Which Digital Operations Belong in Singapore, and Which Do Not

Abstract

Singapore is widely treated by European technology and digital-economy firms as a single destination to be accepted or rejected as a whole. This chapter argues that the question is wrongly framed. Singapore is not one location but a set of distinct fits for distinct digital functions, and the same firm can correctly place its regional command, treasury, and governance functions in Singapore while placing its raw compute, its bulk software development, and its data hosting elsewhere in the region.

The chapter assesses nine categories of digital operation against three constraints that bind in Singapore with unusual force: the rationing of data-centre power capacity under the Infocomm Media Development Authority’s Green Data Centre Roadmap, the cost of resident and foreign technical labour under a rising Employment Pass salary floor, and the content-regulation regime administered under the Broadcasting Act. It sets these against the genuine and durable advantages that make Singapore a strong fit for headquarters, governance, and high-value research functions: a fifteen per cent minimum effective tax rate that has reshaped but not eliminated the incentive landscape, a deep and expensive specialist workforce, and a stable rule-of-law environment. The chapter closes with the recurring mistakes European firms make in reading Singapore as monolithic, and with a worked allocation for a representative European industrial-software group.

2.1 The Question Is Wrongly Framed

Most European firms approach Singapore as a yes-or-no decision. Should we be in Singapore, or not? The question feels natural because that is how site selection usually works for a factory or a warehouse: one site, one decision, one location that either fits or does not. A manufacturer choosing between a plot in Tuas and a plot in Johor is making a single choice about a single piece of ground, and the analysis that supports it is correspondingly singular.

Digital operations do not work this way, and treating them as though they do is the most expensive mistake in this book. A digital business is not a single thing that sits in a single place. It is a bundle of functions: strategic command, financial management, regulatory and legal oversight, customer-facing sales, deep research, applied engineering, routine software production, raw computation, and data storage. Each of these functions has its own cost structure, its own talent requirement, and its own regulatory exposure. Some of them fit Singapore extremely well. Some of them fit Singapore extremely badly. The firm that places all of them in Singapore because the headquarters belongs there will overpay heavily on the functions that do not. The firm that places none of them in Singapore because the data centre does not belong there will have given up a genuine and hard-to-replicate advantage on the functions that do.

The correct frame is allocation, not selection. The question is not whether the firm should be in Singapore. The question is which functions should be in Singapore, and which should sit in the markets around it. For a European digital firm building a presence in Southeast Asia, the realistic answer is almost always a split: a small, high-value, expensive base in Singapore, and a larger, cheaper, function-specific footprint in the markets nearby. Singapore is the apex of a regional structure, not the whole of it. The mental model that serves a European firm well is not “where do we put our Asia operation” but “what is the shape of our Asia operation, and which vertex of that shape sits in Singapore.”

This reframing matters because the two framings produce different errors. A firm that asks the yes-or-no question and answers yes tends to over-concentrate, building a Singapore operation that is too large, too expensive, and burdened with functions that would have been cheaper and just as effective elsewhere. A firm that asks the yes-or-no question and answers no tends to under-commit, running its whole region from a low-cost hub and missing the credibility, the regulatory alignment, and the access to capital and counsel that a Singapore apex provides. Both errors come from the same mistake: treating a bundle of functions as though it were a single object.

This chapter works through the allocation function by function. It is organised around three constraints that shape almost every placement decision, and it is honest about the cases where Singapore is the wrong answer. A reader who finishes this chapter should be able to take their own organisation chart, mark each box, and know roughly where it belongs and why.

2.2 The Three Constraints That Decide Almost Everything

Before assessing individual functions, it is worth setting out the three constraints that drive most of the answers. They recur throughout the chapter, and once a reader holds them in mind, most of the function-by-function conclusions follow almost mechanically. The constraints are electrical, financial, and regulatory: the rationing of power, the cost of people, and the supervision of content.

Power capacity is rationed

The first constraint is electrical, and it is the one European firms most often underestimate, because in most of Europe electrical capacity is a commodity that can be bought rather than a permission that must be won. Singapore is a small island with a single national grid, no domestic fossil-fuel resources to speak of, and limited room to build new generation. Data centres are the most power-hungry buildings in the modern economy, and Singapore has decided as a matter of policy that it cannot let them grow without limit.

Singapore is home to more than seventy data centres with a combined capacity of around 1.4 gigawatts.¹ The government imposed a moratorium on new data-centre development in 2019, lifted it in 2022, and replaced the open-door approach with a system of deliberate rationing.² In May 2024 the Infocomm Media Development Authority launched the Green Data Centre Roadmap, which aims to provide at least 300 megawatts of additional capacity in the near term, with the possibility of more for operators deploying green energy.³ That additional capacity is not allocated on a first-come basis. It is allocated through a qualitative process that prioritises both sustainability and economic value, and it carries binding efficiency conditions: over the next ten years, the Authority intends for all data centres in Singapore, including existing ones, to achieve a Power Usage Effectiveness of 1.3 or lower at full information-technology load.⁴

Two features of this regime are worth drawing out, because together they decide a great deal. The first is that capacity is finite and contested. The 300 megawatts of near-term additional capacity is a national figure shared across the whole industry, and the demand for compute from artificial intelligence, cloud services, and digitalisation comfortably exceeds it. A firm cannot assume that the capacity it wants will exist when it wants it. The second is that capacity is allocated on the state’s criteria, not the firm’s willingness to pay. An operator that can demonstrate high economic value and superior resource efficiency will be preferred over one that simply wants to host a large, undifferentiated workload. This is industrial policy expressed through electricity, and a firm that treats Singapore power as a market commodity has misread the regime.

The practical meaning for a European firm is blunt. You cannot assume you will be able to build or lease the raw compute capacity you want in Singapore at the moment you want it. Capacity is scarce, it is allocated by the state on the state’s criteria, and a firm whose business model depends on cheap, abundant, scalable computation should not plan to host that computation in Singapore. This single fact decides the fit of data hosting and large-scale model training before any other consideration enters.

Technical labour is deep but expensive

The second constraint is the cost of people, and here Singapore’s strength and its limitation are the same fact seen from two sides. Singapore has built a substantial technology workforce. In 2024 the country had around 214,000 tech professionals, up from 208,300 the year before, a growth of 2.7 per cent achieved against a backdrop of cautious global tech hiring and visible layoffs in major markets.⁵ The digital economy generated value added of S$128.1 billion in 2024, equal to 18.6 per cent of gross domestic product, up from 18.0 per cent the previous year.⁶ This is a serious, capable, and growing pool of talent, with particular depth in artificial intelligence, data, and cybersecurity.

It is also expensive, and the expense is structural rather than cyclical. The median monthly wage for resident tech professionals reached S$7,950 in 2024, up from S$7,000 in 2023, against an overall resident median wage of S$4,860.⁷ Tech labour in Singapore is paid roughly two-thirds more than the general workforce, and that premium reflects genuine scarcity of skilled people in a small population rather than a temporary bubble. A firm hiring local engineers in Singapore is hiring into one of the most expensive technical labour markets in Asia.

The cost of importing foreign technical staff is bounded from below by the Employment Pass salary floor. From 1 January 2025, the minimum qualifying salary for new Employment Pass applications rose to S$5,600 a month for general sectors, up from S$5,000, and to S$6,200 a month for the financial services sector, up from S$5,500. Both figures increase progressively with the applicant’s age, and the renewed thresholds apply to Employment Pass renewals from 1 January 2026.⁸ The effect is that a firm cannot use Singapore as a place to import junior or mid-level foreign engineers cheaply. The state has deliberately set the floor at a level pegged to the earnings of the top third of local professionals, managers, executives, and technicians, precisely so that foreign hiring complements rather than undercuts the resident workforce.⁹

The implication for allocation is symmetric. A function that needs a small number of senior, scarce, high-value people is buying exactly what Singapore sells well, and the salary premium is a reasonable price for genuine quality. A function that depends on a large volume of routine, cost-sensitive engineering labour is buying exactly what Singapore sells badly, and the premium is simply waste. The labour constraint does not say “avoid Singapore.” It says “put the senior people here and the volume people elsewhere.”

Content and platform activity is regulated with real penalties

The third constraint is regulatory, and unlike the first two it bites selectively. It bears heavily on consumer-facing content and platform businesses and hardly at all on enterprise, industrial, or business-to-business digital activity. A European firm should understand which side of this line its operation sits on, because the answer changes the analysis completely.

Under Part 10A of the Broadcasting Act 1994, introduced with effect from 1 February 2023, the Infocomm Media Development Authority may designate online communication services and impose codes of practice on them.¹⁰ The Code of Practice for Online Safety for social media services took effect on 18 July 2023 and requires designated services to implement content-moderation systems, user-reporting and resolution mechanisms, enhanced protections for users under eighteen, and the submission of annual online-safety reports for publication.¹¹ A second code, for app distribution services, took effect on 31 March 2025 and extends comparable system-level obligations to the marketplaces through which apps reach Singapore users.¹²

The penalties are not nominal. A designated service that fails its duty under the Broadcasting Act may be ordered to pay a financial penalty of up to S$1 million, and a service that fails to comply with a remedial direction commits an offence carrying a fine of up to S$1 million and, for a continuing offence, a further fine of up to S$100,000 for every day the offence continues after conviction.¹³ These are amounts that show up on a profit-and-loss statement, and the obligations behind them require dedicated staff, systems, and reporting.

For a European firm running an enterprise software business, a logistics platform, or an industrial analytics product, these provisions are largely irrelevant, because the firm does not operate a designated social media or app-distribution service. For a firm running a consumer social platform, a user-generated-content service, or a content-moderation operation, they are a material and ongoing compliance burden that a lighter-touch jurisdiction would not impose. The constraint is real, but it is specific, and a firm should not let a regime aimed at consumer platforms frighten it away from enterprise activity that the regime does not touch.

With those three constraints in hand, the function-by-function assessment is largely a matter of asking, for each function, which constraints apply.

2.3 Strong Fit: Headquarters, Treasury, and Command Functions

Singapore is a strong fit for the apex functions of a regional digital business: strategic management, regional command, and corporate treasury. The reasons are institutional rather than operational. Singapore offers a stable legal environment, a well-regarded judiciary, an extensive network of double-taxation agreements, and a concentration of professional services in law, accounting, banking, and advisory work that a regional headquarters draws on constantly. None of these advantages is about the cost of doing a thing; all of them are about the reliability and quality of the environment in which decisions are made and capital is moved.

The fiscal incentives that historically anchored this case still exist but have changed character. The Economic Development Board administers concessionary-tax-rate incentives through the Development and Expansion Incentive, which following Budget 2024 offers tiers of 5 per cent, 10 per cent, and a new 15 per cent rate against the standard corporate rate of 17 per cent, in each case on qualifying incremental income and subject to substantive activity and expenditure conditions.¹⁴ These incentives once delivered headline tax savings well below the prevailing rate, and for many firms they were the decisive factor in choosing Singapore over alternatives. For the largest firms, that is no longer the whole story, and the reason is the subject of the next section.

But the headquarters case does not rest on the tax rate, and a firm that lets the tax change unsettle the whole decision has misunderstood what it is buying. It rests on the concentration of decision-making, capital, and professional infrastructure that makes Singapore a sensible place to run a region from. When a regional managing director needs to restructure a joint venture, raise local financing, resolve a cross-border dispute, or move treasury balances across a dozen currencies, the value of being in Singapore is the depth and reliability of the people and institutions that do those things. That case is intact and is not vulnerable to a change in the concessionary rate.

The honest qualification is cost. A regional headquarters in Singapore is expensive to staff, because the people who run it are senior and senior people are paid Singapore salaries on top of Singapore housing and Singapore schooling costs. A firm should put its decision-makers and its treasury function in Singapore because that is where they are most effective, not because it is cheap. It is not cheap, and a firm that expects it to be cheap will be unhappy. The right disposition is to accept the cost of the apex deliberately, while refusing to let that cost spread to functions that do not need to sit at the apex.

2.4 The Pillar Two Qualification on the Headquarters Case

The strong-fit conclusion for headquarters carries an important qualification for the largest European firms, and it is worth setting out carefully because it is widely misunderstood and frequently overstated in both directions.

Singapore has implemented the OECD’s BEPS 2.0 Pillar Two framework. Under the Multinational Enterprise (Minimum Tax) Act 2024, two top-up taxes apply for financial years beginning on or after 1 January 2025. The first is a Multinational Enterprise Top-up Tax, corresponding to the Income Inclusion Rule under the Global Anti-Base Erosion model rules, which applies to the low-taxed profits of group entities located outside Singapore. The second is a Domestic Top-up Tax, which applies to the low-taxed profits of group entities located in Singapore and allows Singapore to collect any top-up arising domestically rather than ceding it to another jurisdiction.¹⁵ Both are designed to ensure that in-scope multinational groups pay a minimum effective tax rate of 15 per cent on a jurisdictional basis.¹⁶ The rules apply to groups with annual consolidated revenue of EUR 750 million or more in at least two of the four preceding financial years.¹⁷

The consequence for a large European group is that the old logic of headquarters incentives no longer delivers what it once did. A concessionary rate of 5 or 10 per cent on qualifying income, for an in-scope group, will simply trigger a domestic top-up that brings the effective rate back to 15 per cent. The savings below 15 per cent that once made these incentives compelling are clawed back. This is the single most important fiscal fact for a large European group evaluating Singapore, and a firm that builds its business case on a concessionary rate it will not actually enjoy has built it on sand.

It is equally important not to overstate the point. This does not make Singapore a worse headquarters location than its peers. Pillar Two applies across all participating jurisdictions, so the 15 per cent floor is a feature of the international landscape rather than a Singapore peculiarity. A group that responds to the Singapore top-up by relocating its headquarters to another participating jurisdiction will generally find the same floor waiting for it there. The correct conclusion is not that Singapore has become uncompetitive but that the headquarters decision must now be justified on the non-tax merits set out in the previous section, which remain strong. The non-tax case for Singapore is strong; the tax case has been levelled rather than destroyed.

Two further points soften the picture. First, Singapore’s response to Pillar Two has included instruments designed to be compatible with the new rules. The Refundable Investment Credit introduced in Budget 2024 is structured as a qualifying refundable tax credit, which under the Global Anti-Base Erosion rules is treated more favourably than a tax-rate concession because, unlike a deduction or a rate reduction, it does not reduce the group’s effective tax rate for the purpose of those rules.¹⁸ A firm planning substantial qualifying investment should examine the credit with its advisers rather than assuming the incentive landscape is now barren. Second, the threshold matters in both directions. A European firm below EUR 750 million in consolidated revenue is not in scope at all, and for that firm the older concessionary-rate logic still operates with full force. The Pillar Two qualification is a qualification on the headquarters case for large groups, not a general verdict on Singapore.

2.5 Strong Fit: High-Value Research and AI Governance

Deep technical research, system-level engineering, and the governance of artificial-intelligence systems are a strong fit for Singapore. The fit follows directly from the labour constraint described above, read from its favourable side. A deep, specialised, expensive workforce is exactly what high-value research needs and exactly what routine production does not. A firm staffing a small team of senior researchers and architects is buying precisely the thing Singapore sells well, and the salary premium that makes Singapore a poor home for bulk engineering is a reasonable price for a handful of senior specialists whose quality and scarcity are the whole point.

Singapore has also positioned itself deliberately as a centre for the governance and assurance of artificial intelligence rather than for its raw training. The Authority’s Model AI Governance Framework and its associated testing and assurance initiatives reflect a national strategy of becoming the place where artificial intelligence is made trustworthy and accountable, rather than the place where the largest models are trained on the cheapest power. For a European firm this alignment is unusually valuable. European firms operate under a regulatory tradition that takes data protection and algorithmic accountability seriously, and they face the European Union’s own AI Act. A Singapore base for assurance, safety research, and governance work lets a firm align its regulatory posture across two demanding jurisdictions at once, and lets it build governance capability in a place where the surrounding institutions take the subject seriously.

The qualification here is one of scale, not of fit. Government co-funding for artificial-intelligence work exists, but it is modest relative to the scale of a large multinational’s overall programme, and the relevant expenditure caps are calibrated for the domestic economy rather than for a global firm’s transformation budget. A European firm should therefore treat Singapore’s governance positioning as a strategic and reputational fit, supported at the margin by public funding, rather than as a place where the bulk of an artificial-intelligence budget will be subsidised. The reason to do governance and high-value research in Singapore is that the work is better and better-aligned there, not that it is materially cheaper.

2.6 Split Fit: Artificial Intelligence and Machine Learning

Artificial-intelligence operations do not have a single fit. They split cleanly along the line drawn by the power constraint, and a firm that understands the line can place each part of its artificial-intelligence stack correctly.

Governance, safety research, and system architecture sit at the high-value, low-power end of the stack, and they are a strong fit for the reasons given in the previous section. This is design and oversight work done by small numbers of senior people, and it draws on Singapore’s strengths without running into its power limits.

Artificial-intelligence inference and edge computing occupy a middle position and are an acceptable fit with a caveat. Inference, where trained models serve low-latency predictions to users or systems in the region, benefits from being close to those users, which argues for some inference capacity in Singapore. But inference at scale still draws meaningful power, and the energy cost together with the Power Usage Effectiveness conditions limits how much of it can sensibly sit on the island. The realistic pattern is therefore hybrid: latency-sensitive edge servers in Singapore, sized to what genuinely needs to be local, with heavier batch processing pushed downstream to markets where power is cheaper and capacity is available.

Large-scale model training is a poor fit, and the reason is the power constraint applied at its sharpest. Training modern models requires dense racks of accelerators drawing enormous and sustained power and demanding intensive cooling. Singapore’s capacity rationing, its energy cost, and its efficiency conditions make housing that infrastructure locally unviable for any firm whose training workload is large enough to matter. This is the case where the allocation discipline of this chapter matters most. A European firm that has correctly decided to put its artificial-intelligence governance and its regional command in Singapore must resist the intuition that the training cluster should follow them there. The headquarters belongs in Singapore. The governance team belongs in Singapore. The training accelerators belong wherever power is abundant and cheap, which is not Singapore.

2.7 Acceptable Fit with Caveats: Enterprise Software and SaaS

Business-to-business software and software-as-a-service occupy a middle position, and the right structure is conditional rather than absolute. The two halves of such a business, the commercial half and the engineering half, have opposite fits, and a firm that recognises this can structure itself well while a firm that treats the business as a single unit will mis-site one half or the other.

The commercial functions of an enterprise software business are a good fit for Singapore. Regional sales, customer success, account management, partnership development, and solution consulting all benefit from the same headquarters logic that makes the apex functions fit: proximity to regional customers, a pro-business environment, access to decision-makers and capital, and the credibility that a Singapore base confers when selling to large regional enterprises. For a firm whose customers are major Southeast Asian banks, manufacturers, or logistics operators, having the people who manage those relationships based in Singapore is a real advantage.

The underlying product engineering is a different matter, and here the labour constraint applies in full. Housing the firm’s product managers and software engineers in Singapore is cost-inefficient for the reasons set out in section 2.2: the firm will pay the resident tech median or import staff above the Employment Pass floor, and in either case it will pay far more than the same engineering capability costs in lower-wage regional markets. The viable structure therefore places a small commercial command centre in Singapore and builds the backend engineering team in a lower-cost market, managed remotely. The Singapore office is where the firm sells, contracts, and manages the region. It is not where the firm should write the bulk of its code. A firm that insists on co-locating its engineers with its salespeople in Singapore, out of a preference for everyone being in one building, is paying a substantial and avoidable premium for the comfort of co-location.

2.8 Poor Fit: Data Hosting, Bulk Engineering, and Consumer Platforms

Three categories of digital operation are poor fits, and instructively each is defeated by a different one of the three constraints. Reading them together makes the logic of the whole chapter explicit.

Hyperscale data hosting and raw computation are defeated by the power constraint. The capacity rationing, the energy cost, and the efficiency conditions described in section 2.2 make Singapore an unsuitable home for the bulk of a firm’s data hosting and heavy computation. These workloads are precisely the undifferentiated, power-hungry activity that the Authority’s allocation regime is designed to keep within limits, and a firm whose plan depends on hosting them cheaply and at scale in Singapore is planning against the grain of national policy. They belong in the markets immediately around Singapore, where land is available and power is cheaper, with only the minimum necessary capacity retained on the island.

Engineering and development centres run as a cost play are defeated by the labour constraint. A firm that houses large downstream software development, quality-assurance teams, or front-end application developers in Singapore to save money has misread the market. It will pay the resident tech median of S$7,950 a month, or import staff above the Employment Pass floor, and in either case it will pay far more than the same function costs in lower-wage markets.¹⁹ The point is not that this work cannot be done well in Singapore; it can. The point is that paying Singapore rates for work whose defining requirement is that it be inexpensive is a category error. Singapore is the wrong place to do work whose primary virtue is cheapness.

Consumer-facing content and platform businesses are defeated by the regulatory constraint. The online-safety regime under the Broadcasting Act imposes content-moderation and reporting obligations on designated services and exposes them to penalties of up to S$1 million, with daily continuing-offence fines.²⁰ A firm whose core operation is a consumer social platform or a content-moderation hub takes on a compliance burden in Singapore that a lighter-touch jurisdiction would not impose, with little offsetting advantage for that particular function. Unlike the other two poor fits, this one is about exposure rather than cost, but the conclusion is the same: place this function where the burden is lighter.

In each of these three cases the conclusion is not that the firm should avoid Singapore. It is that the firm should place this particular function elsewhere in the region, while keeping in Singapore the functions for which Singapore is the right answer. The poor fits and the strong fits are two halves of the same allocation, and the firm that gets both halves right is the firm that has understood the chapter.

2.9 Worked Allocation: A European Industrial-Software Group

Consider a representative case, built to make the allocation concrete. A European industrial-software group, take a German firm building control and analytics software for manufacturing equipment, with consolidated revenue below the EUR 750 million Pillar Two threshold, wants to establish a Southeast Asian presence. It has perhaps fifty people to deploy in the region over three years, a product that runs partly in the cloud and partly at the edge on customer equipment, and a customer base of large regional manufacturers. Where does each function go?

The regional headquarters goes to Singapore. The managing director for Asia-Pacific, the regional finance lead, and the treasury function that manages the region’s cash and currency exposure all sit there. Because the firm is below the Pillar Two threshold, the Development and Expansion Incentive’s concessionary rates remain meaningfully available, subject to the activity and expenditure conditions, and the firm should take qualified tax advice on whether it can realistically meet them given its planned headcount and spend.²¹ If it cannot meet the conditions, the headquarters still belongs in Singapore on the non-tax merits.

The regional sales and customer-success team goes to Singapore, co-located with the headquarters. The firm sells to large regional manufacturers, the relationships are senior and consultative, and they are best managed from the apex where the decision-makers and the firm’s regional credibility sit.

A small senior research and governance team goes to Singapore. Three or four architects and a governance specialist work on the firm’s artificial-intelligence assurance and on aligning the product with both European and Singapore regulatory expectations. This is exactly the senior, scarce, high-value work that Singapore’s expensive talent pool supplies well, and the salary premium is worth paying for a handful of people of this calibre.

The bulk software engineering does not go to Singapore. The twenty or thirty developers who write and maintain the product sit in a lower-cost regional market and are managed remotely from the Singapore command centre. The firm would pay roughly double in Singapore for the same engineering capability, and the work does not need the apex.

The data hosting and any heavy computation do not go to Singapore either. They sit in a market with available power and lower energy cost in the region around Singapore. Only the latency-sensitive edge capacity that genuinely needs to be local is retained on the island, sized to what the product actually requires rather than to a preference for keeping everything in one place.

The result is a firm that is genuinely in Singapore in every sense that matters. Its decision-makers, its money, its customer relationships, and its senior research all sit there, and a regional customer or regulator encountering the firm meets it as a Singapore operation. At the same time the majority of its headcount and almost all of its compute sit elsewhere, where they are cheaper and unconstrained. That is the correct shape. A firm that put all fifty people and all its compute in Singapore would overpay enormously and gain nothing for the overpayment. A firm that put none of it in Singapore, running the whole region from a low-cost hub, would have saved money and given up the apex that makes the operation credible and well-governed. The allocation, not the selection, is the decision that matters, and the worked case shows why the answer is almost always a deliberate split.

2.10 What Is Changing Between Now and 2027

Several of the constraints in this chapter sit on announced trajectories rather than at fixed values, and a firm planning a multi-year build should hold the directions of travel in mind rather than treating the present snapshot as permanent.

The Employment Pass salary floor is reviewed annually against the earnings of the top third of local professionals, so the S$5,600 general-sector and S$6,200 financial-services thresholds effective from 1 January 2025 should be expected to rise again over the planning horizon, with the renewed thresholds applying to renewals from 1 January 2026.²² A firm modelling the cost of a Singapore technical team should assume the floor moves upward rather than holding, and should size its Singapore headcount on the assumption that importing technical staff will become more expensive, not less.

The data-centre capacity position will evolve as the Authority allocates the additional capacity promised under the Green Data Centre Roadmap, with the qualitative allocation process and the Power Usage Effectiveness conditions described in section 2.2 governing who receives it.²³ A firm that needs Singapore compute capacity should treat allocation as a process to be entered early and on the state’s sustainability and economic-value criteria, rather than as capacity available on demand at the moment of need. The direction of travel is toward more capacity but also toward stricter efficiency conditions, and a firm planning a compute footprint should plan for both.

The Pillar Two regime is in its first years of operation. The Multinational Enterprise (Minimum Tax) Act 2024 took effect for financial years beginning on or after 1 January 2025, and the administrative and registration machinery, including the registration process and the related returns, is being built out through 2026.²⁴ A firm near the EUR 750 million threshold should watch carefully whether it crosses into scope over its planning horizon, because crossing the threshold changes the headquarters tax calculus set out in section 2.4 from the older concessionary-rate logic to the levelled fifteen per cent floor.

The general lesson is the one stated in the final mistake below. Singapore governs by announced multi-year trajectories, and the figures in this chapter are points on those trajectories rather than fixed constants. A firm that verifies the current values and effective dates with its advisers before acting will avoid being wrong-footed by changes that were entirely foreseeable.

2.11 Eight Mistakes European Firms Make

These are the recurring errors in reading Singapore’s fit for digital operations. A reader who internalises only this section still leaves with the chapter’s practical core.

Treating Singapore as one decision. The most expensive error is the one the chapter opens with: deciding whether to “be in Singapore” rather than deciding which functions belong there. The answer is almost always a split, and a firm that forces a single yes-or-no answer will get the placement of half its functions wrong, either over-concentrating in Singapore or giving up the apex entirely.

Putting the compute where the headquarters is. Firms that correctly site their regional command in Singapore often drag their data hosting and training infrastructure along with it, defeated by an intuition that everything should be co-located. The power constraint makes this a costly mistake. The accelerators belong in a different country from the executives, and keeping them together buys nothing but expense and a fight with the allocation regime.

Justifying a large-firm headquarters on the tax rate. For an in-scope group above the EUR 750 million threshold, the concessionary rates that once anchored the headquarters case are now topped up to fifteen per cent. A firm that builds its business case on a 5 or 10 per cent rate it will not actually enjoy has built it on sand. The non-tax case is the real case, and it is strong enough to carry the decision on its own.

Assuming data-centre capacity is available on demand. Capacity is rationed and allocated by the state on qualitative criteria of sustainability and economic value. A firm that assumes it can lease the compute it wants when it wants it has misunderstood the regime and may find its build delayed or blocked at the point when it matters most.

Using Singapore as a low-cost engineering base. The resident tech median of S$7,950 a month and the Employment Pass floor make Singapore one of the most expensive places in the region to house routine engineering. A firm that locates bulk development there to save money has chosen the wrong market for that function and will pay roughly double for no corresponding benefit.

Underestimating the content-regulation burden for platforms. A firm running a consumer platform or a content-moderation operation takes on real obligations and real penalty exposure under the Broadcasting Act, with fines reaching S$1 million and daily continuing-offence penalties. Firms that model Singapore as uniformly light-touch are surprised by the online-safety regime, while firms running enterprise software sometimes worry about a regime that does not apply to them at all.

Ignoring the EUR 750 million threshold in both directions. A firm below the threshold should not assume the Pillar Two analysis applies to it, and a firm approaching the threshold should not assume its current incentive position will survive crossing it. The threshold is the hinge of the headquarters tax question, and many firms do not know which side of it they will be on across their planning horizon.

Treating the present numbers as fixed. The salary floor, the data-centre capacity position, and the tax regime all sit on announced multi-year trajectories. A firm that models the current snapshot as permanent will be wrong-footed by changes that were announced years in advance and were entirely foreseeable. Verify the current figures and effective dates before acting.

References

Declarations

Competing interests: The author is a licensed real estate agent (Council for Estate Agencies, Singapore) affiliated with OrangeTee & Tie Pte Ltd, and a Singapore Mediation Centre-accredited mediator. The author has commercial interests in industrial and commercial real estate transactions facilitated through OrangeTee & Tie. These interests are openly disclosed. The analysis in this chapter has been written to be useful to the reader irrespective of whether the reader subsequently engages the author’s transactional services.

Funding: This work received no external funding.

Methodology: This chapter assesses the fit of nine categories of digital operation against Singapore’s data-centre, labour, and content-regulation regimes. Each substantive factual claim has been verified against a primary source at the time of writing: IMDA factsheets and the Green Data Centre Roadmap for data-centre capacity and efficiency conditions; the IMDA Singapore Digital Economy Report 2025 for workforce and digital-economy figures; the Inland Revenue Authority of Singapore and the Ministry of Finance for the BEPS 2.0 Pillar Two regime; the Economic Development Board and EY technical alerts for the Development and Expansion Incentive and Refundable Investment Credit; the Ministry of Manpower and Economic Development Board for Employment Pass thresholds; and the Broadcasting Act 1994 with IMDA codes of practice for the online-safety regime. Where a figure could not be verified against a primary or reliable secondary source, it has been omitted rather than asserted. This chapter is a detailed overview for orientation; it is not professional tax, legal, immigration, or investment advice, and readers should obtain qualified advice before acting.

Currency of analysis: The analysis is current as of the date of publication. Singapore’s Employment Pass salary floors, data-centre capacity allocation, and Pillar Two administrative rules all sit on announced multi-year trajectories and will change; readers should verify current figures and effective dates before relying on them.

About the Author

David Hoicka is a Singapore-licensed real estate agent (Council for Estate Agencies) affiliated with OrangeTee & Tie Pte Ltd, with a specialisation in industrial and commercial property for European inbound investment. He is also a Singapore Mediation Centre-accredited mediator, a civil engineer (Bachelor of Science, Massachusetts Institute of Technology), and the founder and publisher of Singapore Mediation Solutions, an academic publisher registered with Crossref (DOI prefix 10.66404) and with the National Library Board of Singapore. He has lived in Singapore as a permanent resident for over twenty-one years.

Scholarly identifiers: ORCiD 0000-0001-9082-0720; Wikidata Q137455251; ISNI 0000 0005 2886 676X; Google Scholar profile available.

About the Publisher

Singapore Mediation Solutions is an open-access scholarly publisher specialising in practical and analytical works for cross-border commercial practitioners with a focus on Asia-Europe industrial and commercial relations. Singapore Mediation Solutions is registered with Crossref (DOI prefix 10.66404), is a Singapore publisher with NLB-assigned ISBNs, and deposits all works in Zenodo for permanent open-access availability and in OCLC WorldCat for library catalogue accessibility.

Confidential Consultation

Readers who would like to discuss the placement of digital operations between Singapore and the markets around it in confidence may contact the author directly. The preferred channels are Signal and Telegram for confidentiality and ease of cross-border communication. Direct email is also available. Contact details are listed on datascienceai.org. Initial consultations are conducted without obligation; the author’s role as principal advisor and the relationship to OrangeTee & Tie transactional execution are set out in a written engagement letter before any onward referrals are made.

---

Chapter DOI: 10.66404/ds.b5.ch2 (to be assigned upon Crossref deposit) Zenodo deposit: pending Published by Singapore Mediation Solutions, Singapore Open access under Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0)