Market and industryNovember 22, 2026Редакция 105.kz
Data as an Asset and the Digital Code: What Statements About Kazakhstan's Digital Economy Mean for Plants
MP Yekaterina Smyshlyaeva spoke about the Digital Code, two commissioned supercomputers and data as an economic resource. We examine which parts of this are working context for Kazakhstan's industrial enterprises and which are still only a stated prospect.
Key takeaway. The source records the general context: data is becoming an independent economic resource, two supercomputers have been commissioned in Kazakhstan, and a Digital Code on relations in the digital environment and protection of personal data is being prepared. There are no direct requirements for industry in it. For a plant this is a framework in which it is reasonable to start with a data inventory and a narrow pilot inside its own perimeter.
What happened
In an interview for a podcast published by Kazinform on 21 November 2026, MP Yekaterina Smyshlyaeva set out several theses about Kazakhstan's digital economy. According to her, artificial intelligence technologies influence labour productivity, help reduce costs and automate work processes. Data, in her assessment, is becoming an independent economic resource, and in the longer term an increasingly large role will be played by the "agent economy" — when some tasks and interactions are performed by digital agents. She calls the development of a "people's digital economy" the logical continuation of the process of forming a digital state: it is important not only to build infrastructure and convert services into electronic form, but also to ensure the participation of citizens in new economic processes.
The second part of the conversation is infrastructure and rules. According to the MP, the digital economy needs not only internet, but also computing capacity, data centres, software solutions and a stable power supply. Two supercomputers have already been commissioned in the country, one more is expected before the end of the year, and the development of data centres and localisation projects of large technology companies are planned. Separately, the Digital Code is announced: it "is aimed at regulating relations in the digital environment" and covers identity confirmation, the authenticity of digital documents, electronic document flow, protection of citizens' rights and confidentiality of information, but, according to Smyshlyaeva, it does not regulate individual technologies and infrastructure processes. She calls personal data a "personal property asset" that cannot be regarded as an ordinary commodity, and points to the problem of excessive collection of information by various organisations. There were also workforce theses: the number of grants for the specialty has been increased to 3,000, but the question of training quality remains open. The theses are linked to the President's Address of 28 October on the transition to a "people's digital economy".
What this means for enterprises in Kazakhstan
The source is an overview and contains no industry-specific requirements. Below is our reading of the statements as applied to production (estimate).
| Thesis in the source | What it means for an enterprise (estimate) |
|---|---|
| A stable power supply is named a condition of the digital economy | New computing loads are an additional argument for keeping observability of main equipment and a capacity reserve |
| Supercomputers and data centres | Large energy consumers; for power utilities — load growth and unevenness, for plants — the question of their own capacity for local data processing |
| The Digital Code and the authenticity of digital documents | Directly concerns work permits, acts, maintenance logs and electronic document flow — where today a document lives on paper and in three accounting systems |
| The special status of personal data | At an enterprise these are employee data in HSE systems, access control and video; it is reasonable to keep the processing of such arrays inside the enterprise perimeter |
| Personnel shortage and quality of training | An argument for the pilot to rely on existing engineers rather than on an external team "on a fly-in basis" |
Energy. The source names power supply as an infrastructure condition of the digital economy. For generation and grids this translates into specific assets: power transformers, feeders, pump and turbine units. The data for monitoring them already exists — SCADA, automated metering systems (ASKUE), vibration, thermal imaging rounds. The maintenance task is to keep the technical availability factor under growing load rather than to deal with the consequences of failures.
Plants. The thesis on the authenticity of digital documents and electronic document flow is closest to processing stages where AI starts not with equipment but with a document: regulations, equipment passports, work orders, shift reports. Document flow at such enterprises is bilingual — Russian and Kazakh — and that is a separate requirement for text processing quality.
Data and the perimeter. The special status of personal data is a signal in favour of on-premise: telemetry, vibration and video stay inside the perimeter, and only results go outside. For national companies this is already the norm, for others it is a matter of information security requirements. Separately, the inherited process control system (APCS) fleet should be mentioned: integration is possible through standard gateways and historical archives, not through replacing the system. Internal standards (ISO, GOST) set the reporting format to which the model's output must be brought.
How to apply: an 8–12 week pilot
- Data and perimeter audit (1–2 weeks). Compile a list of systems: process control systems and SCADA, automated metering, maintenance, electronic document flow, HSE video. Mark where there is personal data, where there are commercial secrets, where an NDA applies. Without this map any pilot turns into a dispute about access.
- Choose one process with a measurable result. The source speaks of routine operations as the first area of change — at an enterprise this is, for example, parsing documentation, preparing shift reports, finding an answer in a regulation (estimate). There must be one process.
- Fix the processing perimeter. On-premise as the base option for enterprise data; for corporate LLMs in a closed perimeter the approach and acceptance requirements are described on the Corporate LLMs page.
- Run an 8–12 week pilot with pre-agreed metrics. For energy, the link to equipment and operating modes is on the Energy page. Metrics are agreed before the start: time per operation, error rate, volume of documents processed. Human verification remains mandatory.
- Acceptance and the decision on rollout. A before/after comparison on the same metrics, and training staff to work with the tool — a direct reference to the thesis on cognitive erosion in the source. Scaling — only after a measured result.
Limitations and risks
- The source is not industry-specific. It contains not a single indicator for industry, no description of AI deployments at enterprises, and no requirements of the Digital Code for production systems. Everything applied in this analysis is our interpretation, not a statement of the source.
- The timing and content of the code are not disclosed. The document's orientation is announced, but not its articles or its date of adoption.
- Infrastructure parameters are not named. For supercomputers and data centres there is neither capacity nor access conditions for industrial customers — computation plans cannot be built on this.
- Statements about AI's effect on productivity and costs are the MP's estimate, not measured data. No guaranteed economic effect follows from them.
- The "agent economy" is designated as a prospect without timelines or mechanisms; for an enterprise this is a horizon to watch, not a project.
- The workforce thesis concerns the number of grants (3,000), not quality; Smyshlyaeva herself raises the question of training quality.
- A pilot needs its own data: equipment passports, failure and operating-hours history, SCADA archives over a sufficient period (estimate). If they are not available in a usable form, data quality work is needed first, and that pushes the result beyond 8–12 weeks.
Sources
- Цифровая экономика: что даст Казахстану развитие ИИ и платформ - Казинформ · 2026-11-21T15:27:00.000Z
Как сообщила депутат, в Казахстане уже введены в эксплуатацию два суперкомпьютера, еще один ожидается до конца года.
- Фото: Diliff — US Capitol dome Jan 2006.jpg (CC BY 2.5, Wikimedia Commons)
This material was prepared with AI from the sources listed above. Spotted an inaccuracy? Write to info@105.kz.