Where Digital Operations Go

8.1 Digital Operations Occupy Several Different Property Markets

A European industrialist who has read this far will be used to thinking of a Singapore property decision as a single question: where does the operation go, and on what terms. For a manufacturer, that instinct is roughly right. A factory is a factory, and the choice is mostly between estates and tenures within one market.

For a digital operation it is wrong, and the error is expensive. A digital firm’s property need is not one market but several, and which one applies depends entirely on what the operation actually does. A regional headquarters that books revenue and houses a sales and management team needs office space, and the choice there runs from the central business district through the business parks to serviced suites. An applied-research or product-engineering team that wants to sit near Singapore’s public research institutes needs the one-north cluster, which is a different market with a different landlord and a different logic. A firm that must run its own compute at scale needs data-centre capacity, which is the most specialised property market in Singapore and the one least like anything the firm has leased in Europe. A small first-presence team testing the region needs flexible space it can walk away from.

These are not points on a spectrum. They are separate markets with separate operators, separate lease structures, separate cost bases, and separate constraints. Take a firm that treats them as one market and signs a long central-business-district headquarters lease when what it needed was a serviced office and a colocation contract. That firm has not overpaid by a margin. It has bought the wrong thing.

The operations this chapter places against property were set out in Chapter 2. This chapter takes each property sub-market in turn, describes what it is and is not suited to, and ends with the lease mechanics and a worked decision. The author writes here closer to the ground than in the institutional chapters, because property is the author’s licensed practice, and because the difference between a good and a bad Singapore property decision is made in operational specifics that no policy document records.

The order is deliberate. The chapter begins with data-centre real estate, the most specialised and most constrained market, because it is the one where a European firm’s assumptions transfer least well and where a mistake is hardest to unwind. It then moves to the purpose-built Punggol Digital District, to the mainstream office and business-park choice, to the one-north research cluster, and to the smaller-format space a first-presence team should usually take. It closes with the mechanics common to all of them: the lease, the fit-out, and the reinstatement obligation that European firms under-budget more reliably than any other cost.

8.2 Data-Centre Real Estate as a Distinct Sub-Market

Data-centre real estate in Singapore is its own institution, and it is worth a European operator’s time to understand why before signing anything, because almost nothing the operator knows about commercial leasing transfers cleanly.

The landlords are, in large part, listed real-estate investment trusts and their sponsors, alongside specialist developers. The operators who run the facilities, meaning the firms a tenant actually contracts with for space and power, are a mix of global colocation specialists, regional players, and the local telcos. The international names operating in Singapore include the large global colocation and hyperscale-adjacent operators; the regional and local names include the telco-affiliated operators. A European firm rarely leases a building. It contracts for capacity, measured in kilowatts or megawatts of IT load rather than in square feet, inside a facility someone else owns and runs.

That distinction drives everything else. A European firm used to thinking of property as floor area at a rent per square foot has to retrain its instinct entirely, because in this market the unit of account is power and the thing being bought is a service. Lease structures run from fully-fitted single-tenant arrangements, through colocation, to powered-shell or shell-and-core models where the tenant takes raw space and power and builds the fit-out itself. The choice among them is a choice about how much of the build risk and the operating control the firm wants to hold. A fully-fitted single-tenant arrangement hands the firm a turnkey facility and the price that goes with turnkey. Colocation puts the firm’s equipment into a shared, professionally run facility and is the right answer for most operations that are not at hyperscale. The powered shell suits a firm large enough and specialised enough to want to build to its own specification, and it carries the build cost and the build risk to match.

Pricing follows the same logic and is quoted on power, not floor area. The common structures are a fixed monthly charge for a reserved block of kilowatts plus operating expenses, an all-in rate that bundles the components, and per-rack billing for smaller colocation footprints. Power is the largest single cost component in running a data centre, and the commercial structures reflect that. In colocation, electricity is typically a direct pass-through to the customer, so the firm carries the energy cost and the energy risk directly rather than seeing it folded into a flat rent. A master-lease tenant taking a whole facility usually contracts directly with the power supplier. For a European firm this is the first place its budgeting instinct misleads it: the headline space rate is not the cost of the operation, and a firm that compares facilities on the space rate alone is comparing the wrong number.

The other cost that dominates is cooling, and cooling is now a regulated matter, not merely an engineering one. Singapore’s data-centre framework pushes operators away from traditional air cooling toward higher-density approaches, including liquid and direct-to-chip cooling, both because the heat density of modern AI workloads demands it and because the state’s efficiency standards require it. The practical implication for a European operator is that the facility it contracts with is operating under efficiency obligations the operator must meet, and that those obligations shape what the operator can offer and at what price. A firm cannot contract around the efficiency regime by asking for a less efficient, cheaper facility, because the less efficient facility is increasingly not permitted to exist.

The single most important structural fact about this sub-market, though, is not in any lease clause. It is that supply is deliberately constrained by the state. That is the subject of the next section, and it is the reason a European operator’s data-centre question in Singapore is rarely “what does it cost” and almost always “can I get it at all”.

8.3 The Capacity Constraint as a Commercial Reality

Singapore manages data-centre capacity as a matter of national policy, and it does so openly. After a period from 2019 during which new capacity was effectively paused, the Infocomm Media Development Authority and the Economic Development Board reopened growth through a managed mechanism rather than an open market. The instrument is the Data Centre Call for Application, a competitive process in which operators bid for blocks of capacity against published criteria.¹

The numbers give the shape of it. The pilot Call for Application in 2023 provisionally awarded roughly 80 megawatts of new capacity to four operators, namely AirTrunk together with ByteDance, Equinix, GDS, and Microsoft, selected for proposals judged to strengthen Singapore’s position as a regional hub.² In May 2024 the Authority launched a Green Data Centre Roadmap, which set out the intention to add at least 300 megawatts of additional capacity in the near term, with more available where green energy is used.³ In December 2025 the agencies launched a second Call for Application, opening at least a further 200 megawatts, with the application window closing on 31 March 2026.⁴

What a European operator must understand is the condition attached to that capacity. Allocation is not awarded to the highest bidder or the fastest mover. It is awarded against sustainability and economic criteria that are, by the agencies’ own framing, among the most demanding in the region. Under the second Call for Application, an applicant must commit to obtaining the BCA-IMDA Green Mark for Data Centres 2024 Platinum certification, achieve a Power Usage Effectiveness of 1.25 at full IT load, a tighter figure than the 1.3 target of the pilot, meet equipment-efficiency standards set out in the Singapore Standard SS 715:2025, and source at least 50 per cent of power from eligible green-energy pathways such as low-carbon hydrogen, low-carbon ammonia, biomethane, or building-integrated photovoltaics.⁵⁶

It is worth being honest about how this constraint frames itself versus how it lands. The state’s framing is stewardship: Singapore is small, power and land are finite, and the policy directs scarce capacity toward operators who will use it efficiently and cleanly. That framing is genuine, and the resulting standards are real engineering commitments, not box-ticking. The commercial effect, however, is a market that is structurally short of supply. Through 2024 and 2025 Singapore’s data-centre vacancy rate sat at roughly one to two per cent, among the lowest, and by several measures the lowest, of any tracked market globally.⁷⁸ A vacancy rate that low is not a tight market. It is a market in which available capacity is, for practical purposes, spoken for.

The consequence for a European firm is direct. If the operation can fit inside colocation space at an established operator, the question is one of price and timing, and both are difficult but soluble. If the operation needs a large dedicated footprint, whether its own facility or a substantial reserved block, then the firm is, in effect, dependent on capacity that flows through the Call for Application process, and that is a multi-year proposition contingent on meeting the sustainability bar. The author’s direct experience of how long, and at what cost, a new operator takes to secure a large dedicated allocation is practitioner knowledge rather than a published figure, and a reader weighing a large commitment should treat the timeline as a first-order risk to be scoped with the operators directly, not as a detail to be settled later. The honest framing for a European board is that a large dedicated Singapore data-centre footprint is not something to be procured on a project schedule. It is something to be pursued, contingently, over a horizon measured in years, and a business plan that assumes otherwise has mis-stated its own critical path.

There is a practical corollary that European firms routinely miss. A firm that needs capacity in Singapore but cannot wait for, or does not qualify for, a dedicated allocation is not therefore shut out. The route in is colocation at an operator who already holds capacity, and the negotiation there is not over whether capacity exists but over price, over the reserved power block, and over the service levels. For most European operations short of true hyperscale, this is the realistic path, and the firm’s energy is better spent securing good colocation terms from an established operator than pursuing a dedicated allocation it is unlikely to win on a timeline it can use. The dedicated-build route is for operators large enough that the multi-year pursuit, the sustainability commitments, and the capital are all justified by the scale of the compute. A European firm should be honest with itself about which category it is in before it commits management time to a process designed for the other.

This is the reason a growing share of regional compute is being built not in Singapore but across the strait in Johor. Cushman & Wakefield’s market analysis ranks Johor as the second most mature data-centre market in the Asia-Pacific, behind only Sydney, on the strength of very low vacancy and the highest individual build capacity of any tracked market, and Johor’s rise is explicitly driven by operators capitalising on proximity to Singapore alongside more readily available land, power, and a lower overall cost of occupancy.⁹ Singapore itself is among the most expensive colocation markets in the region, a direct function of its land constraints and managed supply.¹⁰

The practitioner’s reading of this is not that Singapore has lost the data-centre argument. It is that the argument has split. Latency-sensitive, connectivity-critical, and trust-sensitive workloads, meaning the ones that need to be physically in Singapore for legal, commercial, or network reasons, still command Singapore capacity and pay for it. Bulk compute that does not need a Singapore address increasingly does not get one, and goes to Johor or to other regional markets offering faster build timelines and lower cost. A European operator’s first data-centre decision is to work out honestly which of those two its workload is, because the answer determines whether the Singapore capacity question is worth fighting at all. A firm that fights for scarce, expensive Singapore capacity to house compute that would have run perfectly well in Johor has spent its scarcest resource, which is management attention and time, on a constraint it did not need to accept.

8.4 The Punggol Digital District

Singapore’s most visible recent statement of intent in digital property is the Punggol Digital District, in the island’s north-east. It is a 50-hectare district master-planned and developed by JTC Corporation, the industrial-estate landlord, and it is positioned as the country’s first purpose-built smart district for digital and technology business.¹¹

The substance behind the positioning is real. The district co-locates a JTC business park with the campus of the Singapore Institute of Technology, deliberately placing a university with an applied, industry-facing orientation next to commercial workspace so that students and firms sit in the same precinct.¹² Its declared target sectors are the digital ones, namely cybersecurity, artificial intelligence, robotics, and financial technology, and its early anchor occupiers include the innovation and technology centres of established Singapore institutions, among them an OCBC innovation hub and a UOB technology and innovation centre.¹³ It earned the Building and Construction Authority’s Green Mark Platinum award for districts, the highest sustainability rating at district scale, and at 50 hectares it is the largest mixed-use development to hold that rating in Singapore.¹⁴ When fully built it is projected to support on the order of 28,000 jobs.¹⁵ Tenants began moving in from 2024 and 2025.¹⁶

This is not marketing dressed as planning. The author’s view is that the deliberate design deserves the acknowledgement. The co-location of university and industry is a genuine attempt to engineer the kind of academia-to-firm spillover that happens organically in older clusters and rarely happens at all in greenfield ones, and the district is built to a real environmental standard rather than a nominal one.

The honest part of the assessment is what the district is and is not suited to. The Punggol Digital District suits a firm whose value depends on the specific things the district concentrates: proximity to a pipeline of applied-technology graduates, a working relationship with the university, and a sector fit with the cybersecurity, AI, robotics, and fintech cluster the district is curating. A robotics scaleup recruiting engineers, or a cybersecurity operation that wants a deliberate Smart Nation address, will find the district’s logic sound.

For many European firms it is oversold relative to actual need. A regional headquarters whose work is commercial and managerial gains little from sitting next to an applied-technology campus and loses the central location that a headquarters function usually wants. Punggol is in the far north-east, well away from the central business district and the financial and professional services a headquarters draws on daily. The talent-spillover argument that justifies the district is real only for operations that actually consume that talent and that collaboration; for an operation that does not, the district is simply a business park a long way from town with a sustainability rating the firm is paying for in rent. The district is a strong answer to a specific question. A European operator should be sure it is asking that question before treating Punggol as the default digital address.

8.5 Technology Office and Business-Park Space

For most European digital operations the realistic property choice is not data-centre capacity and not Punggol. It is ordinary office and business-park space, and the live decision is between the central business district and the business parks.

The central business district is where headquarters functions, financial institutions, large multinationals, and an increasing number of frontier technology and AI firms cluster. It is also, at present, expensive and tight. CBRE’s research put core central-business-district Grade A rents at S$12.40 per square foot per month in the first quarter of 2026, the fifth consecutive quarter of growth, with vacancy at a record low of around 3.3 per cent.¹⁷ JLL, measuring on its own definition, recorded Grade A rents at S$12.04 per square foot per month in the same quarter, a seventeen-year high.¹⁸ The cause is straightforward: demand has held up while new supply has thinned to almost nothing. Shaw Towers is the only major office completion scheduled for 2026, and meaningful new supply does not arrive until later in the decade.¹⁹ The market consequence is that landlords hold face rents and negotiate through incentives rather than visible discounts, and that occupiers needing large contiguous floor plates are pre-committing to buildings years before completion.²⁰

A European firm should read those numbers against what a central-business-district address actually buys. The premium is real money, and it should be spent only where it earns its keep. For a headquarters function that hosts clients, recruits senior people, and benefits from being among its peers and its bankers, the premium can be worth paying. The address signals permanence, the building does some of the recruiting, and the proximity to the financial and professional ecosystem has a daily operational value. For a technology operation whose work is internal and whose talent cares more about the commute and the building than the postcode, the same premium usually buys nothing the operation can use.

The business parks are the alternative, and for technology operations they are frequently the better answer. They are developed and largely landlorded by JTC, they offer space designed for research, engineering, and technology use rather than for client-facing prestige, and they sit at a materially lower rent than the central business district. The rent differential between a core Grade A floor and a business-park address is large enough that, for an operation of any size, it funds a meaningful part of the firm’s other costs. The most important of the business-park clusters for technology and research operations is one-north, treated in the next section, but the broader business-park market extends well beyond it, and a firm whose work does not depend on research adjacency has a wide choice of estates at lower cost.

The general geography of Singapore’s industrial and business-park estates was set out in Book 1, Chapter 3, and the digital operator should read this section against that one. The estate logic is the same, and the digital overlay is mainly about which estates concentrate the research and technology tenants a digital firm wants to sit among. A firm that has read the earlier book will recognise the structure; what changes here is the criterion for choosing within it.

The trade-off between the central business district and the business parks is not only cost. It is cost, prestige, talent access, and proximity to the research presence, and the right answer depends on which of those the specific operation actually trades on. A firm should resist the instinct to take the most prestigious address it can afford. The most useful address is the one nearest to whatever the operation depends on, and for a great many digital operations that is research and engineering talent, not a Marina Bay view. The discipline is to identify, before looking at any space, what the operation actually needs to be near, and then to let that answer drive the search rather than letting the available budget drive it toward the most impressive postcode within reach.

8.6 The one-north and Fusionopolis Cluster

A European operation that does research, applied science, or engineering, and especially one that does applied AI, will most likely want the one-north cluster.

one-north is a 200-hectare district in Buona Vista, planned and developed by JTC, built deliberately as Singapore’s research-and-technology hub rather than as a general business park.²¹ It is organised into precincts, each oriented to a cluster: Biopolis for biomedical sciences, Fusionopolis for infocomm technology, media, physical sciences and engineering, Mediapolis for media, and the JTC LaunchPad precinct for startups.²² Fusionopolis occupies roughly 30 hectares within the district and is the precinct most relevant to a digital or engineering operation, offering business-park and research-and-development space including specialised facilities such as clean rooms and vibration-controlled test bays that an ordinary office building cannot provide.²³

The reason one-north matters to a European digital firm is co-location with the public research base. The cluster anchors a substantial part of Singapore’s Agency for Science, Technology and Research, known as A*STAR, and its institutes, including the Institute of Microelectronics and the Institute of Materials Research and Engineering, alongside private corporate research operations.²⁴ Multinationals have located research functions here precisely for that proximity: Procter & Gamble runs its Singapore innovation centre in the cluster, and Wilmar International relocated its global headquarters into Biopolis.²⁵ The connection to the research presence is the same one set out in Chapter 7; the property point is that one-north is where that presence is physically concentrated, and a firm that wants to be near it should be in or adjacent to this district rather than elsewhere.

The property reality is that one-north space is purpose-built and sought-after, and a firm that needs the specialised facilities, meaning clean rooms, test bays, and laboratory-grade services, will find them here and in few other places. For a firm that needs only ordinary technology-office space but values the cluster’s research adjacency, the precinct still works, with the LaunchPad precinct in particular oriented to earlier-stage and smaller technology tenants. The author’s standing advice to a research or applied-AI operation is to treat one-north as the default and to justify any decision to locate elsewhere, rather than the reverse, because the cluster’s concentration of research institutions and technology firms is difficult to replicate by sitting near it but not in it. Proximity in this market is not approximate. A firm two stops away on the train is not in the cluster in the way that matters, which is the informal, repeated, low-friction contact between its people and the research base that the cluster exists to produce.

8.7 Smaller-Format Office: Creative, Shophouse, and Flexible Space

Not every European firm arriving in Singapore is placing a headquarters or a research team. Many are placing a first ASEAN presence, meaning a regional team of a handful of people, testing the market before committing to scale. For these the right property answer is rarely a conventional multi-year lease, and the market offers three formats worth knowing.

The first is flexible and serviced office space. The serviced-office and co-working operators, among them JustCo, The Great Room, The Executive Centre, and The Work Project, offer space on short, scalable terms, with fit-out, furniture, and services included, which suits a firm that needs to preserve capital and retain the ability to grow or shrink the footprint without renegotiating a lease. For a first regional team this is usually the correct first move. It converts a multi-year fixed commitment into a flexible operating cost and buys the firm time to learn what it actually needs before it signs for it. The premium per desk is higher than a conventional lease would be on a per-square-foot basis, but the comparison misleads, because the conventional lease the firm would otherwise sign carries a security deposit, a fit-out cost, a stamp-duty charge, and a reinstatement liability that the flexible arrangement avoids entirely. For a small, uncertain, early operation, the flexible option is usually cheaper once the full cost of the alternative is counted, not more expensive.

The second is the shophouse office. Singapore’s conserved shophouses offer a distinctive office format, with street-level character, a human scale, and an address with identity, and they hold their value as a property class. They suit a smaller creative, design, or client-facing operation that wants a space with personality rather than a tower floor. They carry the practical constraints of heritage buildings: smaller floor plates, conservation requirements that limit what a tenant may alter, and a tenancy market that is narrower and more idiosyncratic than the institutional office market. A firm drawn to a shophouse should understand that it is taking on a building with rules attached, and that the character it is paying for comes with constraints on how the space may be changed.

The third is conventional smaller-format space outside the central business district, where landlords increasingly offer pre-fitted “plug-and-play” units. These are delivered furnished and cabled, on standard gross rent plus an amortisation charge for the fit-out the landlord has funded, which lets a firm occupy quickly without a capital fit-out of its own. For a firm that has outgrown serviced space but is not ready for a full custom fit-out, the pre-fitted unit is a useful middle step, because it gives the firm its own conventional lease and its own door without the capital outlay and the lead time of building from a bare shell.

The practical sequence for a European firm establishing a first presence is to start flexible, learn the market, and commit to a conventional lease only once the operation’s shape and headcount are known. The firms that get into trouble are the ones that sign a conventional headquarters-grade lease on day one against a headcount they have not yet hired and a strategy they have not yet tested. The cost of that mistake is not only the rent on space the firm does not yet fill. It is the reinstatement liability and the lease term the firm is now locked into, against an operation whose actual requirements have turned out to be different from the ones it guessed at when it signed.

8.8 The Lease Structures for Digital Property

The commercial mechanics of a Singapore office lease were set out for industrial property in Book 1, Chapter 8, and for specialised property in Book 3, Chapter 8. The principles carry over; this section adds the digital overlay and treats the data-centre lease as the distinct structure it is.

A standard commercial office lease in Singapore runs for a fixed term, commonly two to three years, often with an option to renew. Rent is quoted on a gross per-square-foot-per-month basis and includes a service charge, broadly in the range of S$0.50 to S$1.20 per square foot per month, covering building services such as air-conditioning during business hours, security, and maintenance, and is subject to Goods and Services Tax where the landlord is GST-registered.²⁶ The tenant pays a security deposit, typically several months’ rent, and is liable for stamp duty on the lease. Stamp duty is a real and frequently overlooked cost. For a lease of four years or less it is charged at 0.4 per cent of the total rent over the term, payable to the Inland Revenue Authority of Singapore within fourteen days of signing a lease executed in Singapore, and an unstamped lease cannot be used in evidence in a dispute.²⁷²⁸ A European firm signing remotely should note the variant: a lease signed overseas must be stamped within thirty days of being received in Singapore.²⁸ The point is not that the duty is large; on most office leases it is modest. The point is that an unstamped lease is worthless as evidence if the firm ever needs to enforce it, and the firm most needs to enforce it precisely when a relationship has broken down, which is the worst moment to discover the document was never stamped.

The data-centre lease is a different instrument. It is governed less by the conventional landlord-and-tenant framework than by a service-level agreement that specifies the performance the operator must deliver: power availability and redundancy, cooling performance, physical security, and connectivity. Such agreements typically provide for acceptance testing before the tenant takes occupancy, periodic facility audits, and structured remedies, meaning service credits or rent abatements, where the operator fails to meet the agreed metrics. They also commonly restrict assignment and transfer more tightly than a commercial office lease, sometimes to the point of requiring an anchor tenant’s consent before the facility itself can change hands. A European operator approaching a data-centre commitment should understand that it is negotiating a performance contract for a service, with a real-estate component attached, rather than a lease with some service terms bolted on. The legal and commercial centre of gravity is the service-level agreement, and that is where the negotiation should concentrate. A firm that sends its real-estate lawyer to negotiate the rent and treats the service levels as a schedule to be accepted has put its effort in the wrong place, because the service levels are the deal.

The general principle a European firm should carry from the earlier books holds here. The lease is not a transaction to be closed and forgotten. It is the multi-year operational foundation of the Singapore presence, and the terms that matter most are the ones that govern what happens later: renewal, expansion, exit, and, above all, reinstatement.

8.9 Fit-Out for Digital Operations

Fit-out is where a European firm’s Singapore property budget most often goes wrong, and it goes wrong in two distinct ways depending on whether the space is office or data centre.

For technology office space the fit-out itself is conventional, covering cabling, partitioning, meeting rooms, and the services a technology team needs, and Singapore now applies a sustainability overlay to it. Newly fitted commercial space falls within the Green Mark for Interiors scheme, which assesses the sustainability and carbon footprint of the fit-out. The fit-out period is commonly delivered rent-free for a span running from a couple of weeks to a few months, against a renovation deposit. None of this is unusual, and a firm that has fitted out office space in Europe will find the process familiar. The one piece of Singapore-specific discipline is to confirm at lease negotiation how the fit-out interacts with the reinstatement obligation, because the more elaborate the fit-out the larger the reinstatement bill at the end, and the two should be planned together rather than separately.

For data-centre space the fit-out is a different order of undertaking. It is a specialised, capital-intensive build of power distribution, cooling, redundancy, fire suppression, and security systems to the operator’s and the regulator’s standards, and a firm taking powered-shell or shell-and-core space is effectively constructing critical infrastructure inside someone else’s building. This is not a budget line to estimate from European experience. It should be scoped with specialist engineers and the operator from the outset, and the firm should expect the build to be governed as tightly as the operation it will eventually house, because the operator’s own service levels to its other customers depend on every tenant’s build meeting the facility’s standards.

The cost that European firms under-budget most reliably, however, and this was true across the industrial and specialised-property chapters of the earlier books and it is true here, is reinstatement. At the end of a Singapore commercial lease the tenant is typically under a strict obligation to return the premises to bare shell, removing the fit-out it installed. This is a real and often substantial cost, scaling with the density and complexity of what was built. For a heavily customised technology fit-out with significant mechanical-and-electrical alterations it can run well above the cost of a plain office reinstatement. For data-centre space, where the fit-out was specialised infrastructure, the reinstatement obligation can be larger still. The mistake is not failing to pay it. The mistake is signing the lease without having priced it, so that a cost the firm always owed arrives unbudgeted at the worst possible moment, at exit, when the operation is already absorbing the cost of moving. Reinstatement should be modelled at the front end, when the lease is signed, not discovered at the back end, when it is due. The discipline is simple to state and routinely ignored: before signing, get an estimate of what it will cost to take the space back to bare shell, and carry that number in the budget for the life of the lease.

8.10 A Worked Property Example

Consider a representative case, in the manner of the worked examples in the earlier books. A German precision-engineering group, the recurring Mittelstand manufacturer of this series, is establishing a Singapore regional operation to serve ASEAN customers. The operation has two parts: a regional headquarters of around fifteen people handling sales, management, and customer engineering, and a modest compute requirement to run a regional instance of the group’s industrial-data and analytics platform, well short of needing a dedicated facility.

The headquarters function does not need a central-business-district address. Its work is internal and regional, its customer engineers travel to client sites rather than hosting clients in a tower, and the rent differential between a core central-business-district floor at the prevailing Grade A rate and a business-park or fringe address is large enough, at the firm’s size, to fund a meaningful part of the operation’s other costs. The firm’s customer-engineering work, however, benefits from proximity to Singapore’s engineering and research base. The decision therefore points away from the central business district and toward a business-park address, with the one-north cluster a strong candidate if the engineering function will draw on the research presence, and a more conventional business-park estate sufficient if it will not.

The scale of the rent decision is worth making concrete. Core Grade A central-business-district space was renting at around S$12.40 per square foot per month in early 2026.¹⁷ A fifteen-person headquarters might occupy on the order of three thousand square feet once meeting rooms and shared space are counted. At the Grade A rate that is roughly S$37,000 a month, or close to S$450,000 a year, in rent alone. A business-park address suited to the same operation sits at a materially lower rate, and the difference over a three-year term runs well into six figures. For an operation of this size that money is not a rounding error. It is the cost of one or two additional engineers, or the first year of the colocation contract, or the reinstatement reserve. The point of the example is not the precise figure, which a firm must obtain for its own specification and which moves quarter to quarter, but the order of magnitude: the address decision for a small headquarters is a six-figure decision over the life of the lease, and it should be made on operational fit rather than on what the firm can afford to sign.

For the first eighteen months the firm should not sign a conventional lease at all. Fifteen people whose number and roles will shift as the operation finds its feet are exactly the profile that serviced or flexible space serves. The firm takes a flexible suite, converts the property commitment into an operating cost, and learns its actual requirement before committing capital. When the operation stabilises, the firm converts to a conventional business-park lease of two to three years, budgeting from the outset for the security deposit, the 0.4-per-cent stamp duty on the total rent, the fit-out against a rent-free period, and, modelled at signing rather than at exit, the reinstatement cost of returning the space to bare shell.

For the compute requirement the firm does not enter the capacity-allocation contest. It does not need a dedicated footprint, so the Call for Application process and its multi-year sustainability commitments are irrelevant to it. It takes colocation space at an established operator, contracting for a reserved block of kilowatts with electricity passed through, and negotiates the service-level agreement, covering power uptime, cooling, security, and the remedies for failure, as the substantive document it is. If the analytics workload turns out not to need a Singapore address for latency, legal, or commercial reasons, the firm should test whether the same colocation is cheaper across the strait in Johor before committing, because for compute that does not need to be in Singapore, it increasingly is not.

The shape of the decision, then, is: flexible space first, business-park lease second, colocation rather than dedicated capacity for compute, and reinstatement priced at the front end. None of it is exotic. All of it is the difference between a clean Singapore entry and an expensive one. The firm that follows this sequence spends its capital when it knows what it is buying, and not before; the firm that does the reverse commits to space, term, and fit-out against an operation it has not yet tested, and pays for the guess.

8.11 The Eight Property Mistakes European Digital Firms Make

Treating data-centre real estate like ordinary commercial leasing. It is not a lease with services attached; it is a service-level agreement with a real-estate component. The negotiation that matters is over power, cooling, uptime, and remedies, not over rent per square foot. A firm that runs the data-centre deal through its standard office-lease playbook will negotiate the wrong terms.

Underestimating the capacity-allocation timeline. A large dedicated data-centre footprint in Singapore depends on capacity that flows through a competitive, sustainability-conditioned application process. That is a multi-year proposition, not a procurement. A firm that assumes it can secure a dedicated facility on a commercial timeline has mis-scoped its single largest risk.

Over-committing to central-business-district space before scale justifies it. Core Grade A rents are at multi-year highs and supply is scarce. A headquarters function that hosts clients may justify the premium; a technology operation whose work is internal usually cannot. Signing a prestige lease against an unproven operation is a recurring and expensive error.

Misjudging the fit-out and reinstatement budget. Reinstatement to bare shell is a strict lease obligation and is the most reliably under-budgeted cost in a Singapore tenancy. For a customised technology fit-out it is substantial; for data-centre space it can be larger still. The firm always owed it. The mistake is not pricing it at signing.

Choosing a location disconnected from the talent and research presence the operation needs. The most useful address is the one nearest whatever the operation trades on. For a research or applied-AI operation that is the one-north cluster; for a headquarters it is somewhere central and well-connected. Picking an address for its prestige or its sustainability rating rather than its fit to the operation wastes the rent.

Under-attending to power and cooling cost in data-centre commitments. Power is the largest cost in running compute, and in colocation it is typically a direct pass-through to the customer. A firm that focuses on the headline space rate and treats power as a detail has not understood where its data-centre cost actually sits.

Copying lease terms without adapting to digital specifics. The general Singapore lease principles from the earlier books hold, but a data-centre service-level agreement, a one-north specialised-space lease, and a flexible serviced-office contract are different instruments with different traps. A firm that applies one template across all of them will be caught by the differences.

Treating the lease as transactional rather than as the multi-year operational foundation. The terms that matter most are the ones that govern later: renewal, expansion, exit, and reinstatement. A firm that closes the lease as a transaction and files it away has deferred its hardest property decisions to the moment it can least afford to face them.

8.12 Conclusion

A European digital firm’s Singapore property commitment is not a single decision and not a single market. It is a set of decisions across several markets, namely data-centre capacity, the purpose-built district, the business parks and the central business district, the one-north research cluster, and the flexible and smaller-format space, and the firm’s task is to match each part of its operation to the market that actually fits it rather than to treat them as one.

The recurring lesson is that the property commitment is the multi-year operational foundation of the firm’s Singapore presence, and that the costs and constraints that matter most are the ones the firm meets later: the capacity it could not secure on its timeline, the rent it overpaid for an address it did not need, the reinstatement it did not budget. Careful attention at the front end, honest about which workload needs a Singapore address, disciplined about not over-committing before scale, and clear-eyed about reinstatement before signing, produces operational stability across the years the lease runs. The next chapter turns from the space the operation occupies to the talent that fills it.

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 combines the author’s licensed real-estate practice in Singapore industrial and commercial property with primary-source verification of the regulatory and market facts cited. Data-centre capacity, allocation criteria, and sustainability standards were verified against published Infocomm Media Development Authority and Economic Development Board materials and the public terms of the Data Centre Call for Application exercises. Office and data-centre market figures were drawn from named brokerage research (CBRE, JLL, Cushman & Wakefield) current to the first quarter of 2026. Lease, stamp-duty, and reinstatement mechanics were verified against Inland Revenue Authority of Singapore guidance and practitioner sources. Estate descriptions for the Punggol Digital District and one-north were verified against JTC Corporation and Economic Development Board materials. Where a market figure could not be verified against a primary or named source, it has been omitted or stated only in general terms. In particular, indicative per-square-metre rents for individual one-north buildings were excluded because they could not be confirmed against a primary or named-brokerage source at the time of writing.

Currency of analysis: The analysis is current as of the date of publication. Singapore office and data-centre rents move quarter to quarter; the capacity-allocation criteria and green-energy thresholds are set by exercise and tighten over successive Calls for Application; and the office-supply position will change as the late-decade pipeline completes. A reader weighing a specific commitment should verify current rents, current capacity availability, and the prevailing allocation criteria at the point of decision.

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 a Singapore digital-operations property decision, whether data-centre capacity, business-park or central-business-district office space, the one-north research cluster, or a first flexible ASEAN presence, 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/de.b5.ch8 (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)