Wand AI adds VLNO robustness layer to sovereign AI offering
Wand AI said it has integrated VLNO’s model robustness technology into its Sovereign AI offering for governments and regulated enterprises using open-weight models.
Wand AI says the addition is meant to help governments and regulated enterprises run open-weight models on sovereign infrastructure with stronger defenses against adversarial attacks.
What happened
Wand AI said it has integrated VLNO’s model robustness technology into its Sovereign AI offering for governments and regulated enterprises using open-weight models on sovereign infrastructure.
In the company’s announcement, Wand said VLNO will act as a model robustness layer that continuously tests models against adversarial attacks, generates training data to harden them, and then re-tests them as models or threats change.
Wand said the joint reference architecture is available now for sovereign programs and regulated enterprises.
Why it matters
The companies framed the move as a response to a security problem that can slow the use of open-weight models in public-sector and regulated settings. Open-weight models are models whose weights are available for inspection or use, but the release said those systems often lack standardized safety evaluations, accountable vendors, and proven adversarial robustness.
Wand said its existing infrastructure already provides inference privacy and confidential computing for the surrounding data path. It said VLNO adds a layer that strengthens the model’s behavior itself.
Background and context
Wand said the offer is aimed at ministries, institutions, agencies, and other regulated organizations that want to keep sovereign control over models, compute, data, policies, and operations while still using open-weight systems.
The company said it was founded in 2023 and is headquartered in Palo Alto, California, with offices in Palo Alto, New York, and Abu Dhabi. It said it serves governments and enterprises and supports production-scale deployments with banks, asset managers, hedge funds, consulting firms, and system integrators.
VLNO said it builds a security training layer for open and agentic models. The company also said it offers a public robustness benchmark that teams can use for free.
What the companies said
Hertzel Kuriel, VLNO’s co-founder and CEO, said sovereign programs have had to choose between models they control and models they can defend. He said VLNO produces data that makes a model stronger.
Cristian Felix, Wand AI’s chief AI architect, said the company is pleased to welcome VLNO into Wand’s sovereign technology ecosystem. He said the addition lets ministries, institutions, and agencies adopt open-weight models on a unified national stack without lowering their security bar.
What happens next
The companies said the reference architecture is ready for use now. The release did not say whether any government or enterprise customers have already adopted the combined offering.
Sources
Editorial note: This article was prepared with AI assistance and includes source attribution for reader review.
Comments ()