A joint venture is not a market entry strategy so much as a decision to share an economic dependency permanently, in exchange for something the investor cannot obtain on its own. The useful question about Kazakhstan is therefore not whether joint ventures work there, but which dependencies still require shared ownership and whether that set of dependencies is growing or shrinking.

Between 2020 and 2026 the answer to that question moved, and the direction of travel is visible in the investment data. Gross foreign direct investment reached $20.5 billion in 2025, up 14.4%, with inflows into manufacturing rising 47.4% while investment into the extractive sector fell by 47%, and by January to July 2026 manufacturing had edged ahead of mining in the industrial output mix at 47.1% against 46.1% on Bureau of National Statistics figures. That crossover deserves less weight than it has been given, because it owes as much to falling oil and gas volumes as to the growth of processing, but the direction of capital is not in doubt. Money is moving toward things that have to be built, staffed and operated inside the country rather than extracted and shipped out of it.

A shift of that kind would normally produce more joint ventures, and the business register says otherwise. On Bureau of National Statistics figures for 1 December 2025, the number of registered legal entities and branches under joint Kazakh-and-foreign ownership grew by 0.6% over the year, while the number under wholly foreign ownership grew by 7.5%. The same divergence holds across every monthly release since late 2024, with foreign ownership compounding at five to eight per cent a year and joint ownership flat or falling.

Both of those things are true at the same time, in that Kazakhstan is attracting more industrial investment while foreign investors are increasingly making that investment alone. What has changed is not the general attractiveness of partnership but its specificity. The joint venture has stopped being the default answer to an unfamiliar market and has become the structure of choice in a much narrower set of situations, where a licence, a deposit, a feedstock stream, an approved production platform, a grid connection or a corridor position cannot be bought, rented or contracted for.

This article sets out what the 2020 to 2026 evidence shows, where joint ventures are actually forming, why the economics of local partnership have shifted, and what will determine the model's role through 2030. It draws on the research base we use when advising clients as a consulting company in Kazakhstan.

A note on what can and cannot be measured

Anyone researching this subject will find that the numbers do not reconcile, and the reason matters for how the rest of this article should be read.

The statistical business register counts legal form, not commercial substance. The Bureau classifies entities as being under state, private, foreign or joint ownership, which means a vehicle held 99% by a foreign parent and 1% by a local nominee is recorded as joint, while a genuine fifty-fifty industrial partnership held through a holding company in the AIFC or the Netherlands is recorded as foreign. The series is therefore a reliable indicator of direction rather than a count of joint ventures.

Foreign investment is reported on at least three bases, and the gap between them is wide enough to change the story. Gross inflow, the National Bank's directional measure, ran at $20.5 billion in 2025, net inflow is far smaller and considerably more volatile, and UNCTAD's stock measure put accumulated foreign investment at $151.3 billion, about 52% of GDP. These answer different questions, and none of them answers how much foreign capital sits inside shared-ownership vehicles.

The gross series, which is the one used throughout this article for scale, runs as follows.

Source: National Bank of Kazakhstan directional series, as compiled and reported in August 2026. These are gross inflows and should not be compared with net or balance-of-payments figures.

Announced projects are not committed projects. A memorandum, a framework agreement, an investment agreement and a final investment decision represent four distinct levels of commitment, and the gap between them can be large. A deep corn processing project in Zhambyl region was valued at $800 million in January 2026 and at $340 million in the government's August review, with no published explanation.

There is no official series counting joint ventures by commercial substance, and constructing one from the available data would produce a number rather than a measurement. This article therefore uses the registry for direction, foreign investment data for scale, and named transactions for structure. Where a figure is derived rather than reported, it says so.

Part I. What actually changed, 2020 to 2026

2020 to 2021: contraction, and a quiet change in the rules

Gross foreign direct investment fell from $24.4 billion in 2019 to $17.2 billion in 2020, then recovered to $23.8 billion in 2021, a movement that is unremarkable when read as a cycle of pandemic contraction followed by rebound.

The more consequential development of those two years did not show up in the flow numbers at all. In 2020 Kazakhstan introduced the utilisation fee on imported vehicles and self-propelled machinery and began building out the industrial assembly agreement framework, and Hyundai Trans Kazakhstan was commissioned in Almaty the same year. Neither event moved the headline investment total, but together they established the principle that has governed industrial policy ever since, which is that the state would not restrict imports but would instead make importing structurally less attractive than producing locally.

For a foreign company that changed the nature of the question, because until then serving Kazakhstan through a distributor had been a purely commercial choice. From 2020 it became a choice with a policy cost attached, and the size of that cost would be set by the government rather than by the market.

2022 to 2023: reorientation, and a boom that was not what it appeared

Gross inflows peaked at $28.2 billion in 2022, the highest figure in a decade, and the number of jointly owned entities rose sharply at the same time. Both movements were widely read as evidence that Kazakhstan had become the natural partner market for companies reorganising around the disruption of Russian trade.

The composition tells a different story, because company formation in this period was dominated by Russian registrations. Roughly 11,390 Russian firms were on the register before February 2022, close to a thousand new entities were being registered each month at the peak of mobilisation, and from June 2024 the trend reversed. Western relocation was modest by comparison, with 41 companies from a government list of around 400 having produced tangible outcomes worth just over $1.5 billion by early 2024.

The residue of that period is still visible in the register, where as at 1 December 2025 Russia accounted for 4,823 of the 11,827 registered jointly owned entities and branches, more than China, Türkiye and Uzbekistan combined. The Bureau does not break those entries down by activity, but the foreign-linked register as a whole is dominated by trade and repair at 35.6%, other services at 14.7% and construction at 8.9%, which helps explain why a surge in registrations should not be read as an equivalent surge in industrial partnership, and why the aggregate later flattened without any change in industrial policy.

The period also introduced a factor that had not previously featured in Kazakh partner selection, in that sanctions compliance became a structuring question rather than a background risk. Third-country entities began appearing on Western restriction lists, and in December 2024 Kazatomprom announced that Rosatom's Uranium One had disposed of its interests in the Zarechnoye mine and the Khorasan-U joint venture. The commercial rationale for those particular transactions was not made public and none should be inferred from their timing. What the episode illustrates is a condition that did not previously apply, namely that the identity of a shareholder and its exposure to sanctions can determine whether a joint venture position remains viable or transferable regardless of how the underlying asset performs. The identity and ownership chain of a local partner acquired a legal significance it had not carried before, and that remains one of the practical realities we cover in more detail in our note on the operating reality of doing business in Kazakhstan.

2024 to 2026: the industrial turn

Gross inflows dipped to $17.9 billion in 2024 before recovering to $20.5 billion in 2025, and the headline figure is less interesting than the redistribution beneath it. Foreign investment into manufacturing rose 47.4%, or roughly $1.4 billion, while the extractive sector lost about $3 billion. Domestic capital moved in the same direction, with fixed capital investment reaching KZT 23.5 trillion, close to double the 2019 level, and manufacturing investment tripling to KZT 3 trillion and rising to 12.7% of the total.

The projects behind those numbers are specific and, in most cases, structurally shared. Kia opened a full-cycle plant in Kostanay in October 2025. KazMunayGas, Sinopec and SIBUR began piling at Atyrau in March 2025 on a polyethylene complex of over $7 billion. TotalEnergies took a final investment decision on a one-gigawatt wind and storage project in April 2026. A China-Kazakhstan container hub at Aktau entered service in July 2026, and an American developer closed a controlling interest in one of the world's largest undeveloped tungsten resources in the April before that.

Two qualifications belong with this, of which the first is that exports remain concentrated, with crude oil and crude petroleum products accounting for 46.5% of the total in the first half of 2026. The second is that the manufacturing sector's own sentiment has been weak, since purchasing managers' expectations for the year ahead fell in July 2026 to their lowest level in more than six years, with working capital financing cited as a binding constraint. Investment is arriving faster than operating conditions are improving, and anyone assessing the commercial case should read the market opportunity in Kazakhstan on its own terms rather than inferring it from the investment headlines.

What the registry actually shows

The clearest single piece of evidence on the joint venture question is the divergence between two series that the Bureau of National Statistics publishes side by side every month.

Source: Bureau of National Statistics, main indicators of the number of entities, 1 December 2025. Of the jointly owned total, 11,217 are legal entities and 610 are branches.

The pattern is consistent rather than a one-month artefact, although the comparable monthly series only runs back to late 2024.

Source: Bureau of National Statistics, monthly main indicators of the number of entities. Figures cover registered legal entities and branches. Across the whole register, jointly owned entities now represent about 16% of all foreign-linked entities.

The breakdown by country of origin adds a second finding, in that German investors hold 489 jointly owned entities against 965 wholly foreign ones, a ratio of roughly one in three, whereas for Uzbekistan the ratio is closer to one in fourteen, with Chinese and Turkish investors between the two at about one in eight. Origin correlates closely with the type of business being done, in that shared ownership is markedly more common among investors whose Kazakh activity involves plant, equipment and long-lived assets, and markedly less common among those whose activity is trade.

Read together, these two findings support a specific conclusion, which is that the joint venture in Kazakhstan is not in decline but is becoming concentrated. It is being used less as a general-purpose entry vehicle and more as a purpose-built structure in a defined set of industrial and infrastructure situations. The rest of this article is about which situations those are and why.

Part II. Where joint ventures are actually emerging

Four questions apply to each of the sectors that follow, and the answers are what distinguish a durable structure from a temporary one. They are what is happening in the sector, why the situation produces shared ownership rather than a contract or a wholly owned subsidiary, what each side actually contributes, and whether the venture would stand up if the policy environment changed.

Petrochemicals: capital and offtake, not technology

The Silleno polyethylene complex in Atyrau region is the largest manufacturing project in the country and the clearest example of a capability-driven joint venture. It is owned 40% by KazMunayGas, 30% by Sinopec Overseas Investment Holding and 30% by SIBUR, with capacity of 1.25 million tonnes a year, investment of roughly $7 billion, and production scheduled from 2029. It sits inside the National Industrial Petrochemical Technopark special economic zone.

The instructive detail is what the shareholders do not bring to it. The steam cracker technology is licensed from Lummus, polymerisation from Chevron Phillips and Univation and the butene unit from Axens, while the engineering is executed by Técnicas Reunidas in consortium with Sinopec Engineering. Technology and construction capability were bought rather than contributed.

What could not be bought is the feedstock, and that is where the structure is explained. Ethane is extracted at KazMunayGas's gas separation complex from Tengiz production and delivered by a pipeline the national company is building, which is the Kazakh contribution and the reason the arrangement is equity rather than an offtake agreement. The plant is economically inseparable from a molecule stream controlled by one party, and where a foreign investor's business depends on a resource that a counterparty controls and cannot credibly commit to over thirty years by contract alone, shared ownership is the rational answer.

Uranium: where the state writes the ownership architecture

Uranium is the sector in which the state has gone furthest in specifying the shape of partnership, and at the end of 2025 it went considerably further still. Of the country's thirteen uranium mining projects, three are wholly owned by Kazatomprom and ten are joint ventures with Rosatom, Orano, Cameco and China National Nuclear Corporation.

On 26 December 2025 the President signed amendments to the Code on Subsoil and Subsoil Use covering the hydrocarbon and uranium sectors, and Kazatomprom set out what they mean. Where a new subsoil use agreement for uranium production is awarded to the national company, it may be transferred onward only to an entity in which the national company holds, directly or indirectly, more than 75%, against a previous threshold of more than 50%. Extending an existing agreement, or increasing production volumes or reserves beyond approved levels, is now permitted only on one of two conditions. Either the national company holds at least 90% of the venture, or the foreign partner transfers uranium conversion and enrichment technology to it or to a jointly established entity. Additional exploration at producing deposits is reserved to the national company or to entities in which it holds at least 90%. Existing agreements are untouched, but only until they come up for renewal.

The significance of this reaches well beyond uranium, because the state has put an explicit price on the renewal of an existing joint venture position and the currency is not capital. A foreign shareholder whose contribution is money and a share of the offtake faces dilution toward ten per cent when its agreement expires, while a shareholder able to transfer conversion and enrichment technology keeps its position. That is the argument of this article written into statute, in that access is not a contribution, capital alone is a weak one, and what protects a partner is holding something the counterparty cannot obtain elsewhere.

The consequence is already being priced into existing positions. CGN Mining holds 49% stakes in two joint ventures with Kazatomprom, with licences running to 2030 and 2031, and has said publicly that it is studying the amendments, having received close to $33 million in after-tax dividends from one of those ventures in 2025. Its negotiating lever at renewal is that it possesses the technology the statute asks for, and a partner without that lever would have none.

Uranium is not the only sector where shared ownership is compelled rather than chosen, since foreign ownership ceilings in air transport, telecommunications and media have the same effect. It is, however, the only case where the statute also fixes the identity of the majority partner and the terms on which the minority may remain, which is a materially different negotiation.

The sector has also produced the best-documented illustration of what minority equity does and does not deliver. JV Inkai is owned 60% by Kazatomprom and 40% by Cameco, and on 30 December 2024 Kazatomprom directed the venture to plan for a halt because updated project documentation had not been approved, with production stopping on 1 January 2025. Cameco's own disclosure states that it was informed of the suspension by its partner, that reports received as recently as 26 December had made no mention of the risk, and that it was surprised, although production had resumed by 27 January. Cameco accounts for the venture on an equity basis and states in its filings that it does not have joint control of it.

Nothing improper occurred, in that the majority partner acted to avoid a regulatory breach and the regulatory relationship was its own. The sequence nonetheless demonstrates the central structural fact about Kazakh joint ventures, which is that where the venture's licence to operate is held through the local partner's relationship with the state, a 40% shareholder has neither operational control nor, in practice, real-time information. Equity and control are different instruments, and in this market they come apart more sharply than in most.

Critical minerals: the inversion

The newest wave of resource partnerships runs in the opposite direction. On 29 April 2026, Cove Kaz Capital closed the acquisition of a 70% controlling interest in Severniy Katpar, with the national mining company Tau-Ken Samruk retaining 30%. The venture holds the Northern Katpar and Upper Kairakty tungsten deposits, together reported at around 1.4 million tonnes of tungsten trioxide, in a project valued at about $1.1 billion, and the same company has a separate joint venture with the national geological exploration company Qazgeology on a rare earth project in Kostanay region.

Here the state takes a minority position and the foreign partner leads development, which reverses the older pattern. The Kazakh contribution is the licence, the geological data and a co-investor whose presence signals that the project has passed a national interest test, while the foreign contribution is capital, technical execution and, decisively, the offtake relationship with Western industrial and defence buyers.

The structure should not be assumed to be permanent, because the government and Tau-Ken Samruk have been seeking to reinstate the national company's priority right over exploration and mining licences for critical minerals, a measure expected in 2026. If it is restored, the negotiating position of new entrants in this sector will move toward the uranium model, in which the terms of participation are set before the commercial discussion begins.

Automotive and machinery: buying access to a platform

Kia Qazaqstan LLP was formed in August 2023 as a joint venture between Kia Corporation and the Kazakh group Allur. The Kostanay plant opened in October 2025 with capacity of 70,000 vehicles a year and full-cycle operations including body welding, painting, assembly and testing. Investment has been reported at figures ranging from $200 million at announcement to $310 million at opening, which is a reminder that project values in this market move as projects do.

Kia had been assembling in Kazakhstan since 2021 at Allur's Kostanay plant, so what the joint venture bought was not the ability to assemble vehicles, which the brand already had through a contract arrangement. It bought a dedicated and controlled full-cycle facility together with a partner that holds the regulatory status, the industrial land, the workforce pipeline and the dealer and service network.

That combination is the recurring feature of machinery partnerships in Kazakhstan. The country's industrial assembly platforms are multi-brand contract operations, and their commercial value lies as much in their registration status under industrial assembly agreements and special investment contracts as in their equipment. Our analysis of Kazakhstan's agricultural machinery market sets out the same mechanism in a sector where policy has pushed further, and where the binding constraint on a new entrant is not manufacturing capacity but whether an existing platform's brand portfolio permits a competitor in the same power band.

The sector as a whole is growing quickly, with vehicle production exceeding 96,000 units in January to July 2026 and mechanical engineering output up 22% to KZT 3.2 trillion, while the share of small-unit assembly reached 28.3%. Component localisation is following, including a joint project between Astana Motors and South Korea's Hands Corporation for aluminium wheel production.

Power: the state as minority co-investor

Renewable energy has produced the most standardised joint venture pattern in the country, and the Mirny wind and battery project is the template for it. The project reached final investment decision in April 2026 with TotalEnergies holding 60% and Samruk-Energy and KazMunayGas 20% each, on an investment of $1.2 billion of which about 75% is externally financed, with a lending consortium including the EBRD, Proparco, DEG, the Development Bank of Kazakhstan, Société Générale, QNB, China Construction Bank and Standard Chartered, and a 25-year power purchase agreement with the government signed in 2023. A comparable one-gigawatt project in Zhambyl region, on which construction began in June 2026, is held 40% by Masdar, 40% by W Solar, 18% by Qazaq Green Power and 2% by the Kazakhstan Investment Development Fund.

The state's minority stakes in these ventures are not control positions and are not intended to be. They align the counterparty to the power purchase agreement, the grid operator and the land allocation with the project's success, and they make the structure bankable for development finance institutions. In effect the state is buying alignment risk out of the project by taking equity in it.

One detail deserves attention because it recurs across these structures. At Mirny the two state holdings add to 40% and at the Zhambyl project they add to 20%, and under the Kazakh partnership rules discussed in Part V a 40% combined holding is enough to block a qualified-majority decision. Two separate minority stakes held by entities under common ownership are not the same thing as a dispersed minority, and the arithmetic should be done before the term sheet rather than after it. The pattern is now close to standard, in that foreign investors hold the controlling economics in Kazakhstan's largest new energy projects while the quasi-state sector holds a minority, with private Kazakh business largely absent.

The contrast with nuclear is instructive, because Kazakhstan's first plant at Ulken is being delivered under an EPC contract signed with Rosatom in September 2026 by Kazakhstan Nuclear Power Plants LLP, with China National Nuclear Corporation selected to lead the second and third plants. At a preliminary cost of around $16.4 billion for two units, the largest energy investment in the country's history is so far structured as an EPC contract with export financing rather than as shared ownership of the asset. Where the state can specify the output and finance the asset, it does not offer equity, and the foreign party does not need it.

Transport and logistics: corridor alignment

The Aktau container hub is a joint venture between KTZ Express, China's Lianyungang Port and Aktau Sea Trade Port. Founding documents were signed in Xi'an in September 2024 and the first phase entered operation on 14 July 2026, covering 9.1 hectares with three rail loading lines. It follows the Kazakh-Chinese terminal at the Xi'an dry port and sits alongside an intermodal terminal agreement in Baku involving the Port of Baku, KTZ and Xi'an Free Trade Port.

The logic here differs again from that of resources or manufacturing. A container moving from Xi'an to Central Europe crosses several national rail systems, a sea leg and multiple customs regimes, and no single operator can contract for reliability across that chain, so equity in the terminals at the handover points becomes the mechanism by which the participants make each other's throughput commitments credible. Kazakhstan Temir Zholy is spending around $10 billion on rail, ports and cargo capacity, and is examining terminal acquisitions in Romania, Hungary and Germany on the same reasoning.

Pharmaceuticals: the counter-example

Pharmaceuticals is the most useful sector in this survey precisely because the same policy pressure produced a different structure. The state wants domestic production, has a single distributor with buying power, and has set a target of 50% locally produced medicines, and it has obtained what it wanted almost entirely without shared ownership, through contract manufacturing at Kazakh sites and the outright acquisition of Kazakh producers by foreign groups.

AstraZeneca signed a long-term agreement with SK-Pharmacy for contract production and technology transfer at Nobel Almaty Pharmaceutical Factory, and Roche entered a similar arrangement covering the localisation of three biotechnological oncology products. At the start of 2026 the single distributor had 83 long-term contracts in force with 31 domestic producers covering 2,022 product names, with sector investment rising 56% in 2025 to $142.8 million.

The reason for the difference is worth stating plainly. In pharmaceuticals the scarce assets are a GMP-certified site and a place on the procurement register, and both can be accessed contractually because the counterparty's incentive to perform is renewed at every contract cycle. In petrochemicals the scarce asset is a molecule stream over thirty years and in tungsten it is a deposit, neither of which can be rented. The choice between a joint venture and a contract in Kazakhstan turns almost entirely on whether the dependency is renewable or permanent.

What the named ventures have in common

Set side by side, the structures that have actually been built since 2024 share a pattern, in that both sides hold something the other cannot buy on the open market.

What is absent from the columns is as telling as what appears in them, since none of these ventures exists because the foreign party needed a guide to the local environment.

Geography: what location actually changes

Regional concentration in Kazakhstan is real, but it is usually described in the wrong terms. Almaty attracted $8.58 billion in gross foreign investment in 2025, with services at 85.9% of city output, while Astana attracted $4.1 billion, 2.6 times the previous year, and hosts an AIFC with more than 6,000 registered companies from over 90 countries and $26.3 billion mobilised by July 2026. Atyrau carries hydrocarbons and petrochemicals, Kostanay carries vehicle and machinery assembly, Karaganda and Ulytau carry metals and critical minerals, and the Almaty and Zhambyl regions carry the new power capacity.

Where a plant goes is of course a strategic question, decided by power, feedstock, logistics, workforce, suppliers and proximity to customers or the border. Outside specific regimes such as the special economic zones, however, location does not generally change the corporate law or the tax treatment that governs the venture itself. What it changes is the scarcity of the assets worth partnering for, and the number of counterparties who hold them. A foreign manufacturer seeking a vehicle programme in Kostanay is negotiating with a handful of platforms, several of which sit within related ownership structures and most of which already carry competing brands, whereas a company establishing a regional commercial headquarters in Almaty faces no equivalent constraint and, in most cases, has no reason to share ownership at all.

The special economic zones are the exception, because there location does carry its own fiscal and regulatory content. Across 18 zones, 577 projects have attracted KZT 11.1 trillion of investment and created more than 39,000 jobs, with resident enterprises generating KZT 865.4 billion in tax against KZT 507.4 billion of state infrastructure spending. An established zone position can therefore represent a specific economic contribution by a partner, although the continuity of individual rights and incentives on a change of ownership has to be verified rather than assumed.

For joint venture structuring, then, geography matters less because it changes the law than because it changes scarcity. In concentrated industrial regions the credible partner set is small, the local side knows it, and that belongs in the timetable and the valuation rather than in the site selection analysis.

Part III. Why the model is changing

Localisation is changing the economics of partnership

Kazakhstan has assembled an unusually dense set of instruments to shift production onshore, and they now work together rather than separately.

The investment agreement has been the anchor of the system since 2021, available for projects above KZT 32 billion, roughly $69 million, and offering stability of key parameters for up to 25 years. Uptake has accelerated sharply, with 66 agreements worth more than KZT 17.8 trillion signed in total, of which six in 2024, thirty in 2025 and twenty-five in the first eight months of 2026. Kazakh Invest currently carries 215 projects worth $78.6 billion, of which 93 worth $32.2 billion are in implementation.

Industrial assembly agreements and special investment contracts provide the sectoral layer, and in vehicles and machinery, where the framework is most developed, they offset the utilisation fee that importers pay in full and exempt components and equipment from import duties and VAT. Special economic zones provide the territorial layer.

Procurement provides the demand. The Law on Public Procurement in force from 1 January 2025 made offtake contracts a ground for single-source procurement, and in subsoil-use procurement, where a domestic producer exists, minimum in-country value requirements of 50% for works and services and 80% for design work apply. Long-term and offtake contracts between subsoil users and the quasi-public sector reached KZT 466 billion in the first half of 2026, and Samruk-Kazyna's purchases from domestic manufacturers exceeded KZT 2.1 trillion in the same period, more than double the prior year. In oil and gas alone, long-term agreements with domestic producers rose from KZT 21 billion in 2024 to KZT 49 billion in 2025, with ten types of original equipment localised and fifteen more planned by 2027.

This is a coherent system and it is the reason industrial investment has moved, but it raises the question that determines whether any given partnership is durable. When does localisation create genuine industrial economics, and when does it merely make local production preferable because policy rewards it?

The test is straightforward to state, if uncomfortable to apply, and it is whether the plant would be viable at this scale if the preference were removed. For Silleno the answer is probably yes, because the core advantage is ethane at a delivered cost no importer can match, although the project also carries special economic zone tax relief. For a vehicle plant sized at 70,000 units against a domestic market of twenty million people, the answer depends entirely on export access to the Eurasian Economic Union and Central Asia, which is a commercial question rather than a policy one. For a component plant sized to a single OEM's local volume, the answer is usually no, and the investment case rests on the OEM's continued presence rather than on the component business itself.

This distinction transfers directly into partner selection, because where the economics are policy-created, a meaningful part of what the local partner contributes is the durability of a policy position, and policy positions can be revised. Where the economics rest on a resource, a corridor or an installed asset base, the contribution is physical and survives a change of government priorities. Investors evaluating projects in this category should read our note on industrial investment in Kazakhstan, which deals with the execution side of the same problem.

The state and quasi-state sector

No account of partnership in Kazakhstan is complete without the state's own economic footprint, which remains the largest single variable in the market.

How large that footprint is depends entirely on what is being measured, and the published figures diverge by a factor of nearly three. State-owned enterprises have been put at around 15% of GDP generation, while the combined economic footprint of Samruk-Kazyna and Baiterek has been estimated at roughly 40% of GDP on IMF work. The two are not measuring the same thing, and neither should be quoted as though it were the state's share of the economy. What is not in dispute is the reach. On the OECD's review, state-owned enterprises operate in at least twenty of the country's thirty economic sectors and account for around 6.2% of national employment, and the IMF's 2025 Article IV consultation, concluded in January 2026, found that the state footprint remains large and constrains private sector development, with planned budget consolidation for 2026 largely offset by expanding quasi-fiscal activity by state-owned enterprises.

The direction of travel is nonetheless toward reduction, with a presidential decree on liberalising the economy signed in May 2025, a moratorium on establishing new state-owned enterprises running to 31 December 2026, and around 500 quasi-public assets worth over KZT 2 trillion due to be transferred to the competitive environment through phased privatisation and listings running to 2030.

For a foreign investor the operative question is what the state's footprint does to the case for choosing a local partner, and the honest answer has changed. Ten years ago a well-connected local partner shortened the path to permits, land and procurement access. Today the state has built its own channels for that, through the investment agreement regime, the Fast Track permitting corridor, the three-tier system under which regional authorities identify projects and seek investors directly, and Kazakh Invest as a single point of contact. Access is increasingly something the state supplies rather than something a partner brokers.

That makes the distinction between a partner and an intermediary sharper than it used to be. A state or quasi-state partner today typically brings something specific and verifiable, such as a subsoil licence, a deposit, a grid connection, feedstock, an existing plant, a power purchase agreement counterparty relationship, or capital and the willingness to fund. Where the proposition is connections, it belongs to a different category of risk, and it is the category whose value has fallen fastest since 2020.

Two contrasting cases frame the period, and neither turned on the ownership percentage. Air Astana was created in 2001 as a 51/49 joint venture between the Kazakh state and BAE Systems, on a combined initial shareholder investment of about $17 million. It ran for more than two decades and listed in February 2024 in London, Astana and Almaty, at which point Samruk-Kazyna moved from 51% to about 41% and BAE from 49% to a holding reported at between 15% and 17%, and BAE sold the remainder in March 2026, completing a full exit through the capital markets. Qarmet, formerly ArcelorMittal Temirtau, was wholly foreign owned for 28 years and was transferred to the state-linked Qazaqstan Investment Corporation in December 2023 on an as-is basis following the Kostenko mine disaster. The consideration was widely reported at around $286 million, against assets ArcelorMittal had carried at $1.8 billion in its own accounts three months earlier.

The comparison is not that joint ventures are safer than subsidiaries, but that the ownership percentage determined neither outcome. What determined them was whether the relationship with the state remained functional, and whether an exit route had been built before it was needed.

Part IV. The legal and institutional framework

What the framework allows

Kazakhstan is a civil law jurisdiction in which the principal vehicles are the limited liability partnership, which is the default for operating businesses, and the joint stock company, used where capital markets access matters. Branches are available for non-trading presence.

There is no general ceiling on foreign ownership, but sectoral limits apply and their scope is narrower and more specific than the shorthand usually suggests. Mass media is capped at 20% of shares or interests held directly or indirectly, and air transport is capped at 49%, with the restriction reaching the exercise of control as well as the shareholding. The telecommunications limit is not a general one, in that it applies at 49% to an entity operating as an operator of national and international long-distance communication which also owns terrestrial transmission lines, and it can be exceeded on a positive government decision taken on the advice of the communications authority in agreement with the national security authorities. Agricultural land is closed to foreign ownership and to Kazakh entities with foreign participation, and land in designated border areas is closed outright. The state holds a pre-emptive right over strategic subsoil assets, and national security review can reach upstream corporate reorganisations executed entirely outside Kazakhstan where they change beneficial control of a qualifying project.

The 2025 privatisation of Qazaq Air shows what a ceiling does to a structure in practice. The bidding terms allowed a foreign investor up to 49% while a Kazakh investor could take 100%, so Vietnam's Sovico Group took 49% through Central Asia Aviation Holdings, with a further 2% going to Kazasia Holdings, an AIFC-registered Kazakh vehicle, and Samruk-Kazyna retained 49% with an option to sell it down over five years. Samruk-Kazyna recorded the transaction as a loss of control and the airline has since been rebranded under the buyer's marque. A cap of this kind does not stop a foreign investor from running the business, but it determines the shape of the cap table and adds a third party whose only function is to satisfy the rule.

Merger control sits in the Entrepreneurial Code, with the notification threshold set by a formula based on ten million monthly calculation indices. June 2024 amendments moved acquisitions of tangible and intangible assets from prior consent to notification and expanded the statutory concept of control. The sixth antimonopoly package provides for lowering the shareholding trigger from 50% to 25%, and the OECD's peer review, published in November 2025, recorded adoption as expected shortly afterwards, so the operative threshold should be confirmed at the time of any transaction rather than assumed. The same review found the regime fragmented and a continuing source of legal uncertainty for business.

Two further changes took effect on 1 January 2026, and both matter for partnership economics.

The investment incentive regime was rebuilt, with amendments to the Entrepreneurial Code abolishing the priority and special investment project mechanisms and replacing them with three contractual instruments, namely the investment agreement, the investment obligations agreement and the simplified investment contract. Ministerial Orders 106 and 107 of 15 October 2025 govern the assessment of preference effectiveness and the setting of investor counter-undertakings, with effectiveness evaluation beginning on 1 July 2026. The regime has moved from status to negotiation, in that an investor no longer qualifies for incentives by category but commits to specific obligations in exchange for them. In July 2026 the Government Project Office reviewed a further package, including proposals to restore priority investment contracts with a full fiscal package.

The new Tax Code raised VAT from 12% to 16%, kept corporate income tax at 20% with a 25% rate for banks and gambling, introduced a progressive personal income tax, and cut the overall stock of tax benefits by around a fifth.

For a foreign corporate shareholder the operative provision is not the personal scale but the withholding tax on dividends, and it is the line that changes a joint venture's return most directly. From 1 January 2026 the base rate is 15%. A recipient holding at least 25% of the capital, directly or indirectly, pays 5% on dividends up to 230,000 monthly calculation indices, roughly $1.75 million, and 15% on the excess, while a recipient resident in a jurisdiction on Kazakhstan's preferential-tax list pays 20%. Treaty relief is available but conditional on a residency certificate and on the recipient being the beneficial owner rather than an intermediary. The practical effect is a low rate on modest distributions and a materially higher one on a joint venture that becomes genuinely cash-generative, which is precisely the case the investment model is built on.

The AIFC provides the common-law option, and its court and arbitration centre had completed and enforced 5,415 cases by 19 August 2026, comprising 289 judgments, 1,155 arbitral awards and 3,971 mediation agreements, with around 90% of arbitration cases having no direct connection to the centre itself.

What the framework does not solve

The gap that most often surprises foreign investors lies not in the ownership rules or the tax treatment but in the enforceability of the shareholders' agreement.

The Law on Limited and Additional Liability Partnerships does not recognise the shareholders' agreement as an instrument at all. On the prevailing reading, Article 1114 of the Civil Code applies Kazakh law to agreements concluded in relation to a legal entity established in Kazakhstan, which restricts the ability to place the arrangement under foreign law. Where a shareholders' agreement conflicts with the foundation agreement or the charter, Kazakh courts have tended to give precedence to the constitutive documents, which are the instruments the law expressly provides for.

The practical consequence is specific and uncomfortable, because in a Kazakh joint venture the enforceable content of the deal sits in the charter and the foundation agreement rather than in the document the parties spent the most time negotiating, so reserved matters, board composition, veto rights and appointment mechanics have to be written into the constitutive documents to be reliable. This is the principal reason sophisticated structures place the shareholder arrangements in an AIFC or third-country holding vehicle above the Kazakh operating entity, and accept the additional layer as the price of contractual certainty.

Where that vehicle is domiciled is a second decision rather than a formality, because the jurisdiction that offers the most convenient contractual regime may be on Kazakhstan's preferential-tax list, in which case the dividend rate is 20% and treaty relief is unavailable. Governance certainty and repatriation cost are decided by the same choice, and they pull in opposite directions often enough that the two questions should be answered together.

The broader point is that Kazakh law defines what a joint venture may look like, but it does not resolve who controls it.

One caveat applies to this whole part, which is that the framework described here moved substantially within twelve months. The Entrepreneurial Code amendments and the new Tax Code both took effect on 1 January 2026, the uranium amendments were signed on 26 December 2025, and a further investment package was under review in July 2026. Everything above is a map of where the constraints sit and how they interact, not a substitute for a position confirmed with Kazakh counsel against the text in force on the day a structure is signed.

Part V. What the market evidence means for joint venture design

The preceding sections set out the evidence, and what follows is interpretation, which is where the analysis becomes a set of decisions.

What 51% actually buys

A controlling stake buys less than most investors assume, and the shortfall is statutory rather than a matter of what was negotiated.

Under Article 48 of the LLP Law, a qualified majority of three quarters of those present and represented is required for a specific and quite short list of matters. They are amendment of the charter, including any change to the size of the charter capital, reorganisation or liquidation, the compulsory buy-out of a participant who has caused the partnership significant damage, and the pledge of the entire property of the partnership. Disproportionate additional contributions by one participant require the consent of all the others, and everything else is decided by simple majority of those present unless the charter says otherwise.

That list is narrower than most investors assume, and the omissions matter more than the inclusions. Appointing and dismissing the executive body is a simple-majority decision, as is the approval of financial statements and the distribution of net income, and so, by default, is the approval of a transaction disposing of property worth 51% or more of the partnership's assets, which sits within the general meeting's exclusive competence but attracts no statutory supermajority.

The consequences run in both directions, in that a 51% holder cannot unilaterally amend the charter or increase the capital, but can, unless the charter provides otherwise, appoint the chief executive, set the dividend and approve the sale of over half the business. A 49% holder can block the first set of decisions and none of the second. Both parties routinely believe they have bought something the statute does not give them.

Attendance is the other half of the mechanism and is frequently overlooked. Because the threshold is calculated on those present and represented, a reconvened meeting is competent regardless of how many votes attend, but where the participants present hold less than half the total votes it may only decide matters that do not require a qualified majority. A minority partner's protection over the charter and the capital therefore survives its own absence, whereas its position on ordinary business does not.

Layered on top of that is the Kazakhstan-specific factor illustrated at Inkai, which is that much of what makes a joint venture here valuable is held in one party's name. The subsoil licence, the industrial assembly agreement, the special investment contract, the special economic zone residency, the entry on the procurement register, the GMP certificate and the power purchase agreement do not move with the shareholding. A partner holding 40% of the equity and none of the registrations is exposed to decisions taken in a regulatory conversation it is not part of.

Three things therefore need to be separated in any Kazakh joint venture, and they are not the same thing. Equity ownership, economic dependence and management control each behave differently in this market, and the questions that determine the third are practical rather than percentage-based. They include who signs the bank mandate, who holds the licences and registrations and what happens to them on a change of shareholder, who appoints the first chief executive and finance director, whether the annual budget and capital expenditure plan are reserved matters in the charter or merely in the shareholders' agreement, whether related-party supply requires board approval, who employs the plant manager, and who owns the customer contracts and the intellectual property. These are the mechanics we work through in governance advisory in Kazakhstan engagements, and they are usually settled badly when they are settled at the end of a negotiation rather than the beginning.

Partner selection has become more important, not less

If access is increasingly supplied by the state rather than brokered by a partner, then the value of a partner has to come from somewhere else, and a single question separates a genuine economic partner from an intermediary.

Would the foreign investor still choose this partner if regulatory access were removed from the equation?

What survives that test in Kazakhstan today is a short and concrete list, running to an existing plant with regulatory status, a distributor network with service coverage in the regions where the volume is, a deposit or a licence or a feedstock stream, a grid connection or a signed power purchase agreement, land inside a special economic zone, capital together with a demonstrated willingness to fund a call on it, and a record of having executed a comparable project. What does not survive the test is introductions.

Beyond that first filter the assessment work is conventional but rarely done properly, and it has to establish whether the partner competes with the venture directly or through affiliates, whether the platform's existing brand arrangements permit the investor's products, who owns the customer relationships and whether they transfer, who owns the intellectual property created inside the venture, whether the venture can operate if the partner withdraws support, whether related-party transactions are the partner's principal source of return, and whether the resource the partner contributes will still be scarce in five years. The brand question is frequently the binding constraint in multi-brand assembly, and the scarcity question is the one most often skipped, since a regulatory position that is scarce today because a licensing regime is new may not be scarce once thirty competitors hold the same document.

The other half of the exercise is verifying that the partner can do what it says it can. Ownership structures in Kazakhstan are frequently layered, decision rights sit with beneficial owners rather than with commercial departments, and published financials often describe only part of a group, so reaching the person who actually decides is a large part of the work and it is not desk research. Where the alternative is buying the platform outright rather than partnering with it, the assessment moves into the territory covered in our note on M&A advisory and buyer-side risk in Kazakhstan.

Governance, deadlock and exit

Disputes in Kazakh joint ventures are unusually predictable, because the pressures that produce them are specific to this market rather than generic to shared ownership.

The first is the capital call, since localisation commitments deepen over time, component content requirements ratchet, and the investment required in year four is routinely larger than the case modelled in year one, so if one partner cannot or will not fund, the dilution formula becomes the most consequential clause in the documents. The second is related-party procurement, and where the local partner also supplies the venture, the price of that supply is the point at which the two sets of interests separate, and it is rarely a reserved matter unless somebody made it one. The third follows directly from the statutory position set out above, in that the reserved matters list has to be drafted into the charter rather than assumed, because budget, capital expenditure, senior appointments, borrowing and dividend policy are all simple-majority decisions by default and every one of them is a live source of dispute.

The fourth and fifth pressures are about what happens at the end. Change of control provisions need to reach the shareholder's own parent, since a transaction executed abroad can change who the partner is without touching the Kazakh entity's documents at all. The exit provisions in turn have to answer what becomes of the licences, registrations, zone status, brand rights and key employees, which is the point most often omitted, because in a market where the venture's regulatory identity may sit in the partner's name, a clause that transfers shares but not status transfers very little.

Exit architecture is part of investment design rather than documentation for the end of the relationship, and Air Astana is the local proof. It ran for a quarter of a century and closed with a listing and an orderly divestment rather than a dispute, because the route out was built into the structure from the start.

Part VI. Outlook to 2030

Forecasting the number of joint ventures in Kazakhstan would produce a figure with no basis behind it. What can be assessed is the set of structural drivers that will determine whether the model expands, contracts or narrows further.

Six drivers

Industrial localisation policy. The direction has been consistent for six years and the instruments have been reinforced rather than relaxed. Against that, the new Tax Code cut the stock of benefits by around a fifth, the investment preference architecture was rebuilt with effect from January 2026 and is already under review again, and the effectiveness assessment beginning on 1 July 2026 gives the state a mechanism to narrow eligibility. Policy support is strong but political rather than structural.

Procurement and offtake economics. This is the most durable of the drivers because it operates through demand rather than subsidy. Samruk-Kazyna's domestic purchasing, the expansion of offtake contracts with subsoil users, and the fifteen additional oil and gas equipment localisations planned by 2027 create predictable volume for anyone with an approved local production position, and predictable volume is what makes a plant financeable.

China and the Middle Corridor. Under a Kazakhstan-China working group commitment, container train volumes on the route are targeted at 600 a year in 2025 and 2026, 1,000 in 2027 and 2,000 in 2029, against KTZ's $10 billion capacity programme and an operational container hub at Aktau. If those targets are approached, Kazakhstan's role shifts from an end market of twenty million people to a production platform serving Central Asia and beyond, which changes the plant economics of every localisation decision.

Investment diversification. The critical minerals memorandum with the United States, the Cove Kaz transactions, and rising flows from the Gulf and Southeast Asia are broadening the investor base beyond the traditional European and Russian sources. Investment from the United Arab Emirates roughly doubled to $1.6 billion in 2025, with Qatar at $1.3 billion and Singapore at $1.4 billion. New entrants tend to arrive without established local relationships, which raises the value of a partner with an installed asset base.

Privatisation and the quasi-state economy. Around 500 state-linked assets are due to move to the competitive environment by 2030, with listings planned for Kazakhtelecom and Qazaq Green Power among others. Privatisation converts some joint venture questions into acquisition questions, and it also creates a new category of partnership, in which the state sells control to a strategic investor while retaining a large minority and an option to sell it later, as it did at Qazaq Air in 2025. The important qualification is that the state is moving in two directions at once and the two should not be read as a single trend. It is withdrawing from competitive sectors through privatisation and listings while tightening its ownership leverage in strategic ones, since the December 2025 uranium amendments and the expected restoration of Tau-Ken Samruk's priority right over critical minerals licences both run the other way. An investor's negotiating position in 2030 will depend less on the general direction of state policy than on which of those two categories its sector sits in.

Tax and regulatory architecture. VAT at 16%, the new dividend treatment, the reduction in benefits and the pending restoration of priority investment contracts will each shift the after-tax return on particular structures. The direction is toward fewer, larger, individually negotiated packages, which favours investors capable of running a structured negotiation with the state.

Three paths to 2030

Access-driven joint ventures are the least durable of the three, because where the partner's contribution is navigation of the local environment, the state's own instruments are steadily displacing the function. The uranium amendments are the sharpest illustration of where that leads, in that a shareholding held for access rather than for capability now has a defined expiry date and a stated price for renewal. This category will not disappear, but it will increasingly be priced as what it is.

Localisation-driven joint ventures will remain significant in specific industries, particularly vehicles, machinery, components, building materials and food processing. Their durability depends entirely on whether the underlying operation would survive the removal of the preference that created it. Some would and many would not, and the latter are effectively policy positions held in corporate form.

Capability-driven joint ventures are the mature form and the one the evidence of 2024 to 2026 most clearly supports. Both sides contribute a real asset, with feedstock set against capital and offtake at Silleno, a deposit against execution and Western market access at Katpar, an assembly platform and dealer network against a brand and technology at Kia, a licensing and grid position against capital and development capability at Mirny, and a corridor node against volume at Aktau. These structures are not built to obtain access but because neither party can produce the outcome alone.

By 2030 the question for a foreign investor is unlikely to be whether local partners are necessary in Kazakhstan, but which specific economic function requires a partner, and whether shared ownership is genuinely the best way to obtain that function rather than merely the most familiar one.

Executive implications

A board considering Kazakhstan has four structures available to it rather than one, namely a wholly owned subsidiary, a contractual relationship with a distributor or manufacturing partner, an acquisition, and a joint venture. The evidence set out above supports a single organising conclusion.

A joint venture should be the consequence of an economic dependency the investor deliberately chooses to share, not the default response to an unfamiliar market.

Six implications follow from that proposition, and each bears on a decision a board actually has to take.

Start from the dependency, not the structure. Identify precisely what the business will depend on that it cannot own or build, whether that is a feedstock stream, a licence, an approved production platform, a grid connection, a dealer and service network or a place on a procurement register, and then ask whether the dependency is permanent or renewable. Permanent dependencies justify equity, while renewable ones are usually better handled by contract, which is what the pharmaceutical sector demonstrates.

Test the dependency against all four alternatives, explicitly. The registry evidence shows wholly foreign-owned entities growing at five to eight per cent a year while jointly owned entities stand still, which means a large number of investors are concluding that they do not need a shareholder. That conclusion deserves to be reached deliberately rather than by default in either direction. Our note on market entry in Kazakhstan sets out how the entry mode decision should be framed, and the capital commitment test applies before the structuring question rather than after it.

If the dependency is regulatory status alone, price the acquisition alternative. A partner whose contribution is an industrial assembly agreement, a special investment contract or a zone residency is contributing a document, and documents can sometimes be bought with the entity that holds them, which avoids twenty years of governance friction. Where they cannot, the terms of the joint venture should reflect that the contribution is finite rather than ongoing.

Treat the charter as the joint venture agreement. Given the enforceability position under Kazakh law, anything that matters commercially has to appear in the constitutive documents, and the statutory protections a minority enjoys by default cover four matters only. Everything else, including the chief executive, the budget, related-party supply, the dividend and even the sale of half the business, is a simple-majority decision unless the charter says otherwise. Where certainty is essential, put the shareholder arrangements above the Kazakh entity in a common-law holding structure, and check the tax cost of the jurisdiction chosen before choosing it.

Assume the policy environment will change at the margin. The incentive architecture was rebuilt with effect from January 2026, a further package was under review by mid-year, and it will be adjusted again. Model the venture without the preference before committing, and know which of the two partners bears the loss if it is withdrawn.

Design the exit before funding. The valuation method, the put and call triggers, the change of control provisions and the disposition of licences, registrations, brand and key people all have to be settled while the parties still agree with one another. In a market where the venture's regulatory identity may sit in the partner's name, an exit clause that transfers shares but not status transfers very little.

None of this argues against joint ventures in Kazakhstan, since the largest and most credible industrial projects in the country are shared-ownership structures, and they are shared for good reasons that would not disappear if the policy environment changed tomorrow. The argument here is narrower and, we think, more useful. Kazakhstan has become a market where the case for a partner has to be made in specific terms, asset by asset, and where the answer is increasingly either a very good joint venture or none at all.

How we work on mandates of this kind

Tretiakov Consulting advises foreign investors and industrial companies on entry structure, partner assessment, joint venture governance and transaction execution in Kazakhstan. Mandates are led directly by Illia Tretiakov and combine market assessment, counterparty qualification and transaction-side management judgement.

If you are weighing a joint venture, an acquisition or a wholly owned position, we would be glad to discuss it with you.