DayOne Data Centers

1.1 Identity, Product, and Footprint

DayOne Data Centers is best understood as a private, Singapore-headquartered hyperscale data center platform rather than a conventional colocation operator with a single-market footprint. The company says it was established in 2022 and develops next-generation, high-capacity, AI-ready campuses for global hyperscalers and enterprises. Its differentiator is a market-creation model: DayOne describes a proprietary SIJORI hub-and-spoke corridor that links Singapore's connectivity with Johor's cost and land advantages and Batam's expansion capacity. Official market pages show a broader platform spanning Singapore, Johor, Batam, Hong Kong, Tokyo, Greater Bangkok and Finland; the June 2026 final Series C release added Spain to the disclosed operation-or-development list. The business model is therefore a capital-intensive capacity-origination platform: secure land, power and permits ahead of demand, package the campuses as high-density AI-ready supply, and convert scarce regional capacity into long-term bookings from hyperscale customers. The strongest public support is for identity, markets and product architecture; named customers, revenue and precise utilization are not public.[CO001, CO002, CO003, CO004, CO005, CO006]

1.2 Leadership and Governance

The public leadership surface is concentrated but credible for an infrastructure company at this stage. DayOne's official site names Lim Ah Doo as chairman, emphasizing more than four decades of experience across telecommunications, technology, data centers, infrastructure, banking and investment, including prior board-level roles. Jamie Khoo has served as CEO since March 2024 and is the central operating voice in the company's financing, TNB renewable-energy and corporate narrative. The most material 2026 leadership change was the appointment of Chengkang (CK) Yan as CFO on June 30, 2026; his background at Hillhouse, KKR and Citi is directly relevant because DayOne's growth depends on recurring access to institutional capital, structured debt, land and power. Governance remains only partially visible: Coatue and Hillhouse are now the two largest shareholders, INA and Achi participated in the final Series C, and GDS remains a minority investor after a US$385 million share repurchase. Board composition, reserved matters, founder economics and customer concentration governance are not disclosed, leaving a key diligence path around control and conflicts as DayOne weighs IPO or strategic alternatives.[CO010, CO011, CO012, CO023, CO024, CO044]

1.3 Capital, Valuation, and Scale

DayOne's capital formation accelerated sharply in 2026. In January, it announced definitive agreements for more than US$2.0 billion of Series C equity led by Coatue with INA support, priced at a 100 percent premium to the prior round and building on US$1.9 billion of Series A/B equity plus an up-to-€1 billion mezzanine facility. By June 5, 2026, DayOne announced a final Series C close of US$4.5 billion, with Coatue and Hillhouse as the two largest shareholders and INA and Achi Capital Partners among notable participants. The traction metrics also stepped up: January materials cited approximately 1GW of secured customer commitments, while the June final close reported more than 1.5GW of total capacity bookings since inception. Independent reporting adds the valuation frame: Data Center Dynamics and Startup Fortune describe a roughly US$20 billion IPO valuation target or reported valuation, while DCD also reported that MGX was exploring an acquisition and that an IPO could still proceed. Treat the US$20 billion number as reported, not company-confirmed, and underwrite it against contracted bookings, power access and execution risk rather than revenue metrics, which remain undisclosed.[CO013, CO014, CO015, CO016, CO017, CO018]

1.4 Infrastructure Partnerships and Regulatory Exposure

The operating story is increasingly about whether DayOne can convert scarce land and power into delivered megawatts faster than competitors. Data Center Dynamics reported more than 500MW in service and under construction and more than 500MW held for future development, while DayOne's own journey cites NTP Johor, HK1 customer move-in, SG1 groundbreaking, Chonburi Tech Park, Tokyo phase one and 34MW at Nongsa Digital Park in Batam. In Malaysia, DayOne signed renewable-energy supply agreements with TNB entities covering roughly 1.5GWp of solar and 2.2GWh of battery storage and reported more than 1GW of renewable energy secured through the TNB partnership. Indonesia is emerging as both a DayOne market and a partner ecosystem: independent reporting describes a Nvidia-backed Firmus plan for a 360MW, 170,000-GPU Batam AI data center with DayOne as infrastructure partner. The adverse side is not a company-specific legal event in the reviewed sources, but a structural risk: data centers are drawing policy scrutiny over grid strain, water, noise, emissions and land use, and Singapore is strengthening sustainability and resilience governance. That makes power contracts, permitting, community engagement and uptime governance core diligence items, not ESG footnotes.[CO019, CO025, CO026, CO027, CO028, CO029]

1.5 Milestones, Adverse Signals, and Evidence Gaps

The milestone record supports a rapid platform build but also exposes how much diligence still depends on private materials. DayOne's journey starts with 2021 land acquisitions in Batam and Johor, then the 2022 Singapore headquarters, 2023 Singapore capacity award, 2024 Japan and Thailand entries, 2025 Finland and financing milestones, and multiple 2026 campus, capital and partnership events. That chronology is internally coherent and corroborated by independent funding and capacity reporting. The biggest gaps are not absence of identity evidence; they are the economics behind the identity. Public sources do not disclose revenue, ARR, run-rate, margins, headcount, named hyperscale customers, contract duration, customer concentration, signed-versus-reserved capacity, project-level power usage effectiveness, or the governance terms around IPO/MGX alternatives. The adverse research also found broad data-center opposition and resource-use risks but no retained source documenting a DayOne-specific lawsuit, sanction, outage or community dispute by the run date. The chapter therefore establishes DayOne as a well-funded, fast-scaling AI-infrastructure platform while explicitly reserving final judgment on customer quality, profitability and execution until data-room evidence is available.[CO032, CO033, CO034, CO035, CO036, CO037]

1.6 Exhibits

2.1 Market boundary: powered hyperscale capacity, not a generic IT TAM

DayOne should be underwritten against the market for powered, networked, AI-ready data-center campuses sold to hyperscale cloud, AI infrastructure and large enterprise buyers, rather than against a loose digital-transformation TAM. The included spend is wholesale colocation capacity, powered shell or build-to-suit campus capacity, interconnection, high-density cooling and site-level energy integration. Excluded spend includes customer-owned GPUs and servers, cloud software revenue, consumer connectivity and general IT outsourcing. This boundary matters because DayOne’s own pitch is not a software workflow but a geography-and-power platform: it clusters Singapore, Johor and Batam into SIJORI and adds Thailand, Tokyo, Hong Kong and Finland as capacity nodes. The market is therefore large, but adoption is gated by power delivery, land, permits, cooling, latency and customer precommitments rather than by simple software adoption curves. The practical underwriting frame is therefore a constrained infrastructure market: customers may want cloud capacity everywhere, but DayOne only earns attractive returns where it can convert site control, grid access and latency into contracted capacity faster than peers.[CM001, CM012, CM023, CM031, CM039]

2.2 Demand and sizing: AI creates a supercycle, but the usable SAM is power-constrained

The strongest market evidence is the convergence of three independent signals: JLL’s 2026 outlook calls for roughly 100 GW of new capacity between 2026 and 2030, CBRE reports tight vacancy and persistent preleasing in major hubs, and the IEA says data-centre, AI and crypto electricity use could double by 2026. For DayOne, the broad TAM is global hyperscale and colocation capacity, while the practical SAM is APAC plus selected European markets where the company can secure power, land and permits. APAC is particularly relevant: JLL forecasts regional capacity rising from 32 GW to 57 GW by 2030, and CBRE identifies Tokyo, Johor and Batam as positioned for AI-related demand or spillover transactions. The attractive feature is scarcity value in powered sites; the adverse feature is that every credible competitor is chasing the same bottleneck. The chapter deliberately avoids a single headline TAM because broad forecasts can overstate opportunity when interconnection queues, substation lead times and customer fit-out budgets determine which megawatts become revenue-bearing supply.[CM002, CM003, CM004, CM006, CM007, CM008]

2.3 Geographic context: SIJORI plus selective APAC and Nordic nodes

DayOne’s most differentiated market logic is SIJORI. Singapore supplies connectivity, enterprise trust and premium pricing, but land and power limits make very large single-market deployments hard. Johor then becomes the scale valve, with MIDA reporting that 78.6% of Malaysia’s operational IT capacity is already in Johor and DayOne claiming hundreds of megawatts across NTP and KTP. Batam adds a lower-cost, SEZ-based node close to Singapore with submarine-cable access. Outside SIJORI, Thailand offers a government-backed digital-economy setting and a large Chonburi campus thesis; Tokyo offers demand density and latency; Hong Kong offers financial-market and cross-border connectivity but limited land; Finland offers renewable-power and waste-heat positioning for European customers. These locations make the market attractive, yet each also carries a distinct bottleneck: grid buildout, water, permitting, customer concentration or execution timing. This creates a portfolio question rather than a single-market answer: Singapore anchors trust and latency, while adjacent or renewable-rich nodes must prove they can absorb workloads without sacrificing performance, compliance or carbon objectives.[CM009, CM010, CM011, CM013, CM014, CM015]

2.4 Buyers, adoption path and constraints: speed-to-power is the purchase trigger

The buyer is rarely a single user. Cloud capacity planners, AI platform teams, network architects, sustainability officers, finance and real-estate infrastructure leaders all influence the purchase, and the budget often spans capacity reservation, construction commitments, power procurement and technical fit-out. Adoption begins when a customer forecasts workload demand, then narrows to latency zones, sovereignty needs, power availability, rack density and carbon requirements. DayOne’s claimed advantages—prefabricated delivery, liquid cooling, SOFC trials, renewable PPAs and multi-market expansion—map well to that funnel. However, public evidence does not reveal contracted customers, occupancy, utilization, achieved pricing or campus-level returns, so a precise SOM remains unavailable. The diligence answer is to treat the market as structurally attractive but execution-sensitive: verify grid milestones, customer preleases, utility contracts, PUE/WUE performance, and the cost of meeting customer sustainability requirements. In diligence, the decisive proof is not another market-size chart but a conversion bridge from forecast workloads to signed reservations, energization milestones, cooling performance and customer-specific deployment dates.[CM028, CM029, CM030, CM031, CM034, CM035]

2.5 Exhibits

3.1 Competitive landscape and peer map

DayOne competes in a crowded but segmented digital-infrastructure field. The closest direct peers are not generic colocation landlords; they are hyperscale campus developers that can secure power, land, modular construction, and renewable-energy pathways in Asia-Pacific markets. PDG, AirTrunk, Bridge Data Centres, and STT GDC fit that job-to-be-done most closely, while Equinix and Digital Realty compete through global platform breadth, customer ecosystems, and interconnection depth. Keppel DC REIT is better treated as an asset-owner and capital-market peer than as a pure operating substitute, because its disclosed strategy focuses on data-centre real estate assets, colocation/single-tenant assets, shell-and-core assets, and disciplined portfolio capital recycling. Status quo and internal build also remain real substitutes for hyperscalers: JLL explicitly expects hyperscalers to keep using both leasing and self-building strategies. The result is a two-axis market: global ecosystem breadth on one side, and power-led APAC hyperscale execution on the other.[CP001, CP018, CP025, CP030, CP031, CP039]

3.2 Geographic footprint, power, and capacity positioning

DayOne’s strongest public footprint evidence is concentrated around SIJORI. The company claims operations or markets across Singapore, Johor, Batam, Greater Bangkok, Hong Kong, Tokyo, and Finland, and its Johor page claims more than 478 MW of operational capacity across Nusajaya Tech Park and Kempas Tech Park with direct Singapore connectivity. That is a meaningful wedge because CBRE says major-market supply constraints are pushing attention toward Johor and other secondary markets, and that Johor and Batam are already capturing sizable multi-MW transactions. However, breadth still favors incumbents and scaled specialists. Equinix reports 63 APAC data centers and interconnection facilities in nine countries; Digital Realty reports 20-plus APAC data centers across 10 metros; PDG has a seven-country pan-Asia platform; AirTrunk discloses about 2 GW of powered land and, through Blackstone, more than 800 MW committed to customers plus land for more than 1 GW of future growth. DayOne therefore looks strongest as a corridor-focused challenger, not as the broadest platform.[CP002, CP003, CP004, CP010, CP015, CP019]

3.3 Capability and customer-focus comparison

Across the peer set, the buying criteria are converging around AI-ready density, fast RFS dates, power certainty, direct connectivity, and sustainability. DayOne claims a modular process capable of up to 500 MW annually, an 8.5-month greenfield delivery record for NTP Johor, more than 300 MW deployed across liquid-cooled sites, and a 130 kW-plus per-cabinet pilot using NVIDIA GB200 NVL72 systems. Digital Realty counters with a publicly described high-density colocation offer from 30 kW to 150 kW per cabinet and liquid-cooling options, while Equinix positions metro AI inference and xScale training as separate deployment patterns. PDG and Bridge emphasize AI-ready greenfield campuses, local delivery, utility partnerships, advanced cooling, and proximity to Singapore. The capability matrix therefore does not support a single winner; it shows DayOne with particularly strong speed/density claims, incumbents with stronger ecosystem breadth, and pan-Asia peers with similar customer-centric campus narratives.[CP005, CP006, CP007, CP012, CP016, CP017]

3.4 Pricing, packaging, and capital access

Public pricing is the weakest comparable dimension. Vendor pages disclose packaging themes—colocation, connectivity, AI-ready high density, build-to-suit, powered land, renewable-ready infrastructure—but not realized prices, discounts, power pass-throughs, customer credits, or fit-out responsibilities. CBRE is useful as a market proxy: Singapore remained one of the most expensive global markets at US$310 to US$470 per kW per month in Q1 2025. Capital access is easier to compare directionally. Equinix and Digital Realty have public-company filing histories and global capital markets access; Keppel DC REIT discloses S$6.3 billion of AUM; AirTrunk gained Blackstone/CPP backing at a transaction value above A$24 billion; PDG discloses US$1.8 billion of green and sustainability-linked loans; DayOne lists more than US$2.0 billion of Series C equity financing plus earlier large facilities. This supports underwriting DayOne as capital-credible, but not uniquely advantaged versus AirTrunk, public REITs, or global incumbents.[CP008, CP013, CP020, CP027, CP032, CP035]

3.5 Moat durability, switching costs, and adverse evidence

DayOne’s moat is most plausible where site control, power access, modular delivery, liquid cooling, and SIJORI network effects compound before peers can secure equivalent capacity. Yet the same evidence base raises adverse points. CBRE calls limited power availability the prime inhibitor of growth and notes preleasing and construction timelines stretching to 2027 and beyond. JLL says grid-connection waits in primary markets exceed four years and estimates a US$3 trillion industry investment requirement by 2030. Those conditions can help a fast operator if it has secured power, but they also attract better-capitalized incumbents and infrastructure funds. Switching costs are real once a hyperscaler preleases multi-MW capacity and integrates cloud/network connectivity, but they are not absolute: hyperscalers can multi-home and self-build, and the largest peers can finance power-adjacent strategies. The diligence priority is therefore not whether DayOne has differentiation; it is whether that differentiation is already locked through customer commitments, utility arrangements, and RFS dates.[CP033, CP036, CP037, CP038, CP039, CP044]

3.6 Exhibits

4.1 Revenue Model and Backlog Proxies

DayOne looks financially less like a software startup and more like a private hyperscale infrastructure platform whose economic engine is contracted power, space, cooling, connectivity and facility operations. The public record does not show an SKU card, take-rate schedule or GAAP revenue bridge. Instead, the strongest revenue signals are capacity, market placement and customer-intent proxies. DayOne's own materials frame the buyer as hyperscalers and enterprises that need speed, scale, latency and high-density AI infrastructure; GDS's public descriptions of outsourced data-center customers reinforce that this business is usually sold to cloud, internet, financial and telecom buyers under long-duration capacity agreements. For underwriting, the most important proxy is therefore not website traffic or customer count but committed megawatts and customer move-ins. The SIJORI strategy is financially meaningful because Singapore capacity is scarce and expensive while Johor and Batam offer expansion land, power and lower-cost operating environments close to Singapore. Still, no retained source discloses DayOne's contracted rent per kW, renewal terms, minimum commitments, pass-through power charges, customer concentration or backlog conversion schedule. The 1.5 GW bookings figure reported by Startup Fortune would be a major backlog signal if proven, but it remains a secondary-source claim until customer contracts, lender packages or management schedules corroborate it.[CI001, CI002, CI003, CI004, CI021, CI022]

4.2 Financing History and Valuation Signals

The financing file is stronger than the operating-metric file. DayOne's own timeline discloses a rapid capital stack: approximately $587 million of Series A funding, $1.2 billion of Series B funding, up to EUR 1 billion of mezzanine financing for Finland and global expansion, RM15 billion of multicurrency green financing for Malaysia, and over $2.0 billion of Series C equity financing in 2026. GDS's SEC-filed January 2026 release is the cleanest valuation signal because it is not just a press headline. GDS said DayOne would repurchase $385 million of shares at the same price as the Series C preferred issue; that transaction recycled about 95% of GDS's principal invested in DayOne at nearly 6.5x money-on-money and left GDS with a stake implied to be worth more than $2.2 billion. Those facts imply meaningful private-market appreciation even without a fully disclosed post-money valuation. Startup Fortune reported a larger $4.5 billion Series C, a $20 billion valuation and a possible $7 billion loan, but those figures are treated as lower-confidence because the official DayOne page and GDS filing retained here corroborate only the over-$2.0 billion Series C and GDS sale mechanics. The conclusion is that investor demand is real, but headline valuation should not be relied upon without cap table, preference stack and debt-facility detail.[CI013, CI014, CI015, CI016, CI017, CI018]

4.3 Capex Intensity and Unit Economics

DayOne's public financial story is dominated by capital intensity. JLL's 2026 forecast of $11.3 million per MW for shell-and-core construction, plus AI fit-out that can reach $25 million per MW, makes clear why DayOne is pairing equity with green financing, mezzanine facilities and likely project-level debt. The company may have a build-speed advantage: it claims one of the world's largest modular solutions, up to 2,500 modules and 500 MW of annual fabrication capacity, and an 8.5-month greenfield delivery timeline for the Johor NTP project. It also claims more than 300 MW of liquid-cooled sites deployed since 2020 and pilots above 130 kW per cabinet using NVIDIA GB200 NVL72 systems. Those details matter because AI density turns cooling, power delivery and supply-chain control into direct margin drivers. However, the public file still does not disclose utilization, realized rent per kW, power pass-through margin, energy procurement cost, depreciation, maintenance capex, land basis, construction cost by campus or stabilized yield-on-cost. Public comps such as Digital Realty and Equinix disclose these categories in audited filings; DayOne does not. The best underwriting posture is therefore scenario-based: treat signed capacity as the revenue leading indicator, model construction and fit-out costs using market benchmarks, and require management to provide site-level development budgets and customer contracts before underwriting free cash flow.[CI025, CI026, CI029, CI030, CI031, CI032]

4.4 Financial Verdict and Diligence Blockers

The financial verdict is attractive but not yet underwritable from public evidence alone. DayOne has high-quality signals: institutional financing, a GDS-validated Series C price, large disclosed capacity footprints, a Singapore-adjacent market thesis, and credible infrastructure-market tailwinds. It also has serious underwriting gaps. The Johor page contains an internal capacity inconsistency, the private company discloses no revenue or margin metrics, and key figures such as cash on hand, monthly burn, runway, debt maturity, interest cost, customer concentration and backlog are unavailable. Power and grid constraints are not theoretical; JLL, Cushman, Uptime and DOE all point to power availability, cost, staffing and AI density as sector-wide constraints. DayOne has some mitigation evidence through TNB, water-treatment, supplier ecosystem and liquid-cooling initiatives, but those do not replace project-level economics. A financing-led bullish case should therefore be conditional: require audited or management-prepared revenue, EBITDA and cash-flow schedules; signed customer commitments by site and MW; debt-facility terms; construction budgets; power procurement terms; and stabilized yield-on-cost by campus. Until then, DayOne should be treated as a scarce, well-financed AI-infrastructure platform with positive capacity optionality and material disclosure risk.[CI005, CI006, CI007, CI008, CI009, CI010]

4.5 Exhibits

5.1 Product definition and asset map

DayOne's product is best understood as hyperscale data-center capacity packaged as a repeatable regional platform, not as a software-only infrastructure product. Customers buy or reserve space, power, cooling, connectivity and operating reliability in campuses that DayOne places near constrained demand pools. The public evidence shows an explicit SIJORI architecture linking Singapore, Johor and Batam, then a broader expansion layer across Hong Kong, Tokyo, Greater Bangkok and Finland. The service workflow starts with market selection and capacity commitment, moves through land and power securing, uses prefabricated delivery and high-density cooling, and ends in an operated facility with monitoring and sustainability claims. The strongest product assets are therefore campus rights, utility interconnections, modular construction capability, liquid-cooling know-how, and customer-specific high-density readiness. Public materials do not prove that every hall is 30 MW-plus; the better-supported claim is that selected campuses and power blocks are large enough for hyperscale AI capacity while individual hall design remains a diligence item.[CE001, CE002, CE003, CE012, CE016, CE019]

5.2 Architecture, delivery and operations

The operating architecture has a clear pattern across markets. DayOne first secures land, grid access, regulatory support and connectivity; then it shortens deployment with modular prefabrication; then it configures the technical envelope around power density, cooling mode, resilience and management software. The company claims that its modular approach can fabricate up to 2,500 modules and up to 500 MW annually and that a Johor project reached 8.5-month greenfield delivery. Public filings from GDS independently echo the same architecture by describing DayOne as using high-density, liquid-cooling-enabled designs, renewable and low-carbon pathways, and rapid-deployment prefabricated models. The infrastructure-management platform adds a software layer for live energy, thermal and system-health monitoring, but the public file does not expose APIs, uptime telemetry, fault-domain design or site-level PUE. As a result, the delivery model looks unusually specific for a private data-center platform, while operational evidence still requires customer diligence packs.[CE002, CE003, CE004, CE005, CE006, CE027]

5.3 AI density, power, cooling and network dependencies

The AI-ready story is credible but dependency-heavy. DayOne is an NVIDIA Data Center Partner and says its verified facilities are built to NVIDIA Reference Design standards; its solutions page also claims a pilot above 130 kW per cabinet using GB200 NVL72 systems. That supports a high-density direction, but it does not by itself disclose how many racks, halls or campuses are ready for those densities. The power system is the binding constraint. Singapore depends on a tightly regulated 20 MW allocation and a SOFC proof of concept; Batam depends on phased 2026-2027 grid delivery under a 511 MVA PPA; Malaysia depends on CRESS renewable execution; Thailand depends on Amata, AMBRE and PEA-linked renewable integration; Finland depends on grid connection, waste heat and permitting. Cooling is equally differentiated by geography: Singapore tests tropical hybrid/liquid cooling, Johor and Batam target high-density regional scale, and Finland leans on naturally cool climate and zero freshwater cooling. The architecture is attractive because it is multi-market, but each market moves only as fast as utilities, permits and cooling validation allow.[CE005, CE007, CE008, CE009, CE010, CE011]

5.4 Sustainability, trust and compliance

DayOne's sustainability posture is more developed than its public reliability posture. Malaysia has the most concrete renewable-energy evidence, with DayOne describing about 1.5 GWp of solar and 2.2 GWh of battery storage under CRESS and more than 1 GW of renewable energy secured through TNB. Singapore adds a Sembcorp PPA, SOFC proof of concept, NUS tropical data-center R&D, and planned LEED/BCA Green Mark Platinum applications. Johor has a visible LEED track record, including NTP1 Platinum and several 2026 Platinum certifications. Finland contributes waste-heat recovery and zero freshwater cooling. Trust controls are thinner in public: Hong Kong lists ISO 9001, ISO 27001 and SOC2, DayOne publishes a data-security policy and supplier code, but there is no site-by-site SOC2 report, SLA, incident ledger, cybersecurity packet, or PUE history in the retained sources. For underwriting, sustainability is a credible customer-facing feature; reliability and security require private diligence.[CE013, CE014, CE015, CE019, CE023, CE024]

5.5 Roadmap, maturity and technical risks

The maturity map is uneven by design because DayOne is scaling across several development stages at once. Hong Kong and Johor provide the strongest proof of operating maturity, with certifications and in-service campus claims. Batam NDP is opening from 2025 while Kabil adds a very large grid-backed expansion. Singapore SG1 is a strategic testbed with 20 MW approved capacity and first-phase timing disclosed as 2026/2027 across sources. Thailand has construction and a 300 MW/1 GW platform ambition. Finland has large planned capacity, local-city support and a 2027 operations target, but permitting and tenant execution remain relevant. Tokyo is a later-dated 2028 project. The most material risks are not whether demand exists for AI compute; they are whether DayOne can synchronize power interconnection, renewable procurement, high-density cooling validation, modular quality control, and compliance evidence across seven markets without creating delivery slippage or uneven customer experience.[CE018, CE020, CE021, CE022, CE023, CE024]

5.6 Exhibits

6.1 Who pays, and where demand clusters

DayOne’s publicly visible customer base is best understood by segment and geography rather than by logo list. The official positioning is hyperscale-first: DayOne says it builds for customers that need deployment speed, scale, and access to strategic underserved markets, while its Series C announcement ties roughly 1GW of secured commitments to AI and cloud infrastructure demand. The customer mix therefore appears weighted toward global hyperscalers, cloud platforms, AI-native GPU operators, and regulated enterprises that need either local Singapore compute or nearby overflow capacity. Geography matters to the customer thesis. Singapore provides scarce, policy-supported local compute; Johor offers Singapore-adjacent scale and direct connectivity; Batam supplies expansion power and land; Finland adds renewable European optionality. This segmentation is supportable, but the revenue mix is not: the retained sources do not disclose customer count, top-customer concentration, or segment revenue.[CU001, CU002, CU003, CU004, CU005, CU006]

6.2 Committed capacity and backlog quality

The strongest public demand metric is the approximately 1GW of secured customer commitments cited in DayOne’s January 2026 financing materials and repeated by independent market coverage. Backlog quality improves where those commitments connect to enabling facts: Singapore SG1 has a 20MW Phase One RFS target in 2026 and a renewable-backed Sembcorp PPA; Batam has a 511MVA, roughly 450MW, PLN Batam power arrangement phased over 2026-2027; Johor has operating capacity and direct Singapore connectivity; Finland has 281MW across Lahti and Kouvola plus a larger Nurmijärvi option that remains approval-dependent. The important underwriting distinction is that capacity and power are not the same as revenue. Public evidence supports large demand and delivery intent, but the conversion from commitment to signed lease, fit-out, utilization, and renewal remains mostly private.[CU004, CU007, CU009, CU010, CU012, CU013]

6.3 Named proof is led by Firmus, while most customers remain unnamed

Firmus is the clearest named customer-demand signal found in public sources. Tech Wire Asia and Certified Strategic report that Firmus will rent DayOne-built Batam capacity for a 360MW NVIDIA-backed AI factory, with up to 170,000 GPUs planned for delivery from 2027 into early 2028 and committed offtake claims in the tens of billions of dollars under the NVIDIA partnership. That proof matters because it links DayOne capacity to a specific AI-native use case rather than to generic hyperscale marketing. It also sharpens concentration risk: if one named AI infrastructure project becomes a large demand block, DayOne needs private evidence that economics, cancellation rights, power obligations, and customer diversification are robust. Beyond Firmus, public sources refer to HK1 customer move-in and secured commitments but generally omit tenant names and lease terms.[CU017, CU020, CU021, CU022, CU023, CU024]

6.4 Retention is not disclosed, but switching costs look structurally high

No retained source disclosed DayOne NRR, GRR, churn, renewal rates, SLA credits, or satisfaction scores, so retention must be diligenced directly. Still, the product and policy evidence suggest switching costs can become high after deployment. DayOne describes bespoke in-house builds, liquid-cooled high-density infrastructure, live monitoring, and customer-specific operational controls; its data security policy restricts Customer Data use, Customer Equipment access, security incident notification, and shared-responsibility boundaries. These are the kinds of controls that become embedded in customer operating models once workloads, power design, cooling design, and monitoring are live. That said, switching-cost logic is not retention proof. The diligence path is to obtain contract duration, renewal/cancellation terms, SLA-credit history, incident logs, support metrics, and customer reference calls by campus.[CU025, CU026, CU027, CU028, CU029, CU036]

6.5 Customer risks: concentration, delivery, power, and AI resilience

The customer-risk profile is concentrated in what public evidence does not show. EY’s sector work flags the structural risk that data center operators can depend on a small number of hyperscalers and cloud providers, while power access, construction delays, and customer churn can directly affect growth. Uptime and Data Center Knowledge add that AI-driven density, cooling load, grid instability, and power-system complexity are increasingly important reliability questions. None of these sources names a DayOne-specific outage or failed deployment, but they are directly relevant to a company selling high-density AI-ready capacity. DayOne has credible mitigants: PPAs, modular build claims, government-backed capacity awards, security controls, and partner ecosystems. The orchestrator should still treat customer concentration, signed lease quality, and service-level performance as data-room priorities before relying on the 1GW commitment figure as durable revenue.[CU031, CU032, CU033, CU034, CU035, CU040]

6.6 Exhibits

7.1 Severity ranking and thesis impact

DayOne’s top risk is not a single jurisdictional defect; it is the concurrency of a hyperscale platform buildout across power-constrained and policy-sensitive markets. The SG1 permit and SOFC pilot are strategically valuable, but they introduce unproven onsite-power execution into a market where EMA is actively managing generation adequacy. Johor is already large enough to become a concentration point for both capacity and grid/water scrutiny, while Batam, Chonburi, Fuchu, Hong Kong and Finland each add local permitting, land, utility and disclosure dependencies. The investment implication is a research-more posture on risk: valuation upside depends on converting committed power and bookings into billable, permitted, renewable-backed capacity without overruns or policy backlash. The ranking also weights source quality: company disclosures establish the project map, while regulators and filings establish the constraints that can invalidate optimistic ramp assumptions.[CR001, CR002, CR004, CR005, CR010, CR012]

7.2 Power, grid and permitting risk by jurisdiction

Power is the binding constraint behind most of the chapter’s risk register. Singapore’s government is explicitly managing generation adequacy and solar intermittency, so a 20 MW DC-CFA allocation and a first-of-kind SOFC proof of concept should be treated as a monitored milestone, not a de-risked capacity entitlement. Malaysia is more expansionary, but MIDA’s own materials show why power and water are becoming policy variables: Johor dominates operational IT capacity, national data-centre demand could grow sharply by 2035, and grid-reinforcement costs are described as very high. Hong Kong and Europe show the other side of the same constraint: facilitation and legal transparency help, but land disposal, reporting, waste-heat and sustainability KPIs become formal obligations.[CR021, CR022, CR023, CR035, CR036, CR037]

7.3 Hydrogen/SOFC, supply chain and ESG claims risk

The SOFC strategy is credible enough to monitor but too early to underwrite as a solved decarbonization pathway. DOE and Bloom support the technical proposition that fuel cells can deliver efficient onsite power, yet hydrogen’s climate value turns on production pathway and delivered fuel availability; DOE’s statement that most hydrogen is produced from natural-gas reforming makes fuel-source proof a diligence gate. DayOne’s sustainability roadmap includes renewable PPAs and 2030 targets, but the inaugural sustainability report is scheduled for 2026, so investors still need audited site-level PUE, WUE, renewable-matching, emissions and waste-heat evidence. Supplier controls reduce construction and compliance risk, but the supplier code itself shows that audits, sanctions, export-control and environmental evidence will have to scale with the build program.[CR014, CR015, CR017, CR018, CR024, CR025]

7.4 Financing, IPO and customer concentration risk

DayOne’s reported financing story is both a strength and a risk amplifier. Large private rounds, reported IPO ambitions and possible multi-billion-dollar debt facilities can fund land, substations and cooling infrastructure, but they also expose the company to rate, listing-window, disclosure and valuation-reset risk. The most decision-useful public comparator is GDS’s own filing: it documents DayOne’s standalone status and committed/billable IT power, while also showing the concentration and pricing-leverage risks inherent in hyperscale data-center models. DayOne’s exact customer concentration is not publicly disclosed in the retained sources; therefore, concentration, contract termination rights, billing ramp and top-customer credit should be treated as a diligence blocker before relying on bookings or valuation marks.[CR027, CR028, CR029, CR030, CR031, CR032]

7.5 Mitigations, monitoring indicators and kill criteria

The risk posture is manageable only if DayOne can turn company-claimed mitigations into verifiable operating evidence. The critical monitors are: Singapore SOFC fuel source and uptime; Johor grid and water approvals; Batam SEZ and cable/interconnection milestones; Chonburi and Fuchu utility and construction schedules; Hong Kong land and certification continuity; Finland EU reporting and waste-heat compliance; and group-level customer, debt and supplier disclosures. A thesis break would occur if any flagship campus loses grid allocation, if SOFC claims rely materially on fossil hydrogen without transparent emissions accounting, if debt or IPO terms imply distressed dilution, or if the top-customer mix proves too concentrated to sustain pricing power. These indicators should be refreshed before each investment committee because several are volatile and tied to 2026 financing, policy and construction events.[CR016, CR019, CR020, CR033, CR034, CR038]

7.6 Exhibits

8.1 Bottom line: track the asset, do not underwrite the headline blindly

DayOne screens as a scarce hyperscale platform, but the valuation decision is not the same as the company-quality decision. Public evidence confirms a January 2026 Series C priced at a 100 percent premium, a June 2026 final close at US$4.5 billion, and third-party reporting around a possible US$20 billion IPO valuation. That is enough to frame the opportunity, not enough to clear a buy recommendation at the headline mark. The recommendation is track / research-more with high risk and a stretched valuation stance until booked MW economics, customer concentration, debt terms, and preference stack are verified. Upside exists if DayOne converts more than 1.5GW of bookings into power-secured profitable campuses and reaches the IPO window; downside is material if execution or financing costs lag the narrative. This posture also preserves optionality while avoiding false precision around private economics not yet disclosed in public materials.[CV001, CV002, CV007, CV008, CV030, CV031]

8.2 Current valuation context and entry discipline

The January round gives the cleanest primary baseline: more than US$2.0 billion of Series C equity, led by Coatue and supported by INA, at a 100 percent premium to the prior round. The same filing points to earlier US$1.9 billion Series A/B equity and up to €1 billion of mezzanine financing, while the June close lifts total Series C gross proceeds to US$4.5 billion. Public reports then describe a potential US$20 billion IPO valuation and in some accounts a US$5 billion raise. This chapter therefore treats US$20 billion as an estimated marker, not a disclosed post-money. Entry discipline should require a discount or structure unless diligence proves bookings are binding, power is deliverable, project returns are attractive, and senior claims do not dilute common-equity upside.[CV001, CV003, CV004, CV005, CV006, CV007]

8.3 Comparable method: scarcity premium with imperfect public anchors

The relevant comp set combines private APAC hyperscale M&A, listed global data-center platforms, and market data on power scarcity. AirTrunk is the closest strategic precedent because it was an APAC hyperscale platform with more than 800MW committed to customers and land for more than 1GW of future growth; its A$24 billion transaction validates that large pools of infrastructure capital will pay for scaled, power-and-land-secured growth. Public REITs provide disclosure discipline and market-cap anchors, but Equinix, Digital Realty, Iron Mountain, and GDS are not pure DayOne twins. The method should triangulate value per booked/committed MW, quality of power and land, customer credit and concentration, construction execution, and whether public markets will accept private-growth marks.[CV015, CV016, CV017, CV018, CV019, CV020]

8.4 Scenarios and sensitivity

The bull case is not simply “AI demand remains high.” It requires DayOne to convert bookings into revenue-producing campuses, preserve access to debt and equity, deliver Finland and APAC builds on time, and approach public markets while data-center scarcity remains acute. The base case credits the platform’s momentum but treats the US$20 billion marker as fair only with strong private evidence. The bear case matters because valuation can move faster than construction: power interconnection delays, capex inflation, customer concentration, or weaker IPO appetite could pull the mark back toward the earlier US$10 billion reference. Sensitivity is highest to booked-MW conversion, delivered cost per MW, contracted yield, debt cost, and terminal market appetite for AI infrastructure.[CV010, CV013, CV014, CV025, CV028, CV029]

8.5 Exit readiness and final diligence asks

Exit readiness is credible but incomplete. Reuters/Straits Times and Bloomberg/Yahoo reporting point to US and potential Singapore listing routes, and the AirTrunk sale demonstrates strategic infrastructure appetite if IPO timing deteriorates. However, no public prospectus, audited standalone DayOne financial package, customer schedule, or detailed preference stack was retained. The IC should therefore treat a 2026 IPO as an option rather than the base underwriting exit. Before investing, diligence must obtain contracted revenue and utilization by campus, named customer concentration, capex-to-complete and interconnection schedules, debt covenants, liquidation preferences, and sensitivity to an IPO clearing below US$20 billion. Failure to produce those materials is itself a thesis-break signal.[CV036, CV037, CV038, CV040, CV041, CV042]

8.6 Exhibits