Data sovereignty is becoming an architectural problem as enterprises push artificial intelligence inference toward their most sensitive information. Pipelines assembled from separate tools struggle to keep embeddings, permissions and source data aligned, leaving many valuable data sets out of AI’s reach.
The shift from model training to inference is changing what AI infrastructure has to deliver, and it has fueled demand for Vast Data Inc., which raised about $1 billion at a $30 billion valuation in April 2026. Platforms built for scale and speed now must also handle user identification, access control and regulated data, according to Vaughn Stewart (pictured), vice president of systems engineering at Vast.
“There’s kind of different philosophical concepts in the market, right? One is the bespoke model … a best-in-breed of every piece in an agentic pipeline,” Stewart said. “And what we see there is just too much exposure to having data that’s maybe not in sync between being updated versus the vector embeddings, or maybe not in sync between the user access controls on the source data versus what’s provided through an inference query.”
Stewart spoke with theCUBE Research’s John Furrier at the AI Data Pipeline Forum, during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed data sovereignty, integrated data platforms for AI factories and how confidential computing can open regulated data to frontier models. (* Disclosure below.)
Why data sovereignty strains the best-of-breed pipeline
AI builders such as neocloud operators have little patience for infrastructure that underperforms. They will wipe and replace software that fails to deliver, a sharp break from the storage industry’s era of lock-in, Stewart noted. Vast’s disaggregated, shared-everything architecture combines the roles of storage systems and data-processing tools, he explained.
“No one ever referred to a Hadoop data lake, or a Kafka event bus as a storage platform, but it stored petabytes or tens of petabytes of data. So, we’re able to conflate the functions of these two systems in a tightly integrated way that delivers performance,” Stewart said.
Asked about Vast’s competition, Stewart pointed not to a rival company but to “a legacy mindset” — assembling an agentic pipeline from swappable best-of-breed components. The trade-off is that data must move from system to system, and access controls get lost along the way, he noted.
“I think this construct of, ‘I can build and move this massive volume of data and all these connections through a pipeline and have it be near-real-time,’ I think, is a fallacy when we start to talk about the needs for serving parallel requests, parallel functions at mass scale,” Stewart said.
About half of the world’s data still sits in corporate data centers or colocation facilities because it is sensitive and in regulated industries, according to Stewart. Much of it legally can’t reach a public cloud frontier model, which makes data sovereignty a hard constraint. Vast’s response is DataEnclave, a confidential AI runtime announced in September 2026 with partners including Nvidia Corp. and Cohere Inc. It runs models in secured virtual machines, requires remote cryptographic attestation before keys are released and logs all access to a tamper-proof database, he explained.
“If there’s no attestation, there’s no access to the keys for the model and/or the data,” Stewart said. “This is what we believe is going to allow us to bring frontier models to these highly secure datasets today and kind of bridge that last mile of opening up AI tools, technologies and capabilities.”
Here’s the complete video interview, part of SiliconANGLE’s and theCUBE’s coverage of the AI Data Pipeline Forum:
(* Disclosure: ScaleFlux sponsored this segment of theCUBE. Neither ScaleFlux nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)