Direction:

GovTech

From Digital State to Agentic State: What Changes Inside Ukraine’s Ministry of Digital Transformation

Publication date and time:

Feb 04, 2026, 18:24

Reading time:

5 min

In his first interview in a new position, Bornyakov outlines a strategic shift: Ukraine is no longer focused simply on digitalisation of services. The goal is to build what he calls an agentic state — a model where artificial intelligence becomes a core layer of governance, embedded into how public services operate, how decisions are made, and how the digital economy is regulated.

What already exists: Ukraine’s digital foundation

Ukraine already operates one of the most advanced digital public service ecosystems in the world.

The Diia platform serves as a unified gateway to government, covering:

• digital identity documents,
• business registration,
• social services,
• tax and administrative procedures.

This infrastructure is supported by a nationwide network of Chief Digital Transformation Officers (CDTOs) embedded in ministries, responsible for internal digital systems and reform implementation.

According to Bornyakov, this phase of “digital state” is largely complete. The next challenge is not automation, but intelligence: moving from digitised workflows to AI-driven systems.

Why the national LLM is delayed

Ukraine’s national large language model (LLM) was expected to launch earlier, but the project faced major structural constraints. The primary issue is legal, not technical.

As a government institution, the ministry cannot train models on copyrighted or non-consented data. Using scraped datasets — common in private-sector AI — would expose the state to legal risk and potentially invalidate any public services built on top of the model.

Private Ukrainian teams already operate Ukrainian-language models, but the state model must be: legally compliant, transparent in data provenance, and safe for use in public services.

To address this, the ministry is building a legal framework allowing training on publicly available and consented datasets. This includes books, archives, and public web sources.

Technically, the model will be based on Google’s open-source Gemma architecture. A custom Ukrainian tokenizer is nearly complete, and training will proceed once the legal dataset framework is finalised.

The national LLM is intended to support:

• AI assistants in Diia,
• education tools,
• citizen support services,
• future agentic public services.

From software to industry: what Diia.City 2.0 means

Diia.City 2.0 represents a major evolution of Ukraine’s digital economy model.

Originally designed as a special legal and tax regime for IT companies, Diia City is now being restructured into a sector-based industrial policy tool.

The focus shifts from software services to strategic technological industries, including:

• robotics and automation,
• microelectronics,
• bionic prosthetics,
• deep tech and applied AI,
• advanced manufacturing.

Each sector must demonstrate measurable economic impact: export potential, job creation, and contribution to GDP growth. Diia.City 2.0 is explicitly not designed to subsidise low-impact sectors.

Bornyakov describes this approach as “tailored regulation”: each industry receives a regulatory and fiscal framework adapted to its economic profile.

Chips and microelectronics: closing the last dependency

One of the most strategic topics in the interview is semiconductor production.

Ukraine has made major progress in localising defence manufacturing, particularly in drones. However, Bornyakov identifies microchips as the last critical dependency.

The government does not aim to compete with cutting-edge fabs at 1–5 nanometres. Instead, the focus is on:

• stable production of 20–120 nm chips,
• suited for drones, sensors, robotics, and defence systems.

The idea is to build domestic chip production facilities — potentially underground for security — financed by private investors under a state-created regulatory framework.

The state’s role is not to own factories, but to:

• aggregate demand from manufacturers,
• define technical specifications,
• create investment conditions via Diia.City 2.0.

Defence tech and industrial scaling

Parallel to Diia.City, Ukraine is developing Defence.City, a separate regime focused on large-scale industrial production in the defence sector.

Defence.City targets factories, warehouses, physical infrastructure, and export-oriented arms manufacturing.

Diia.City, in contrast, remains focused on intellectual and technological products.

Bornyakov emphasises that both models are complementary: one for industrial mass production, the other for innovation-driven sectors.

From digital services to AI governance

The overarching vision is that AI becomes state infrastructure — similar to electricity or telecommunications.

This includes:

• AI-based public service delivery,
• automated administrative processes,
• intelligent decision-support systems,
• and eventually autonomous agentic workflows inside government.

Bornyakov argues that AI should not remain an experimental layer. It must become the operational core of public administration.

Strategic implication

Ukraine is positioning itself as one of the first countries attempting to build a legally compliant national AI ecosystem integrated directly into governance, combined with industrial policy and defence innovation.

Rather than copying Western digital models, Ukraine is using wartime constraints to accelerate structural experimentation — turning digital infrastructure into economic and geopolitical capital.

The transition from digital state to agentic state marks a shift from efficiency to architecture: from digitising services to redesigning how the state itself operates.