Kazakhstan has 10 industrial robots per 10,000 manufacturing workers, according to the International Federation of Robotics. To address an estimated labour shortage of 40,000 to 50,000 workers while maintaining production growth, that figure may need to rise to around 70 robots. Even that would leave Kazakhstan far behind the world’s most automated manufacturing economies. South Korea and Singapore have more than 1,000 industrial robots per 10,000 workers. Germany, Japan, the United States and China have between 670 and 850.
Kazakhstan therefore faces more than a modest technology gap. Its manufacturing sector would need to increase the use of industrial robots several times over simply to approach a level consistent with its labour needs. The usual explanation is financial. Industrial equipment is expensive, long-term loans are limited, leasing costs are high and many manufacturers lack the capital needed for modernisation.
Kazakhstan’s response has been to subsidise machinery purchases through state business support programmes. Under one mechanism, the lending rate is calculated using the National Bank of Kazakhstan’s base rate plus five percentage points. The business pays 6% annually, while the state budget covers the difference.
Subsidies are available for five years for investment projects and three years for working capital, without extensions. A separate programme provides leasing through a subsidiary of the Development Bank of Kazakhstan. In small towns and rural areas, the value of a financial lease can reach KZT 1.5 billion per entrepreneur.
The logic is straightforward: if businesses cannot afford modern machinery, the state should reduce the cost of financing it. But money may not be the only constraint. Kazakhstan may be subsidising equipment without developing the engineering services needed to integrate that equipment into production.
What the Hungarian study reveals about robot adoption
The CEPR study examines Hungary, an open economy that relies on imported technology and does not have a major domestic robot manufacturing industry. In that respect, its industrial structure offers a useful comparison for Kazakhstan. The researchers gained access to administrative data covering the entire Hungarian economy between 2015 and 2021. The dataset included tax and customs records, business-to-business invoices and company financial statements collected by the national statistics agency.
This allowed them to track not only companies importing robots directly, but also domestic intermediaries that selected, configured, installed and commissioned automated equipment.
The study calls these specialised engineering companies integrators. Their role proved far more important than customs data alone would suggest. Among Hungarian companies that adopted industrial robots, only one in 20 imported the equipment directly. The other 19 out of 20 worked through local integrators.
Foreign trade statistics therefore captured only a small share of the companies actually introducing robots into production. That distinction matters for Kazakhstan. If equipment imported by an engineering intermediary is recorded as a single transaction, statistics may not reveal how many factories ultimately received, installed and used that equipment.
Why integrators matter more than machinery purchases
Buying an industrial robot is not the same as introducing automation. A manufacturer that imports a robot directly needs engineers who can install it, adapt it to the production process, write or configure the necessary software, connect it to other equipment and maintain it after commissioning.
Building such a team creates a high fixed cost. Large manufacturers may be able to absorb it. Most small and medium-sized companies cannot. An integrator removes part of that cost. The intermediary studies the production process, selects suitable equipment, installs it and makes sure it works within the factory’s existing systems.
The client pays a margin for those services, but avoids maintaining a permanent in-house engineering department. This creates a clear division in the market. Large and capital-intensive companies can buy automation equipment directly. Smaller manufacturers depend on specialised intermediaries because their production volumes do not justify their own integration teams.
Without an integrator, a robot can remain an expensive machine sitting in a warehouse. With an integrator, the same robot becomes part of a functioning production line.
Hungary’s integrator market was substantial. The researchers identified 352 engineering intermediaries. The number operating in the country increased from 234 in 2015 to 309 in 2021. Over the same period, the number of businesses purchasing equipment and services from them rose from 1,766 to 3,216. That was an increase of 82% in seven years.
The integrators were larger, more productive and more capital-intensive than the average company in their industries. These were not small repair businesses. They were technology companies earning revenue by converting capital expenditure into productive capacity. Manufacturing generated 65% of the total value of purchases made through integrators, although wholesale and service companies also used their services.
What Kazakhstan’s statistics may be missing
The Hungarian findings raise a measurement problem for Kazakhstan. If a large share of automation takes place through local intermediaries, customs data cannot show the full path from an imported machine to the factory where it is ultimately installed. Direct import statistics identify the company bringing the equipment into the country, but not necessarily its final industrial user.
This does not automatically mean that Kazakhstan’s official robot-density estimate is wrong. It means that statistics based heavily on direct imports may offer an incomplete picture of how automation spreads through the economy.
Kazakhstan’s 10 robots per 10,000 manufacturing workers could therefore describe only the visible part of a larger market. The size of the missing part would depend on how many local companies already install, configure and service automated equipment.
The problem is that Kazakhstan does not track these engineering intermediaries as a distinct part of the economy. The annual business innovation survey conducted by Kazakhstan’s state statistics agency found that 11.9% of companies were innovation-active in the latest reporting period cited in the article.
For companies that did not innovate, the survey grouped the reasons into three broad categories: weak demand, insufficient financing and high implementation costs. It did not ask whether a business had been unable to find a company capable of configuring, installing and commissioning the equipment.
That omission shapes policy. When the statistical system recognises only financing and demand constraints, government programmes naturally focus on reducing borrowing costs and subsidising purchases. A shortage of engineering capability remains invisible because businesses are not directly asked about it.
Automation infrastructure is unevenly distributed
The CEPR paper also shows that access to integrators has a geographic dimension. In Hungary, engineering intermediaries and their customers were concentrated around the capital more heavily than manufacturing activity as a whole.
The Herfindahl-Hirschman Index, used in the study to measure geographic concentration, was 0.19 for integrators, 0.13 for companies buying robots and 0.11 for manufacturing overall. The infrastructure supporting automation was therefore more centralised than the factories that needed it.
This has direct relevance for Kazakhstan, where industrial enterprises are spread across large distances. A factory in a regional industrial centre may qualify for subsidised machinery, but still struggle to find engineers who can install, programme and maintain it.
The financial support may reach the company while the technical capability does not. In practice, that would make industrial automation easier in a limited number of major cities and industrial hubs. Regional businesses would face higher travel costs, longer commissioning periods and greater dependence on engineers based elsewhere.
A policy designed to support regional manufacturing could therefore reinforce geographic inequality if it subsidises equipment without considering the location of integration services.
Robot purchases are rare, not routine
The Hungarian evidence also challenges the assumption that companies invest in automation regularly. Robot adoption was usually a large and episodic investment. Around two-thirds of companies purchased automation equipment unevenly and in bursts. More than half of all buyers, or 54%, made exactly one purchase during the seven-year observation period and did not return to the market. For most businesses, automation is not a routine annual expense. It is a rare decision that may shape production for several years.
That changes how the effectiveness of state support should be assessed. The central question is not only whether the subsidy helped a business buy equipment. It is whether the machinery was successfully installed, integrated and used to increase output. A poorly integrated machine can remain idle even when its financing terms are generous. A well-integrated machine can improve productivity even if the subsidy itself is relatively modest.
The quality of implementation may matter more than the difference between one subsidised interest rate and another.
Kazakhstan’s dependence on foreign technology suppliers
The Hungarian study found another layer of dependence. Every integrator in the sample imported at least some equipment. Around 80% obtained more than half of their supplies from a single country. For industrial robots, the main supplier countries were Germany, Austria, Sweden, Japan and the Netherlands.
When production expanded in one of these supplier countries, Hungarian integrators dependent on that market increased their own employment and sales. The researchers traced how changes in exports from individual countries affected the growth of domestic engineering intermediaries.
This means that national automation depends on more than domestic financing policy. It also depends on foreign equipment suppliers, logistics, technical standards, software access and the ability of local engineers to work with imported systems. For Kazakhstan, developing an integrator market would not eliminate external dependence. It would make that dependence more manageable by building local expertise around foreign technology.
The country may not manufacture industrial robots, but it can develop companies that understand how to deploy, adapt and maintain them.
How Kazakhstan could test the integrator hypothesis
Kazakhstan already has much of the information needed to find out whether integrators influence the results of state-funded modernisation projects. Loan subsidies, leasing programmes and financing provided through the state development bank generate contracts linked to specific companies, investment amounts and equipment purchases. These records could provide a documented list of businesses that received support for industrial modernisation.
Kazakhstan has also operated a mandatory electronic VAT invoicing system since 2019. It records transactions between VAT-paying businesses, including payments for equipment, installation, commissioning and technical services. In principle, these datasets could reconstruct the relationships between three groups:
- businesses receiving state support to purchase machinery;
- suppliers importing or selling the equipment;
- engineering companies installing and commissioning it.
The government could then compare those relationships with changes in output per worker after the equipment was introduced. The analysis should answer several practical questions. Did the company work with an integrator? Was the integrator located in the same region? Did the project include installation and technical support? Did labour productivity increase after commissioning? Did the business continue using the equipment?
This would separate projects that produced functioning industrial capacity from those that resulted only in the purchase of machinery. The approach would also allow Kazakhstan to identify regions and industries where integration services are missing. Support could then be directed towards engineering companies, technical training and regional service networks rather than machinery alone.
Industrial policy should finance capability, not just equipment
Kazakhstan does not necessarily need to abandon machinery subsidies. It needs to determine whether those subsidies are addressing the main barrier to automation. If the decisive difference between successful and unsuccessful projects is the availability of a qualified integrator, industrial policy has another lever to use.
The state could support the development of engineering intermediaries, include installation and commissioning in eligible project costs, require post-investment productivity assessments and create incentives for integrators to work outside the largest industrial centres. It could also adjust business surveys so that companies can report shortages of engineering, installation and technical support services directly.
The broader lesson is simple. Industrial modernisation does not happen when a company buys a machine. It happens when that machine begins producing.
Kazakhstan already helps businesses finance equipment. The next question is whether it has enough companies capable of making that equipment work.
Shyngys Yerbolat, EconomyKZ Research Group, exclusively for EconomyKZ.org
