Trust Stamp Inc (NASDAQ:IDAI, ISE:AIID) has delivered its first sovereign large language model to an international telecommunications company, a project it says marks a shift in strategic direction.
The Atlanta-based technology firm, known for its identity authentication tools, said the compliance-focused model runs entirely on the client's own servers.
Sovereign artificial intelligence refers to building and running systems with independence over data, technology, operations and legal structures.
Chief executive Gareth Genner said the model has no external access and no reliance on any third party for future operation.
Because it runs on-premises and is cut off from the internet, Trust Stamp said it removed client concerns about data poisoning, a technique where corrupted inputs are used to manipulate an AI model's behaviour.
The approach also addressed privacy worries about sharing multinational commercial and personal data with large cloud-based AI providers.
Genner said the only ongoing costs were modest incremental power charges, in contrast to the uncertain future pricing attached to frontier AI providers.
He argued those cost concerns were already prompting some commercial users to scale back usage or switch to smaller, cheaper models.
The deal follows Trust Stamp's acquisition of Lexverify in February, which the company said was intended to broaden its range of language model products alongside its identity work.
Genner described the telecoms project as an early milestone, framing sovereign deployments as offering energy-efficient models tailored to specific client needs.
He said such systems avoided the capital, energy and water demands associated with large-scale AI infrastructure.
Trust Stamp said it was focused on European Union and African markets, where it sees rising demand for data and technology sovereignty.
The company cited a forecast that annual spending on such deployments across those markets could grow to between $15 billion and $30 billion by 2030, describing the estimate as conservative.
Genner said he would host a shareholder call on 17 July to set out the scale and significance of the strategy.
Access details for that call will be published closer to the event.
