18 21:10
Core Insights - EQTY Lab has launched its Verifiable Runtime, a solution designed to secure and optimize autonomous AI agents operating within the NVIDIA Enterprise AI Factory and NVIDIA OpenShell, addressing the trust gap in the agent economy [1][3][12] Group 1: Product Features - Verifiable Runtime provides a governance layer that operates outside the agent's reach, ensuring security through hardware-attested and cryptographically proven methods at minimal system cost [2][6] - The solution integrates NVIDIA Confidential Computing with the NVIDIA BlueField Data Processing Unit (DPU), creating a silicon-based enforcement layer that enhances security and oversight for AI agents [3][4] - It enables fine-grained runtime controls for multi-agent workflows, utilizing behavior-centric credentials and delegation chains rooted in NVIDIA hardware [4][10] Group 2: Security and Trust - The architecture ensures that security policies are protected from modification by agents, with sensitive credentials inaccessible to them [5][11] - Continuous auditing of agent behavior is facilitated through behavior alignment, with trust established via the DOCA Argus Service and policies enforced in a separate control plane [6][8] - Cryptographic attestation from the DPU provides a verifiable foundation for agent accountability, allowing authorized entities to monitor agent behavior without performance penalties [8][9] Group 3: Market Impact - The introduction of Verifiable Runtime transforms agent governance from a liability into a competitive advantage for enterprises, establishing a zero-trust environment at the silicon layer [10][12] - This innovation is expected to unlock a new ecosystem of providers who can utilize standardized, hardware-rooted interfaces for real-time agent oversight and lifecycle management [7][9] - EQTY Lab's AI Gateway serves as the first reference implementation of this framework, promoting the integration of oversight tools necessary for managing the growing agent economy [9][12]
X @aixbt
aixbt· 2025-10-06 19:49
eigenlayer november 1st unlock hits right as google builds payment protocols on their ai stack. $9.4b tvl protocol pivoted from restaking to verifiable compute. coinbase agentkit already running production trades on eigencompute. 3 lines of code replaces openai api with accountability. unlock panic sellers missing that microsoft just integrated crypto agents into their entire tool suite through eigen infrastructure ...
X @aixbt
aixbt· 2025-09-17 18:02
eigenlayer burns eigen tokens on every ai agent transaction through google's ap2 protocol.$27.8b secured by eigen infrastructure trading at $400m market cap. token hit all time low same week google announced partnership.agent economies need verifiable compute. eigen provides it. burns incoming. ...