DayOne Data Centers

1.1 Identity, model, and operating footprint

DayOne describes a capital-intensive infrastructure model: it develops and operates hyperscale data centers for hyperscalers and large enterprises, rather than selling a software subscription. The company identifies Singapore as its headquarters and presents SIJORI—Singapore, Johor, and Batam—as a connected regional platform. This is a company description, not independent evidence that the network has achieved the claimed operating advantages. Its public narrative traces the business to GDS International and a GDS Holdings spinout; available reports differ in how they describe the timing of the transition to the DayOne name and independence. The sources reviewed do not establish a verified total for customers, employees, or all-time funding. The snapshot therefore distinguishes reported figures from unknown or unreconciled metrics rather than treating a round's gross proceeds as cash on hand.[CO001, CO002, CO003]

1.2 Leadership, governance, and ownership

Available company-facing pages describe an experienced leadership team but do not identify founders, name an executive roster, or provide board composition. That leaves key-person dependence, governance quality, and control rights unassessed; the absence of names in this assigned source set is not evidence that no such disclosures exist. Reporting identifies Coatue and Hillhouse as the two largest shareholders after leading the Series C, and identifies the Indonesia Investment Authority and Achi Capital Partners as new participants. GDS Holdings remains relevant as former parent: one report says it sold US$385 million of shares back to DayOne in January 2026, while another reports a continuing non-controlling stake. Obtain the current cap table, board roster, voting arrangements, and related-party terms before relying on headline investor names as a proxy for control.[CO003, CO005, CO006, CO016]

1.3 Financing and public-market status

ISI Markets and Data Center Dynamics report a US$4.5 billion gross Series C close in June 2026; ISI says the round began with a US$2 billion tranche in January. Data Center Knowledge also reports a December 2025 mezzanine facility of up to US$1.2 billion. These are financing announcements and reported facility size, not a reconciled total of capital raised, debt outstanding, or cash available. On October 5, Bloomberg reported that DayOne had filed for a US IPO, while other coverage described a roughly US$5 billion raise and a possible US$20 billion valuation. On October 21, Bloomberg reported that the company was weighing a delay or an offering reduced to US$2.5 billion–US$3 billion. Neither target is a completed transaction or a confirmed valuation. The changing reported timetable is a material execution and market- window uncertainty.[CO004, CO007, CO008, CO009]

1.4 Scale, concentration, and diligence limits

Bloomberg reports US$512 million of revenue for the six months ended June 30, 2026, versus US$151.5 million in the comparable prior-year period. The same article reports an US$81.9 million net loss, while Startup Fortune reports US$77.2 million for that period; the discrepancy should be reconciled to the filed statements rather than averaged. Startup Fortune also reports that one customer represented 69.4% of 2025 revenue and 69.2% in the first half of 2026. This is a single-source concentration disclosure and makes customer identity, contract duration, renewal terms, and related capacity especially important. The available evidence does not establish customer count, a current employee count, total raised across all rounds, or a complete debt and covenant schedule. The Johor land transaction and Finnish development plans show expansion activity, but do not establish powered, permitted, or revenue-generating capacity.[CO010, CO011, CO012, CO013, CO014, CO015]

2.1 Market boundary and evidence-constrained sizing

The relevant market is leased or contracted data-center capacity, measured by location, power, availability, and rent—not the broader revenue pools for cloud software, AI models, chips, or digital services. This keeps operator capacity distinct from the workloads that create demand. CBRE reported 13.4% year-over-year inventory growth in Q1 2026 across its Singapore, Tokyo, Hong Kong, and Sydney basket, alongside 608.9 MW of regional net absorption; those are growth and quarterly flow measures, not an absolute Asia-Pacific TAM (CM001, CM004, CM015). The region's reported 7% vacancy also masks substantial market variation, from 2% in Singapore to 18% in Hong Kong (CM002). Singapore's average asking rent was $403 per kW/month for 250-to-500-kW capacity, a price proxy for a defined capacity band rather than total market spend (CM003). No retained source supplies a consistent regional inventory denominator, total addressable spend, or a defensible DayOne-specific SAM/SOM. The sizing table and figure therefore preserve distinct measures and leave the monetary market size unquantified (CM015).[CM001, CM002, CM003, CM004, CM015]

2.2 Buyers, demand signals, and adoption path

Public market evidence identifies hyperscalers, neocloud providers, and AI-focused operators as demand cohorts, but does not establish their relative Asia-Pacific spend or the budget holder for an individual deployment (CM005). DayOne's market-facing proposition is long-duration capacity for hyperscale cloud and AI workloads; AInvest describes take-or-pay arrangements, but this is secondary reporting and not a disclosed contract schedule or independently tested retention measure (CM011). DayOne's 2.3 GW of booked capacity as of September 20, 2026, which it expected to deliver nearly entirely by the end of 2028, is evidence of its own contracted pipeline—not the size of the market or realized utilization (CM008, CM009). AInvest separately reports 4.6 GW of power resources and 962 MW in service on that date; these are distinct capacity stages and should not be added together (CM009). Earlier coverage reported more than 1.5 GW of bookings since launch in 2022, without enough method detail to reconcile its cumulative basis precisely to the later snapshot (CM013). The buyer map marks budget ownership and adoption triggers as unverified where the sources do not disclose them.[CM005, CM008, CM009, CM011, CM013]

2.3 Supply constraints and investment conditions

Demand growth does not translate automatically into deliverable capacity. CBRE attributes longer development timelines in Asia-Pacific core hubs to power availability, higher construction costs, and regulatory complexity, while its Q1 2026 report says availability across major Asia-Pacific markets fell 43% year over year (CM006). Singapore's low vacancy and elevated asking rents are consistent with tight supply, but neither establishes that every project can secure power or achieve the quoted rent. Water is another site-level input: CNA reports that data centers rely heavily on water for server cooling and describes one provider's exploration of analytics-led conservation, which is not a DayOne result or a sector-wide performance benchmark (CM007). Mission Media separately described more than 500 MW of DayOne capacity across Asia-Pacific and Europe, but gave no as-of date or comparable denominator; it should not be reconciled with newer resource, booked, and in-service measures (CM014). Market financing is also exposed to timing risk; Bloomberg reported that DayOne was considering a delay or a smaller US IPO amid data-center market volatility, while The Straits Times described a broader operator IPO pipeline (CM010, CM016). The sector's growth case is therefore conditional on power, permits, water, construction delivery, and financing; project-level costs, customer concentration, and returns cannot be inferred from regional rent or inventory data.[CM002, CM003, CM006, CM007, CM010, CM012, CM014, CM016]

3.1 Direct hyperscale peers and scale signals

DayOne's company announcement positions it as a multi-market, high-density platform and reports approximately 1 GW of secured customer commitments; those are company claims, not independent operating benchmarks. Data Center Index tracks only two named DayOne campuses and 72 MW across them, explicitly warning that its coverage is not every building. AirTrunk is a direct regional alternative: its reproduced company announcement describes more than 280 MW of planned Johor capacity at JHB3 and JHB4, while saying JHB1 and JHB2 exceed 420 MW and are nearly fully contracted. These figures use different geographies, dates, and pipeline definitions, so they indicate competitive investment rather than a like-for-like capacity ranking.[CP002, CP003, CP004, CP005, CP006]

3.2 Incumbents, adjacency, and buyer access

Digital Realty is an adjacent incumbent rather than a direct scale match in the available Malaysia evidence: Tech Wire Asia reports three Cyberjaya assets with an approximately 32 MW IT-capacity target, including a 15 MW AI/HPC facility and a future 14 MW site. Its enterprise pitch is differentiated around private AI environments, multi-cloud access, and interconnection; its ServiceFabric platform is planned for Malaysia in the second half of 2026 subject to licensing approvals. Equinix's Johor page emphasizes colocation, Equinix Fabric, Network Edge, and a broad global network ecosystem, but its customer and network-provider totals are company-reported global figures, not Johor metrics. The map therefore scores evidence visibility, not service quality or actual market share.[CP009, CP010, CP011, CP012]

3.3 Pricing, packaging, and procurement tradeoffs

The retained sources do not establish comparable list prices, realized rates, contract lengths, power commitments, or discounting for DayOne, AirTrunk, Digital Realty, or Equinix. Data-center deals are shaped by capacity, delivery timing, power availability, connectivity, and service scope, but the assigned evidence does not quantify those commercial terms; no price-per-megawatt comparison is defensible here. Digital Realty's reported enterprise-private-AI use case and Equinix's interconnection products show alternatives buyers may evaluate alongside hyperscale capacity, while the evidence does not show which provider wins on total cost. Pricing and contract-level customer evidence should be obtained before asserting a cost moat or undercutting advantage.[CP016]

3.4 Switching costs, moat durability, and displacement risk

The plausible moat is execution at scale: secured customer commitments, nearly contracted campus capacity, and access to power, land, financing, and delivery partners can make a provider difficult to replace once a facility is built and commissioned. The current sources substantiate some announced commitments and investment, but they do not disclose DayOne contract tenor, customer concentration, renewal behavior, switching costs, or site-level delivery performance. Bloomberg's October report that DayOne was weighing an IPO delay or a smaller offering amid data-center volatility, alongside its March report on a potential loan increase, makes financing conditions a material competitive variable rather than proof of distress. Competitor announcements show substantial supply expansion; they do not prove displacement or establish an independent benchmark.[CP002, CP006, CP007, CP013, CP014, CP015]

3.5 Exhibits

4.1 Monetization and reported traction

Light Reading reports that DayOne defines bookings as IT-power capacity committed by clients under legally binding long-term contracts, and places reported bookings at approximately 2.3 GW across ten Asia-Pacific and European markets. It says seven global hyperscale and technology customers drive most bookings, with about 1.4 GW concentrated in Johor; those are capacity and concentration indicators, not recognized revenue. The same report gives US$512 million of revenue for the six months ending June 30, 2026, without a contract-level bridge from booked capacity to billed or recognized revenue. Startup Fortune reports FY2025 revenue of US$484.3 million versus US$178.1 million in FY2024, but the assigned evidence does not provide revenue mix, realized price per unit, or recognition schedules by service. The tables therefore distinguish reported revenue and contracted capacity from undisclosed monetization details.[CI001, CI002, CI003, CI004, CI008]

4.2 Cost structure and unit economics

Light Reading reports a US$81.9 million net loss for the six months ending June 30, 2026, while Startup Fortune reports US$77.2 million for H1 2026; the assigned material does not reconcile the discrepancy (CI005, CI012). Startup Fortune also reports FY2025 adjusted EBITDA of US$180.7 million at a 37.3% margin and H1 2026 adjusted EBITDA of US$205.9 million at a 40.2% margin, alongside FY2025 share-based compensation of US$341.8 million within a US$367.1 million net loss (CI014-CI015). These secondary-source adjusted measures are not audited gross margin, operating cash flow, or cash burn. Channel NewsAsia describes water-intensive cooling across Singapore's data-center sector (CI011), but not DayOne's usage or tariffs. Company power, water, utilization, working-capital, and capex-per-MW inputs remain gaps.[CI004, CI005, CI011, CI012, CI014, CI015]

4.3 Capital adequacy and financing dependency

DayOne's June 5, 2026 release announced US$4.5 billion of gross Series C proceeds and said it could consider additional equity or debt depending on funding needs and market conditions (CI006-CI007). An October 2025 ICEF roadmap reported US$3.58 billion (SG$4.6 billion) in green financing for Malaysian development (CI016). ISI Markets later reported that the company was said to be negotiating an expansion of a corporate loan facility to as much as US$7 billion, not that a facility had closed or been drawn (CI010). Bloomberg reported that a delay or a smaller US$2.5-3 billion IPO was under consideration, with no final decision (CI009). These gross, historical, or conditional financing figures do not establish current liquidity; cash, burn, committed capex, debt terms, and runway remain unavailable in the assigned sources.[CI006, CI007, CI009, CI010, CI016]

4.4 Underwriting verdict

Reported revenue growth and contracted capacity demonstrate scale, but they do not establish revenue quality, durable margins, or capital adequacy. Startup Fortune reports that one customer represented 69.4% of FY2025 revenue and 69.2% in H1 2026 (CI013); customer-level pricing, renewal sensitivity, and sales efficiency remain unquantified. The conflicting H1 loss figures require reconciliation to the primary statements before relying on adjusted margins. Do not infer ARR from bookings, cash burn from accounting loss, or cash available from gross financing proceeds. Underwriting needs executed customer contracts and a billing bridge, site-level delivered capacity and costs, audited cash-flow and margin reconciliations, and a current debt and capex schedule. Until then, the financial picture is promising on scale but incomplete on economics and funding runway.[CI005, CI008, CI009, CI013, CI014, CI015, CI016]

4.5 Exhibits

5.1 Product and customer workflow

DayOne describes itself as a Singapore-headquartered platform that develops and operates next-generation data centers for hyperscalers and enterprises; a provider directory also describes cloud, AI, and high-performance workloads. In practical terms, the offered product is campus capacity and associated colocation infrastructure, not a disclosed standalone software product. DayOne's June 2026 release says it had secured more than 1.5 GW of total bookings since inception, but bookings are not commissioned capacity, utilization, or independently verified customer deployments. The company also calls its delivery model proprietary without publishing the technical or operating procedures that would establish how it differs from other operators. The module and workflow tables keep those claims separate from demonstrated customer outcomes.[CE001, CE003, CE016]

5.2 Campus footprint and maturity

Public descriptions span Singapore, Malaysia, Indonesia, Thailand, Japan, Hong Kong, Finland, and Spain, but the sources mix operating, under-development, and planned assets. The clearest facility-level description is the July 2025 announcement for SG1: a 20 MW Singapore facility with about 40,000 square meters of gross floor area and Phase One targeted to be ready for service in 2026. That is a dated forward schedule, not evidence that the phase is now commissioned. Directory and media capacity figures use different labels and do not provide a reconciled, current site-by-site status. Product diligence should therefore distinguish booked demand and campus count from live IT load, energized capacity, and customer-accepted deployments.[CE002, CE003, CE004, CE010, CE011]

5.3 Architecture and deployment model

The Singapore announcement describes high-density GPU support with hybrid air and liquid cooling, a ten-year power purchase agreement backed by bundled renewable energy certificates, and a solid-oxide fuel-cell proof of concept. It also describes a research partnership with the National University of Singapore to test cooling solutions in a tropical environment. These are concrete design and development signals, but the sources provide no measured PUE, rack density, cooling performance, energy mix, or reliability results. The architecture figure groups only the customer interface, campus design, and energy/research features explicitly described in the pool; it is not a verified engineering schematic.[CE005, CE006, CE007, CE008, CE015]

5.4 Differentiation, trust, and diligence limits

DayOne's materials emphasize rapid deployment, high-density designs, low-carbon power pathways, and an integrated delivery model. The assigned sources do not establish comparative performance or disclose the methods behind these claims. The Singapore announcement says LEED Platinum and BCA Green Mark Platinum certifications would be applied for; a directory separately lists LEED statuses for several Johor buildings. Neither statement establishes certification of SG1. A peer operator's site descriptions are useful only as competitor-issued context, not independent benchmarks. Bloomberg's report of possible IPO delay or size reduction, alongside Startup Fortune's framing of the offering as a test of investor appetite for AI infrastructure, adds capital-timing uncertainty, not evidence of a technical failure. Current site-level security, redundancy, incident, and service-level evidence remains a material diligence request.[CE009, CE012, CE013, CE014, CE015, CE016]

6.1 Buyer segments and the regional proposition

Public descriptions place DayOne between infrastructure development and customer operations: it secures and operates powered, cooled facilities, while customers install their own computing and network equipment. Indexed.vc characterizes its target market as hyperscalers, cloud providers, and enterprises; AI in Asia also describes cloud and AI operators as users. These are reported target segments, not a verified account roster. YesPress presents Singapore, Johor, and Batam as a connected SIJORI proposition, but does not identify which customers use each location or disclose customer outcomes there.[CU001, CU002, CU012]

6.2 Bookings are not yet a retention or utilization record

Two publications report more than 1.5 GW of customer bookings since DayOne's 2022 inception, with YesPress attributing the figure to a June 2026 company statement. Separately, YesPress reports 1,250 MW of committed and 444 MW of billable IT power at December 31, 2025. Those are different measures and reporting dates; the available evidence does not reconcile them into a common cohort or show how much booked demand became energized, occupied, and invoiced. The gap is commercially important for a capital-intensive operator, but it is not a calculated conversion rate.[CU003, CU004, CU005]

6.3 Concentration is visible; named deployment proof is not

Startup Fortune reports that the largest customer represented 69.4% of 2025 revenue and 69.2% of first-half 2026 revenue. Mukhbir, summarizing Reuters Breakingviews, attributes 84.3% of first-half 2026 revenue to ByteDance and Oracle together. The attribution is secondary reporting rather than a company-named customer reference: Startup Fortune says the preliminary prospectus did not name the largest customer. The assigned sources do not provide named site-level deployment status, contract duration, customer-quoted outcomes, or an independently validated reference.[CU006, CU007, CU008]

6.4 Retention and workload alternatives remain diligence priorities

None of the reviewed customer-focused sources supplies NRR, GRR, churn, renewal, cohort, or contract-duration metrics, so booking scale and revenue concentration cannot establish customer durability. Public cloud alternatives also matter to buyer choice: Oracle lists live regions in Batam, Kulai, and Singapore, while Microsoft's November 2025 announcement described a planned Johor Bahru region and named Malaysian Azure users. These are market alternatives and Microsoft customer examples, not evidence that those organizations are DayOne customers or that the services are interchangeable for every workload.[CU009, CU010, CU011]

7.1 Capital-market timing and funding dependence

The financing thesis is exposed to public-market timing as well as construction capital needs. Bloomberg reported on October 21, 2026 that DayOne was weighing a delay or smaller US IPO and that a smaller offering could raise US$2.5–3.0 billion, roughly half the original target (CR001, CR002). Earlier coverage by Mission Media described an approximately US$5 billion target, so the newer report is evidence of a changed range under volatile market conditions, not proof that an offering was cancelled or repriced (CR002). DayOne announced definitive agreements for more than US$2.0 billion of Series C equity in January 2026; its release said this followed US$1.9 billion of Series A and B equity; the company separately described an up-to-€1 billion mezzanine facility secured in 2025 (CR003, CR004, CR016). ISI Markets later reported that an expansion of a corporate loan facility to as much as US$7 billion was being negotiated, not closed (CR015). These amounts describe financing raised, announced, or under discussion; they do not establish current unrestricted cash, drawn debt, covenants, or funding available for specific projects.[CR001, CR002, CR003, CR004, CR015, CR016]

7.2 Customer concentration and contract visibility

The assigned evidence does not quantify customer concentration. DayOne's January 2026 release said its expansion investments would support approximately 1GW of secured customer commitments (CR005), while ISI Markets later reported more than 1.5GW of customer bookings since the company launched in 2022 (CR006). Those differently named totals do not disclose a common measurement date, definition, customer split, or contract status, so they should not be added or treated as a reconciled contracted-MW backlog. Neither source provides customer-by-customer revenue, contracted MW, renewal protection, or cancellation rights. A large counterparty deferral could still reduce utilization and cash recovery, but its magnitude cannot be established from aggregate bookings. Diligence should reconcile booking-to-contract conversion and customer exposures against the campus capex schedule before relying on demand visibility.[CR005, CR006]

7.3 Permitting, power, and resource constraints

DayOne's Singapore buildout sits within a deliberately limited capacity-allocation regime. Data Center Dynamics reported that SG1 is a 20MW project with Phase I scheduled for ready-for-service in 2026, and that DayOne received capacity through a pilot scheme in which four operators shared 80MW (CR007, CR008). Reed Smith describes Singapore's measured growth, land, power, and water constraints, noting that no new capacity had been allocated since the 2023 pilot and that new projects face stringent approval and efficiency standards (CR011). This establishes regulatory and resource exposure, not a finding that DayOne lacks a permit or allocation. For Thailand, a July 2026 legal update describes a PUE threshold of 1.3 for high-efficiency data centers seeking tax privileges, alongside security, redundancy, water-management, and workforce conditions (CR012). It also reports government attention to energy, water, and community impacts (CR013). The assigned sources do not establish which DayOne sites have completed each local approval.[CR007, CR008, CR011, CR012, CR013]

7.4 Commissioning, reliability, and operational execution

A schedule milestone is not operating proof: the SG1 Phase I date is described as scheduled for 2026, while the assigned sources do not confirm commissioning, utilization, uptime, or achieved energy performance (CR007). Data Center Dynamics reports a 10-year power purchase agreement with Sembcorp backed by bundled renewable-energy certificates (CR009). The agreement is a visible mitigation, but it does not establish delivered power volumes, firm grid capacity, or production- scale cooling results. A DayOne mechanical-engineer listing hosted by Simplify describes preventive and corrective maintenance, vendor coordination, safety audits, and 24/7 operations at Lahti (CR014). This is company-issued evidence of stated responsibilities, not evidence of a staffing shortage or an incident. Portfolio growth across countries increases the importance of site-by-site commissioning, maintenance, and regulatory controls; the evidence does not permit a claim that any outage or safety event has occurred.[CR007, CR009, CR014]

7.5 Residual exposure and decision triggers

The risk ranking below is qualitative and reflects the magnitude of disclosed exposure and the visibility of mitigation in the assigned evidence, not a measured probability model. Bloomberg's report of possible IPO delay or downsizing is the clearest current market signal; site delivery and regulatory risks remain contingent on approvals, power, and commissioning evidence. Aggregate customer commitments and bookings do not establish concentration or contract durability. The company describes equity funding and customer commitments, and Data Center Dynamics reports renewable-power procurement, but sources do not disclose current cash availability, debt covenants, site-specific approvals, or independently verified operating KPIs (CR001, CR003, CR005, CR006, CR009, CR015, CR016). Re- underwrite if a material IPO delay or smaller proceeds leaves committed projects without identified financing; if customer-level exposure or protections prove concentrated; or if a campus misses a disclosed ready-for-service milestone without replacement power or financing. These are diligence and thesis-break tests, not predictions of failure.[CR001, CR003, CR005, CR006, CR007, CR009, CR015, CR016]

7.6 Exhibits

8.1 Investment thesis and anti-thesis

DayOne's investment case rests on reported demand and access to capital: coverage describes a US$4.5 billion Series C, more than 1.5 GW of capacity bookings, and first-half 2026 revenue growth. Those are meaningful scale signals, but reported bookings are not recognized revenue and revenue is not cash generation. The same reporting gives a first-half net loss and warns of dependence on a limited number of hyperscale customers. On the supply side, CBRE and the IEA identify power availability, grid connections, and approvals as constraints on where and how quickly data-center capacity can be delivered. The thesis therefore depends on converting demand into energized, profitable capacity without losing access to customers or financing. (CV001, CV006-CV012)[CV001, CV006, CV007, CV008, CV009, CV010, CV011, CV012]

8.2 Price discipline and public-market evidence

The approximately US$20 billion figure in the assigned reporting is a reported target or market estimate, not an official company or investor valuation. The reported US IPO proceeds target also needs to be kept separate from enterprise value, equity value, and investor return: Bloomberg later reported that DayOne was weighing a delay or an offering raising US$2.5 billion to US$3 billion. Neither report supplies an offer price, share count, or a complete capital structure. GDS Holdings' market capitalization is a contextual public-market observation, not a like-for-like DayOne valuation multiple; the companies' business perimeters and ownership differ. With no audited standalone earnings, net debt bridge, or final terms in this evidence set, an entry price or target return would be false precision. (CV002-CV005, CV013-CV016)[CV002, CV003, CV004, CV005, CV013, CV014, CV015, CV016]

8.3 Scenario framing and exit readiness

The scenarios below are conditional decision cases, not forecasts or probability-weighted price targets. The bull case requires disclosed financials, reliable power delivery, customer durability, and a receptive IPO window; its only available valuation reference is the reported approximately US$20 billion target. The base case allows for a resized or delayed transaction while capacity and financing execution remain central. The bear case is a delay or repricing if market access, customer concentration, losses, or power constraints worsen. Bloomberg's reported US$2.5 billion–US$3 billion alternative refers to potential IPO proceeds, not company valuation. No numerical return, probability, or downside range is supportable until primary offering terms and audited standalone financials are available. (CV004-CV009, CV011, CV012, CV015, CV016)[CV004, CV005, CV006, CV007, CV008, CV009, CV011, CV012, CV015, CV016]