MeetKai signs Albania MOU as James Kaplan pushes sovereign AI into government workflows
The July 6th deal is a planning agreement with AKSHI, with no disclosed value, model architecture, or timetable.
By Ryan Merket ยท Published
Why it matters
MeetKai is trying to convert sovereign AI from policy language into government deployment. Albania gives James Kaplan a European test case, but the MOU leaves the commercial terms, technical architecture, and governance model unresolved.
MeetKai co-founder and CEO James Kaplan signed an MOU with Albania's National Agency for Information Society on July 6 to advance plans for a joint venture that would build Albanian-language reasoning models, sovereign infrastructure, and AI applications across government.
The key word is plans. The July 6 PR Newswire release says MeetKai and AKSHI signed a memorandum of understanding to move toward a joint venture focused on a national sovereign AI system for Albania. It does not say the joint venture has been legally formed. It does not disclose a contract value, ownership split, procurement structure, deployment schedule, compute provider, base model, security standard, or data governance terms.
That gap matters because Kaplan is selling governments a version of AI that is less about chatbots and more about control. MeetKai's pitch is that countries should run AI on national terms: in local languages, on local or sovereign infrastructure, with institutions able to operate and expand the systems themselves. In Albania, the proposed workstreams cover sovereign deployment architecture, Albanian-language model development, public sector integrations, and knowledge transfer to local teams.
Kaplan founded MeetKai in Los Angeles in 2018 with Weili Dai, who is listed by MeetKai as co-founder and executive chairwoman. MeetKai's current about page names Kaplan as co-founder, CEO and CTO, and Dai as co-founder and executive chairwoman; a 2024 MeetKai release said the pair had built MeetKai across AI-enabled digital twins and interactive products for clients including BYD, Silicon Box, Sony Pictures, the National Basketball Players Association, and the Charlotte Hornets. Albania shows the second act Kaplan is trying to write: moving MeetKai from immersive and conversational AI work into national AI infrastructure.
What MeetKai says it will build
MeetKai and AKSHI say the proposed joint venture will focus on sovereign infrastructure, large language and reasoning models built for Albania, AI integrations across ministries, and new AI systems for public services and priority sectors of the economy. Early areas are expected to include digital government services, citizen interaction, education, and healthcare.
The release also says the venture will fund local research and talent development. That is the part of the announcement that turns this from a software integration into a capacity-building claim. MeetKai says the goal is an AI base that Albania can operate and expand on its own, rather than a set of foreign-hosted tools bolted onto ministry workflows.
Kaplan framed the project around Albania's existing digital government push. In the release, he said Albania has made AI a serious part of its digital governance agenda and that the next step is making systems "secure, locally grounded, and truly native to Albanians." AKSHI general director Igli Tafa said the goal is a framework Albania can deploy across government in a coordinated and trusted way.
MeetKai's own platform language matches that government-first pitch. MeetKai describes its sovereign AI product as a multilingual, multimodal platform built to run on a nation's infrastructure, trained on national data, and fine-tuned in local languages. The Albania MOU is the kind of reference account Kaplan needs if MeetKai is going to be judged by deployed government systems rather than by AI branding.
Albania brings both proof and risk
Albania is a useful target for MeetKai because AKSHI already runs a centralized digital-government program through e-Albania. The release says AKSHI leads Albania's digital government services, IT systems, and national digital infrastructure, including e-Albania and government technology modernization.
Albania has also put AI into public services via Diella, which the release cites as part of the existing base. Separate AP coverage in 2025 and 2026 sparked debate over consent and accountability around Diella's deployment, including a legal challenge by an Albanian actor over the use of her image and voice.
For MeetKai, the Diella episode is relevant because sovereign AI is ultimately a public-trust business. Local language models and in-country deployment answer one class of concern: where data sits, who controls infrastructure, and whether foreign providers can shape core systems. They do not, by themselves, answer questions about consent, accountability, procurement authority, model auditing, or how citizens challenge automated decisions.
The Albania MOU includes governance language, but the release leaves the hard mechanics unstated. There is no public detail on which datasets would train or tune Albanian-language reasoning models, how personal data would be excluded or controlled, which ministries would be first to deploy, or how model behavior would be audited once services go live.
Kaplan is chasing a real government AI budget line
MeetKai's Albania announcement fits a wider repositioning. MeetKai lists Albania among a run of 2026 government-facing announcements. In late June, the company said its Brazilian affiliate had won SERPRO's sovereign AI tender for Brazil (MeetKai Brasil named the final winner of SERPRO's sovereign AI contract). MeetKai has also announced sovereign AI work tied to Pakistan and multi-country pilots through regional organizations.
That cadence is the story behind the MOU. MeetKai is trying to become an implementation partner for governments that want national control over AI but do not have the internal stack to build, tune, deploy, and govern local-language systems alone. Albania gives Kaplan a European public-sector case study if the MOU turns into a functioning joint venture.
The market is moving in his direction. McKinsey estimates sovereign AI could represent a $600 billion market by 2030, driven by public-sector and regulated-industry workloads. The European Commission's Cloud and AI Development Act, published in June 2026, frames cloud and AI capacity as a sovereignty and competitiveness issue, with proposals for an EU-wide sovereignty framework, public-sector procurement tools, and a target to at least triple EU data center capacity over five to seven years.
MeetKai is not competing on the same basis as frontier-model labs with multi-billion-dollar training budgets. Its Albania pitch is narrower: local-language reasoning, deployment architecture, ministry integrations, and institutional knowledge transfer. That can be a defensible wedge if governments prioritize working systems inside their own legal and operational boundaries over access to the largest general-purpose model.
The next evidence will be contractual, technical, and operational. A signed MOU tells the market that AKSHI and MeetKai want to work together. It does not prove that Albania has procured the system, that MeetKai has secured the required compute, or that ministries will adopt the resulting tools. Kaplan has put MeetKai in the room with a government already willing to test AI in public administration. Turning that into durable infrastructure is the harder part.