Vast Data Inc. today introduced DataEnclave, a confidential computing environment designed to let companies run advanced artificial intelligence models against sensitive data without exposing either the data or the model’s intellectual property.
DataEnclave is a new capability within the Vast DataEngine and broader Vast AI Operating System that uses Nvidia Corp.’s confidential computing technology to create hardware-isolated environments in which proprietary models and enterprise data can be decrypted and processed securely.
The technology targets financial institutions, healthcare providers, government agencies and other organizations that cannot move regulated or sensitive information to externally hosted AI services. It also addresses model developers’ reluctance to distribute proprietary model weights into infrastructure they do not control.
“Model weights are fast becoming the most valuable intellectual property in the world,” said Jeff Denworth (pictured), Vast Data’s co-founder. “Base weights define the value of frontier models, while fine-tuned weights will increasingly represent the proprietary intelligence of AI-driven enterprises.”
DataEnclave uses trusted execution environments spanning central processing units and graphics processing units. It encrypts guest memory, graphics processing unit memory and NVLink traffic while isolating active models and data from administrators, infrastructure operators and other tenants.
A cryptographic attestation process verifies the hardware, software and governing policies before models or data are decrypted. Enterprises and model developers retain control of their respective encryption keys through integrations with their own key management systems.
“Up until now, there hasn’t been a great solution for ensuring that the memory space that people do this computing in is equally protected,” Denworth said. “Once you get to the final computing destination, is that environment safe enough to actually decrypt the world’s most sensitive data?”
Flexible deployment
DataEnclave can be deployed in customer data centers, AI clouds and fully air-gapped environments. Its attestation services are based on the open-source Cloud Native Computing Foundation Trustee project or can be provided through Fortanix Inc.’s Confidential AI offering.
The system also logs attestation events, key releases and enclave lifecycle activity in the Vast DataBase data warehouse. These records provide an audit trail showing which workloads ran and under what policies without revealing the protected data or model weights.
Initial model and software partners include Cohere Inc., CrowdStrike Holdings Inc., Deepgram Inc., The San Francisco AI Factory Inc., Fundamental Research Labs Inc. and TwelveLabs Inc. Cisco Systems Inc. and Super Micro Computer Inc. are among the hardware partners that will offer integrated systems.
Vast said the model allows a developer’s base weights and a customer’s fine-tuned weights to operate together while remaining under separate ownership and key control. That could help enterprises create customized models without disclosing their proprietary training or operational data to the base-model provider.
“Both of those can interoperate within this multitenant enclave where we can support the trust requirements of both organizations simultaneously,” Denworth said.
Vast is initially concentrating on organizations with particularly strict security and data residency requirements. Denworth identified large banks, intelligence and defense agencies and pharmaceutical companies as early target customers.
“In all three cases, they have said that this solves major pains for them in that they’re not allowed to send certain data products up to different services that they want to use,” he said. “It’s basically the most regulated and air-gapped customers of the world.”
The company also argues that DataEnclave could change the economics of enterprise AI. Customers could run proprietary models on infrastructure they own or select a regional AI cloud, reducing model developers’ need to finance the underlying computing capacity.
For Vast, the capability will be included in the existing AI Operating System rather than sold as a separate product. “We don’t charge anything extra for this,” Denworth said. “It’s just part of the AI operating system stack.”
DataEnclave is available in preview and is scheduled to ship in the first quarter of 2027 through Vast and participating hardware partners, including Cisco and Supermicro.
