Vast Data Introduced Confidential Computing Service

The company launched DataEnclave to allow AI model training on sensitive data using secure hardware-isolated environments.

Updated on Sept. 22, 2026 in Artificial Intelligence

Bold flat-color editorial illustration in navy, cream, and orange, depicting a stylized steel server chassis as a secure geometric metaphor. — Vast Data has launched DataEnclave, a secure, hardware-isolated computing service designed for training artificial intelligence models on highly sensitive or regulated information. AI Illustration. Upload story photo >

Live Poll

Do security concerns currently stop you from using advanced artificial intelligence tools at your work?

Vast Data has introduced DataEnclave, a service designed to enable the use of artificial intelligence models on highly sensitive or regulated data. By utilizing Nvidia confidential computing technology, the platform creates isolated environments to protect data and proprietary model weights.

Why it matters

The service addresses critical concerns for organizations that are unable to move regulated information to external clouds. It also allows developers to utilize AI infrastructure without the risk of exposing proprietary model weights.

DataEnclave provides hardware-isolated environments that encrypt guest memory, GPU memory, and NVLink traffic. The system uses cryptographic attestation to verify hardware, software, and governing policies before any data is decrypted.

The players

Vast Data

This technology company specializes in data infrastructure and artificial intelligence storage solutions.

Nvidia

This multinational technology corporation is a leader in graphics processing units and confidential computing hardware.

Cisco Systems

This global networking and hardware manufacturer provides enterprise-grade infrastructure components.

Super Micro Computer

This company designs and manufactures high-performance server and storage systems for large-scale data centers.

The details

The platform integrates directly into the existing Vast Data AI Operating System and logs all enclave activity within the company's proprietary database warehouse. Partners contributing to the service ecosystem include Cohere, CrowdStrike, Deepgram, The San Francisco AI Factory, Fundamental Research Labs, TwelveLabs, Cisco Systems, and Super Micro Computer.

Vast Data introduced the DataEnclave service on September 22, 2026.
The service is scheduled to ship to users in the first quarter of 2027.

The Tech Race

This development reflects a broader industry shift toward securing the model training pipeline as organizations prioritize data sovereignty. It signals a move away from legacy cloud models toward isolated, verifiable compute environments that offer better protection for intellectual property.

Developers and organizations using the Vast Data AI Operating System will gain access to these security features at no additional charge when the service launches. This allows teams to process sensitive workloads locally or in secure environments without requiring changes to their existing hardware infrastructure.

The takeaway

Confidential computing is becoming an essential component for enterprises that need to balance innovation with data privacy and regulatory compliance. Users should prepare their data pipelines now to take advantage of these integrated hardware-level security protections upon the 2027 release.