SM+ Jakarta Data Center Opens With 21MW for AI Workloads

SM+ opened its SMX01 facility on November 1, adding a reported 21 megawatts of IT capacity in Jakarta’s central business district. The SM+ Jakarta data center supports liquid cooling and conventional air cooling, positioning it for different generations of enterprise and AI equipment. Its opening establishes facility availability, while customer deployment schedules remain a separate question.

SM+ says the facility has reached Ready for Service, meaning it is available to accommodate customer workloads. Taiwanese infrastructure provider Konst has separately identified a substantial lease there. However, its project page still describes customer access and service milestones extending into early 2027.

That distinction matters as established operators expand around Jakarta, including NTT DATA and Princeton Digital Group. Buyers now have another facility to evaluate, but building capacity, contracted capacity, and usable computing services describe different stages of delivery. SMX01’s significance rests on how those stages connect.

What the SM+ Jakarta data center opening changes

SMX01 moves SM+ from a development announcement into operating a flagship Jakarta facility designed to accommodate demanding computing equipment.

The opening changes what prospective customers can evaluate. A development proposal offers engineering plans and a projected delivery date. An operating facility permits a more concrete discussion about access, installation, acceptance testing, and support.

That shift matters even before every customer has installed equipment. Buyers can begin evaluating the physical environment against their requirements rather than relying entirely on future construction milestones. However, opening a building does not establish that every proposed configuration is already available.

SM+ developed SMX01 with Korea Investment Real Asset Management, known as KIRA, and LG Sinar Mas. The company identifies SM+ as owner and manager, KIRA as investment partner, and LG Sinar Mas as the design, construction, and technical operations partner.

These responsibilities create a useful starting point for accountability. They also leave customer-specific questions that an opening announcement cannot answer. A buyer still needs to establish who approves its equipment layout, who accepts installation work, and who handles operational incidents.

The facility’s stated capacity describes an electrical provision for IT equipment. It does not directly describe the number of customers, available AI services, or completed computing jobs. Those depend on what tenants install and how they operate it.

This distinction prevents a common analytical mistake: treating every facility announcement as an immediate expansion of publicly accessible AI computing. Some tenants purchase space and infrastructure for private systems. Others build services that outside customers can rent.

For an enterprise buyer, those arrangements produce different choices. Leasing a facility environment requires responsibility for hardware and software, whether handled internally or through partners. Renting a computing service places more responsibility with the service provider.

SMX01’s opening therefore broadens the infrastructure options around Jakarta without resolving which arrangement suits an individual workload. The relevant question is what a particular supplier can deliver from the site.

The announcement also should not be read as proof that businesses must relocate existing systems. Migration decisions involve application dependencies, network connections, staffing, and contractual commitments. A newly available facility becomes an option within that calculation.

The strongest interpretation is straightforward: SM+ has delivered a physical asset that buyers can investigate. The commercial consequences will emerge through tenant installations, service availability, and operating experience.

That makes the launch meaningful without requiring a claim that Jakarta’s competitive order has already changed. Delivery creates an opportunity to compete. Customer outcomes determine how much that opportunity is worth.

Konst’s lease puts demand behind the launch

Konst provides a named customer commitment, but its lease should not be counted as fully installed or actively consumed computing capacity.

Data Center Dynamics reported on November 3 that Konst had taken 15MW of liquid-cooled capacity at SMX01. Its SMX01 launch coverage also distinguished the new 21MW announcement from earlier figures describing 27MW total capacity and a 6MW first phase.

Those figures should remain attached to their original descriptions. The reporting does not establish a consistent basis for combining them into a new total. It also does not explain whether every figure measures the same delivery stage.

For buyers and investors, the missing reconciliation matters more than a larger headline number. A capacity statement needs a defined scope, a date, and an explanation of what is ready. Without those details, comparisons can exaggerate immediate availability.

Konst’s own SMX01 project details identify the 15MW lease and describe access beginning December 26, 2026. The page lists a first-quarter 2027 target for formal service commencement. These remain company-stated milestones rather than independently verified delivery outcomes.

The schedule does not automatically contradict the building’s November opening. A facility can become available before a particular tenant completes installation and validation. However, Konst’s published schedule needs confirmation against the newer operational announcement.

It would be premature to describe the tenant timeline as either delayed or accelerated. Establishing either requires a current schedule and a clearly defined milestone. The available documents support a narrower conclusion: facility opening and tenant service delivery follow different calendars.

Konst also describes a distinction between the facility environment and customer IT responsibilities. Its project materials identify customer equipment and operational tasks, with optional assistance available for deployment and maintenance. The arrangement illustrates why a capacity lease alone cannot establish service readiness.

A customer bringing its own servers needs a different delivery package from a customer seeking managed computing. Both might use the same building. Their procurement processes, acceptance criteria, and operational responsibilities differ.

The commercial value of a named lease is nevertheless substantial. It identifies a concrete prospective use for the infrastructure rather than leaving the entire demand case dependent on general market expectations.

The next question concerns execution. Equipment must arrive, connect to the facility, pass validation, and become available under the agreed service arrangement. Each milestone adds information that a lease announcement cannot provide.

For the SM+ Jakarta data center, Konst’s commitment strengthens the demand story. Its published schedule also provides a practical reminder to track contracted capacity separately from commissioned equipment and customer-accessible services.

Liquid cooling changes the engineering conversation

The facility’s AI positioning depends on matching equipment requirements across power, cooling, and installation, rather than adding an AI label to conventional space.

SMX01’s reported support for racks up to 130 kilowatts describes the electrical demand accommodated within an equipment cabinet. That specification makes rack configuration central to the commercial discussion. Buyers need to know which layouts can receive the advertised support.

Liquid cooling uses circulating fluid to remove equipment heat. It introduces requirements that extend beyond placing servers in a room. The customer’s equipment and the building’s cooling provision must work together under defined operating conditions.

That creates a practical dependency between hardware selection and facility design. A buyer cannot assess the building entirely separately from the systems it intends to install. Procurement teams need a shared understanding of the proposed configuration.

For example, a company evaluating a dense GPU deployment must establish what cooling interfaces its selected servers require. A GPU, or graphics processing unit, performs parallel computing used in many AI workloads. The suitability of the installation depends on the complete system.

Electrical supply, cooling connections, equipment dimensions, and maintenance access all belong in that assessment. A favorable specification in isolation cannot settle whether the proposed deployment fits.

SM+ says SMX01 can accommodate liquid cooling throughout its data halls alongside air cooling. That describes design flexibility. It does not establish that every hall currently contains liquid-cooled customer equipment or operates at the maximum advertised rack density.

The distinction matters for mixed deployments. An enterprise might retain conventional servers while introducing more demanding AI equipment. Supporting both environments can create options, provided the proposed installation receives the necessary engineering approval.

Flexibility also carries operational responsibilities. Teams need procedures for maintenance, equipment changes, and incident response across the configurations they support. Buyers should examine how those responsibilities appear in delivery documents and service agreements.

LG Sinar Mas CEO Donghyup Han captures the engineering issue in the launch statement: “Power, cooling, and structural load capacity must be planned together from the outset.” That is a more useful claim to assess than a broad assertion of AI readiness.

The question for customers is whether that planning extends to their actual equipment. A building can meet its general design objectives while a particular installation still requires adjustments. Those adjustments affect timing and responsibility.

Buyers should therefore resist translating rack density into an assumed GPU count. Different server designs, networking equipment, and operating configurations consume capacity differently. Published maximums do not establish a universal deployment template.

The mechanism behind SMX01’s proposition is credible at the design level: coordinate the building around demanding equipment requirements. Its practical value will depend on validated configurations that tenants can install, maintain, and operate.

That shifts competitive discussion toward engineering evidence. A supplier able to explain the supported configuration and acceptance process gives a buyer something more actionable than a capacity headline.

Jakarta buyers gain options as established operators expand

SMX01 enters a market where competitors are also delivering capacity, making connectivity and execution central to supplier selection.

NTT DATA announced the opening of its Jakarta 2 Annex facility, JKT2A, on November 3. The company says the Jakarta campus expansion adds 12MW of IT capacity and connects directly with its existing JKT2 environment.

NTT DATA emphasizes access to the network and cloud connections already associated with that campus. Its announcement also introduces connectivity between its Jakarta facilities. The proposition combines additional space with relationships and infrastructure that customers may already use.

That offers a useful comparison with SMX01 without establishing a winner. A buyer expanding an existing installation might prioritize continuity with established connections. Another buyer might prioritize a proposed equipment configuration or an alternative operating arrangement.

Princeton Digital Group illustrates a different strategic choice. In October, it announced an agreement to sell five Indonesian enterprise facilities to SM+, while retaining its Greater Jakarta hyperscale portfolio. Hyperscale facilities accommodate large computing deployments.

Data Center Dynamics’ Indonesian portfolio coverage describes the enterprise transaction as a definitive agreement. It places the expanded SM+ footprint after closing, an important qualification when interpreting portfolio claims.

The transaction suggests different priorities rather than a simple retreat by an incumbent. PDG is retaining large-scale projects while SM+ expands the range of assets associated with its platform. Their strategies overlap geographically without becoming identical.

These developments put pressure on operators to explain the customer problem each facility solves. Available electrical capacity matters, but a procurement team must also evaluate connections, support, migration requirements, and installation timing.

Carrier neutrality means a facility permits customers to choose among network providers. It creates the possibility of choice, rather than guaranteeing that every desired connection is immediately available. Buyers still need confirmation of the providers and routes relevant to their deployment.

This is particularly important when an application depends on systems outside the building. Moving its computing equipment does not remove dependencies on databases, identity services, or external partners. The network path remains part of the application.

The competitive pressure also reaches service providers leasing space inside these facilities. They must explain what customers receive beyond the underlying building. That might involve equipment, operating support, or a managed service, depending on the offering.

For buyers, the expanding selection supports more specific comparisons. Requests can move beyond asking whether a supplier is present in Indonesia. They can address which configuration is available, which party operates it, and when it can enter service.

The SM+ Jakarta data center adds another candidate to that evaluation. Its competitive position will become clearer through the deployments it supports and the customers it retains, rather than its opening date alone.

Reliability and efficiency claims need operating evidence

Design descriptions provide a basis for evaluation, but measured performance and verified certification require their own evidence.

A November 3 release from parent company DSSA describes SMX01’s power and cooling systems as designed in accordance with up to Tier IV standards. These standards concern infrastructure resilience. That wording should remain distinct from an independently verified certification claim.

The same SMX01 operating announcement describes four network entry points, municipal water use, rainwater harvesting, and a recirculating chilled-water loop. These are company-reported design and resource-management features.

The release also discusses Power Usage Effectiveness, or PUE, which compares total facility energy use with IT equipment energy use. It does not provide a measured operating result that establishes SMX01’s performance under sustained customer load.

That leaves an evidence gap rather than proof of a problem. A new facility has limited operating history by definition. Buyers should distinguish the information available at opening from the information accumulated through continued operation.

An efficiency claim becomes more useful when its measurement boundary and reporting period are clear. Without those details, readers cannot determine whether competing figures describe comparable conditions. A favorable adjective contributes little to that assessment.

The same reasoning applies to water. Recirculation describes an approach within a cooling system. It does not independently establish total site consumption, replenishment requirements, or performance under different operating conditions.

Electricity procurement requires separate scrutiny. A cooling design does not establish the source of the electricity powering the building. Likewise, an energy-efficiency statement does not establish the environmental impact of the supply.

Reliability needs equally careful treatment. Redundant infrastructure provides alternative components or paths when equipment fails or requires maintenance. Customers still need to understand which protections cover their installation and how staff operate them.

A supplier’s service agreement should make that discussion concrete. Buyers need clear responsibilities, incident communication procedures, maintenance arrangements, and remedies. A general design description cannot replace those terms.

There is also a distinction between facility reliability and application reliability. A customer’s application depends on its own equipment, software, and architecture alongside the building. Strong facility infrastructure does not remove every potential interruption elsewhere.

None of these questions establishes that SMX01 underperforms. They identify the evidence required to evaluate its claims fairly. Applying the same scrutiny to every candidate also improves the comparison.

For enterprise procurement, maintaining a searchable knowledge base can help organize specifications, contract revisions, and acceptance records. The purpose is to preserve decisions and their supporting documents as supplier materials change.

The appropriate skeptical position is specific: accept the reported opening, attribute the advertised capabilities, and reserve performance judgments for evidence. That approach recognizes delivery while keeping unresolved questions visible.

Three signals will show how deployment progresses

The most useful updates will connect the building’s opening to tenant access, accepted installations, and operating results.

The first signal is confirmation of Konst’s delivery schedule. Its published December access milestone and first-quarter service target give readers concrete dates to follow. The parties should clarify whether those dates remain current after the facility opening.

A confirmed access milestone would strengthen the case that the lease is progressing toward deployment. A revised schedule would need context, including what changed and which customer activities it affects. Either update would be more informative than another statement of market interest.

Access alone would still represent an intermediate milestone. Permission to enter a facility does not establish that equipment has arrived or passed acceptance testing. Readers should look for an explanation of what the milestone enables.

The second signal is evidence of completed tenant installations. Useful disclosures would identify the commissioned environment and distinguish it from reserved space. They would also explain whether the resulting service is private, customer-managed, or available to outside buyers.

Such evidence would strengthen the commercial interpretation of the launch. It would show how physical infrastructure becomes a usable offering. A continuing emphasis on potential capacity would leave that transition less visible.

Public disclosure is not the only source of evidence for buyers. A prospective customer can request relevant acceptance documents and evaluate them under appropriate confidentiality arrangements. The absence of public detail does not establish that deployment has failed.

The third signal is operating evidence from the occupied facility. Measured efficiency, documented maintenance practices, and verified resilience information would help buyers assess the advertised design. The reporting boundary matters as much as the headline result.

Comparable evidence would strengthen confidence in SMX01’s operating proposition. Selective figures without dates or scope would leave the assessment unresolved. This is especially relevant when comparing a newly opened building with facilities carrying longer operating histories.

Together, these signals form a practical sequence. Customer access enables installation. Accepted installations enable service delivery. Continued operation generates evidence about performance and support.

That sequence also gives readers a way to interpret subsequent announcements. A partnership, an equipment shipment, and a live service each add different information. They should not be treated as interchangeable proof of the same outcome.

For developers, the relevant milestone is availability of the computing service their application requires. For enterprise infrastructure teams, it is an accepted environment with defined responsibilities. For investors, it is the progression from commitments toward productive use.

The SM+ Jakarta data center opening establishes a concrete addition to Indonesia’s infrastructure options. It gives the partners an operating asset around which to deliver customer projects, while leaving tenant schedules and performance claims open to verification.

Before choosing a supplier, ask for the current delivery schedule, the supported equipment configuration, and the acceptance criteria. Establish who owns each task between site access and production service. Compare network requirements against the application’s actual dependencies, then review the operating evidence behind reliability and efficiency claims. Keep those answers attached to dated documents so later changes remain visible. The same discipline applies whether buying facility space, managed equipment, or a computing service. SMX01 now belongs in that evaluation, with a confirmed opening and identifiable customer demand. The decision should follow the deployment a supplier can commit to, the responsibilities it accepts, and the evidence it provides. What, precisely, will be ready for your workload on the date you need it?