A foreign technology company that wants a development team in Belarus — to build for the group, or to serve its own Western clients — usually arrives at the idea of a High-Tech Park subsidiary through its tax benefits, and usually with a particular picture in mind: a low-cost entity that bills the parent a modest fee, leaving little taxable in Belarus and most of the value where the parent sits. The benefits are real, and the model is a legitimate and common one. But that particular picture gets the advantage in the wrong place. The tax saving is not in setting a low intra-group price — that is precisely the thing the rules exist to adjust — and a company that builds its plan around underpricing the service is building around the one feature that will not survive scrutiny. The genuine advantage is the regime itself, applied to an operation that is real and priced at market. This article is about how that works, and what it asks of you.
It sets out the model — a Belarusian HTP subsidiary that services the parent or the parent’s clients — and what the HTP regime actually gives it, stated with the qualifications rather than as a headline. It then turns to the part that matters most: why the intra-group fee is governed by transfer pricing and has to be set at arm’s length, why the subsidiary needs real substance rather than a nameplate, and who ends up owning the intellectual property in a service model. And it is honest about a question the current environment forces, which is whether servicing Western clients from a Belarusian entity is realistic at all. Throughout, the theme is the one that runs through all sensible cross-border structuring: the efficient structure and the defensible structure are the same structure, and building for the second is how you keep the first.
Start with the shape of the thing, because it determines everything that follows about how it is taxed and priced.
The model is a captive service centre — a development centre owned by a foreign parent and set up as a Belarusian company, which becomes a resident of the High-Tech Park and provides software development and IT services either to the parent itself or to the parent’s clients, and is paid for those services. It is a service arrangement, not an asset one: the subsidiary does the work and is compensated for it, rather than owning what it builds.
That distinction matters, because it separates this model cleanly from the alternative in which a Belarusian subsidiary owns and licenses intellectual property, which carries a different tax and transfer-pricing profile and which we cover separately in our writing on moving and holding IP in an HTP subsidiary. The two are a genuine choice, and this piece is about the service side of it — the arrangement most foreign groups actually want when what they need is engineering capacity rather than an IP-holding vehicle.
The benefits are substantial and worth setting out plainly — with the qualifications that keep the picture honest, because “nothing to pay” is not quite the truth.
An HTP resident is exempt from profit tax on its core activity, which for a development centre means its service income is not taxed in the ordinary way — but this is not quite “zero,” because in place of profit tax a resident pays one per cent of its gross revenue as a contribution to the Park administration. Its turnover within HTP activity is not treated as subject to VAT, and services exported to a foreign parent or foreign clients fall outside VAT — with one qualification worth knowing, that buying services from abroad can create an import-VAT obligation on the subsidiary as a tax agent, which needs checking case by case. Dividends paid up to the foreign parent carry a five per cent withholding, reduced further where a double-tax treaty applies on the usual conditions. And employees, including foreign specialists, are taxed at a favourable personal rate and benefit from simplified work and residence arrangements.
Taken together this is a genuinely light regime for a services operation — the fuller withholding and treaty picture is in our writing on corporate tax in Belarus, and the regime is administered by the High-Tech Park. The important thing for planning is what this regime does not do, which is the subject of the next section.
This is the pivot of the whole article, and the point most worth grasping before any structure is built.
The instinct many parents bring is that the tax play is to have the subsidiary charge them as little as possible, so that little profit lands in Belarus and most of the value stays with the parent. That instinct is not just unnecessary here; it is actively the wrong move, because setting an intra-group price below its market level is the exact pattern that transfer-pricing rules are built to catch and reverse. The saving a company is reaching for by underpricing does not materialise — the price gets adjusted back to market, and tax and penalties follow — while the saving it could have had, from the favourable regime itself, was available all along without any of that risk.
Put plainly: you do not reduce your tax by underpaying the Belarusian subsidiary for its services. You reduce it through the regime, applied to a fee set where an independent party’s would sit. The company that understands this stops trying to engineer the price and starts making sure the price is right, which is both safer and, once the regime is in play, no less efficient.
With the principle clear, the practical question is how the fee is actually set — and there is a standard, defensible way to do it.
Because a parent and its subsidiary are related parties, the fee the subsidiary charges the parent falls squarely within Belarusian transfer-pricing rules, and it has to be set at arm’s length — the price an independent service provider would command for the same work. For a captive development centre the usual and accepted method is cost-plus: the subsidiary is compensated for its costs plus a margin that reflects what a comparable independent provider would earn, so that the Belarusian entity books a proper, market-level return on the services it performs rather than a suppressed one.
That margin needs to be supportable, and the pricing needs to be documented, because in a transfer-pricing review the burden is on the taxpayer to show the fee reflects market value; for larger operations, an advance pricing agreement with the tax authority can fix the approach in advance and remove the uncertainty. A fee built to move profit out of Belarus is what an adjustment exists to unwind, with the tax and penalties that come with it; a fee built to reflect the value of the work is what stands. The transfer-pricing rules sit with the tax authority, within the framework set by the Tax Code.
Pricing is one half of a defensible structure; the other is that the operation behind it is real.
The HTP regime is built for genuine technology companies, and residency reflects that — it requires a Belarusian legal entity actually carrying out the Park’s permitted activities, supported by a business project that has to demonstrate real innovativeness and export potential, not a shell that exists to route invoices. Beyond the residency requirement, substance is what makes the whole arrangement withstand scrutiny: where the development work is genuinely performed, whether there are real engineers doing real work in Belarus, whether there is a true commercial operation rather than a conduit — these are the questions that determine whether the structure holds when it is looked at.
A company that builds an actual development centre, with people and work located in Belarus, has substance on its side; one that sets up a billing entity and hopes the regime does the rest has the weakest possible position if the arrangement is ever examined. The regime rewards a real service operation. It does not reward a nameplate, and it was not designed to.
A service model answers the IP question in a particular way, and it is worth being clear about it up front.
In a pure service arrangement, the intellectual property the subsidiary develops generally belongs to the parent. The work is done for hire — the deliverables are the parent’s, produced under the service relationship — and the subsidiary is compensated for performing the work rather than for owning its output. That is, in fact, one of the things that makes the service model cleaner than the alternative: there is no transfer of IP from the subsidiary to the parent to price and defend, because the IP arises in the parent’s hands from the start, so the transfer-pricing question is confined to the service fee rather than extending to an asset transfer. The service fee still has to be at arm’s length, as above; but the group avoids the separate and more demanding exercise of valuing and moving intellectual property between related parties.
Here the question turns from tax structure to a harder, more current one — who the subsidiary can actually bill, and whether the answer you want is available in 2026.
There is a transfer-pricing difference between the two audiences, and then a practical reality that sits on top of it. Servicing the parent is an intra-group transaction, so the fee is governed by transfer pricing and has to be set at arm’s length in the way already described. Servicing the parent’s Western clients directly is a third-party transaction, which is cleaner in transfer-pricing terms — a price agreed with an unrelated client is arm’s length by its nature, with none of the related-party pricing burden. On tax alone, then, direct client work is the more straightforward of the two.
But tax is not the whole question in the current environment, and it would be dishonest to present it as though it were. Whether Western clients will contract with a Belarusian entity at all, how payments would be routed, and what sanctions exposure the arrangement carries are real and material questions in 2026, and for many businesses they, rather than the tax position, will decide whether the direct-client model is viable. This is not a reason to abandon the idea, but it is a reason to assess those practical and legal realities honestly and specifically before building a business around servicing Western clients from Belarus, rather than to assume the tax efficiency settles the matter. It does not, and a clear-eyed look at the practicalities is part of doing this properly.
None of the regime is available until the subsidiary is inside it, so the way in is worth knowing.
The benefits described here belong to HTP residents, and residency is something the subsidiary has to obtain rather than something that comes with incorporation. It requires a Belarusian legal entity whose activities fall within the Park’s permitted list — software development and the related technology activities — and a business project that satisfies the Park that the operation is genuinely innovative and export-oriented, submitted through the Park’s application process. For a foreign group this is a step to plan for deliberately, and it can be approached from abroad; the detail of the application, the criteria, and how a foreign company works through it is the subject of our writing on HTP residency for a foreign IT company, and the wider HTP framework is set out through the Ministry of Economy and the consolidated legislation on ETALON-ONLINE. Setting up the underlying company is covered in our writing on establishing a subsidiary in Belarus.
Indicative — the figures and their scope depend on the structure and current law; confirm each with the firm before acting.
| Item | What applies | The catch |
|---|---|---|
| Profit tax | 0% | Replaced by 1% of gross revenue to the HTP administration |
| VAT on exported IT services | 0% | Import VAT as a tax agent can arise on services bought abroad |
| Dividend withholding to the parent | 5% | Reduced further by treaty, on the usual conditions |
| The intra-group service fee | Market price | Governed by transfer pricing — this is the discipline |
| Employee income tax | 13% | Foreign specialists included, with simplified permits |
Not literally, though it is light. An HTP resident is exempt from profit tax on its core activity but pays one per cent of gross revenue to the Park administration instead; its exported IT services fall outside VAT; dividends to a foreign parent carry five per cent withholding, treaty-reduced; and employees are taxed at a favourable rate. So it is a genuinely favourable regime rather than a zero one, and the picture is worth understanding with its qualifications rather than as a headline.
No — and it would work against you. Setting the fee below market is the exact pattern transfer-pricing rules catch: the price is adjusted back to market and tax and penalties follow, so the saving does not materialise. The saving you actually want comes from the favourable regime itself, applied to a market-level fee. You do not reduce tax by underpaying the subsidiary; you reduce it through the regime, with the fee set correctly.
At arm’s length — the price an independent provider would charge for the same work. For a captive development centre the standard method is cost-plus: the subsidiary’s costs plus a defensible margin reflecting a comparable independent return. The pricing must be documented and supportable, since the burden in a review is on the taxpayer, and for larger operations an advance pricing agreement with the tax authority can fix the approach in advance.
You need a real operation. HTP residency requires genuine activity and a business project with real export potential, not a shell, and substance — real engineers doing real work in Belarus — is what makes the structure withstand scrutiny. A billing entity with no substance is the weakest possible position if the arrangement is examined. The regime is built for genuine development centres and rewards them; it does not reward a nameplate.
In a service model, generally the parent. The work is done for hire, the deliverables are the parent’s, and the subsidiary is paid for performing the work rather than for owning it — which is one reason the service model is cleaner than holding IP in the subsidiary, since there is no IP transfer to price. If you later want the IP held in Belarus, that is a different structure with its own consequences, covered in our separate IP writing.
In transfer-pricing terms this is cleaner than servicing the parent, because a price agreed with an unrelated client is arm’s length by nature. But tax is not the whole question in 2026: whether Western clients will contract with a Belarusian entity, how payments would be routed, and what sanctions exposure the arrangement carries are real and material, and for many businesses they decide viability. It is worth assessing those realities honestly before building around direct client work.
A Belarusian legal entity whose activities fall within the Park’s permitted list — software development and related technology activities — and a business project satisfying the Park that the operation is genuinely innovative and export-oriented, submitted through the Park’s application process. It is a deliberate step rather than automatic on incorporation, and a foreign company can work through it from abroad. The detail is in our separate writing on HTP residency.
A Belarusian HTP subsidiary that services a foreign group is a genuine advantage, but it repays being understood correctly. The advantage is the regime — the profit-tax exemption for a one-per-cent revenue contribution, VAT-free exported services, light withholding and favourable employee tax — applied to an operation that is real and to a fee that sits at market. It is not, and cannot be made into, a saving engineered by underpricing the intra-group service, because that is the one move the rules reverse. Set the fee where an independent party’s would be, put genuine substance behind it, and the regime does the work it was designed to do.
Two things sit alongside that. In a service model the parent generally owns the intellectual property, which keeps the structure simpler than holding IP in the subsidiary and confines the transfer-pricing question to the service fee. And the question of servicing Western clients directly, cleaner though it is on tax, turns in 2026 on practical and legal realities that deserve an honest and specific assessment rather than an assumption. Get the pricing right, the substance real, and the client question answered clearly for your own circumstances, and you have a structure that is efficient because it is defensible — which is the only kind worth building. To work through the model for your group — the pricing method, the residency application, and whether direct client work is viable for you — contact our team.
A foreign company with a Belarusian supplier or customer lined up and a contract ready to sign usually has a quiet question in the background: how much can it really know about the party on the other side? The reassuring answer is: more than it expects. Belarus has a public company register, a counterparty’s details […]
Choosing the form for a Belarusian company, a foreign founder often feels a pull towards the joint-stock company. It sounds more substantial than a limited liability company — more like a real corporation — and there is a quiet assumption that the more formal vehicle must be the better one. It is worth resisting that […]
Registering a company in Belarus is quick — a matter of a day once the documents are in order — and foreign founders who have just been through it often expect the bank account to follow at the same pace. Then the account takes weeks, and the contrast is jarring enough to feel like something […]
