> ## Content Index > Fetch the complete content index at: https://adikumar.co/llms.txt > Use this file to discover other available public pages before exploring further. # Retrofit versus greenfield: the DC-DC decision framework (7/13) - URL: https://adikumar.co/dc-dc-transition-07-retrofit-greenfield/ - Published: 2026-08-20T19:49:56.000Z - Updated: 2026-10-07T05:39:07.000Z - Description: Retrofitting an AC facility to 800 VDC is technically possible but rarely economic. Greenfield beats retrofit on 10-year TCO by 30-50 percent under typical assumptions. Retrofit only makes sense in three narrow cases; most operators should scope a satellite greenfield instead. - Author: Adi Kumar - Tags: The DC-DC Transition, AI Infrastructure, Data Centre Power, 800V DC, #Essay [The DC-DC Transition series](https://adikumar.co/dc-dc-transition/) · Part 7 of 13 The DC-DC Transition · Part 7 of 13 # Retrofit versus greenfield: the DC-DC decision framework Retrofitting an existing AC facility to 800 VDC spans several economic categories. The retrofit depth (sidecar / hall / facility / full) changes the capex and disruption profile discontinuously. Under typical assumptions the ten-year TCO comparison materially favours greenfield; the exact spread depends on facility age, cooling readiness, workload mix and land availability. Retrofit is lower-TCO in a narrow set of cases where facility age, cooling readiness and land constraint align. For many operators, scoping a satellite greenfield alongside the existing site becomes the lower-TCO outcome under the TCO comparison. Part 7/13 in the DC-DC Transition series. Independent analysis, no advisory conflict on any named party. The DC-DC Transition · thirteen essays for data centre power architects 1. **1/13.** The real reason data centres are going DC 2. **2/13.** Two architectures wearing the same name 3. **3/13.** The architecture map 4. **4/13.** Arc behaviour + insurance 5. **5/13.** Grounding + ground-fault protection 6. **6/13.** Battery integration at 800 VDC 7. **7/13.** Retrofit vs greenfield *(you are here)* 8. **8/13.** 800 VDC and liquid cooling co-emergence 9. **9/13.** Power quality + grid interaction 10. **10/13.** Standards: OCP, IEC, NEC, IEEE 11. **11/13.** Vendor economics: who wins the transition 12. **12/13.** Commissioning, skills, operational readiness 13. **13/13.** The ten-year view ## 01Why retrofit is structurally hard An existing AC-native data centre is a set of physical constraints an 800 VDC upgrade has to work around. The dominant constraints are: the existing MV\-to-LV substation infrastructure sized for AC distribution, the cooling plant sized for the prior rack-density profile, the real-estate footprint fixed at build time, and the operational continuity requirement that constrains what can be de-energised for how long. Each of these translates to a specific retrofit cost. Rebuilding the MV-to-LV distribution to feed 800 VDC racks requires either wholesale replacement of the LV switchgear (expensive) or interposition of an AC-to-800 VDC conversion stage between the existing LV and the new DC bus (defeats much of the efficiency argument). Increasing rack density to the levels that motivate an 800 VDC architecture means where the existing facility was designed around substantially lower rack densities, cooling capacity can become the dominant retrofit constraint. As an example, a 30 kW/rack air-cooled hall targeting 120 kW/rack liquid-cooled represents a 3-4x thermal-capacity step-change. The real-estate footprint constrains whether new equipment fits at all; some retrofits require external containerised additions. And the continuity requirement means the retrofit is typically phased over 12-24 months with the facility running throughout, which multiplies commissioning cost. ## 02Retrofit capex composition Retrofit capex is dominated by the electrical distribution rebuild. Understanding this composition is what separates a well-scoped retrofit business case from a poorly-scoped one. #### Chart 1\. Retrofit capex composition per MW at 800 VDC, illustrative Electrical distribution rebuild is the dominant bucket (author scenario range: 35-45%; subject to scope of existing MV/LV reuse). Cooling plant upgrade is the second-largest bucket (author scenario range: 20-30%; depends on liquid-cooling readiness). Rack-level equipment (the visible part of the upgrade) is only 12-18 percent. Commissioning premium: the cost of executing the retrofit while the facility operates. Is 8-15 percent of total. Author scenario ranges informed by vendor engineering references and observed retrofit cost structures across author DD evaluations 2024-2026; not a statistical sample of publicly disclosed projects. The transition tax measured in Supplement A is the premium for moving from incumbent AC to 800 VDC; retrofit economics measure the additional cost of executing that transition inside an operating legacy facility. The commissioning premium is the line item most often under-estimated. Executing electrical work on an operational facility requires isolation procedures, temporary power arrangements, work-window scheduling around production maintenance, and additional safety inspection. Vendor commissioning quotes for retrofit work can run materially above the equivalent greenfield quote for the same equipment scope. ## 03Greenfield capex composition Greenfield capex has a different composition. The electrical distribution can be sized natively for 800 VDC without accommodating existing infrastructure. The cooling plant can be sized for the target rack-density profile without oversizing to accommodate legacy air-cooling. The real-estate footprint can be optimised for the target facility layout. And the commissioning happens once, on an empty building, without operational-continuity constraints. #### Chart 2\. Greenfield capex composition per MW at 800 VDC, illustrative Site preparation and shell construction dominate at 25-35 percent (land, foundations, building shell). Electrical distribution runs 20-25 percent, cheaper than retrofit because it is sized natively. Cooling plant sized to target density runs 15-20 percent. Rack-level equipment is 15-20 percent. Commissioning is 5-8 percent of total, roughly a third of retrofit commissioning as a share. Composite from published greenfield project economics, Modular Datacenter Stack essay estimates, and vendor engineering references. ## 04Ten-year TCO comparison The retrofit-vs-greenfield economics diverge at ten-year total cost of ownership even when day-one capex looks close. The TCO model assumes: 200 MW facility, 10-year hold within 20-year facility life, retrofit efficiency penalty ≈ 6 percentage points absolute end-to-end electrical efficiency, for the base-case hybrid retrofit architecture (author stress-case assumption; the resulting energy costs are model inputs, not an observed retrofit benchmark), retrofit insurance uplift ≈ 25 percent (author estimate; carrier-specific), electricity $0.08/kWh, utilisation 80 percent, discount rate 8 percent (10-year annuity factor ≈ 6.71). Illustrative auditable PV bridge per MW over 10 years (author scenario, not a market benchmark): | PV $M/MW | Retrofit | Greenfield | | ------------------------------------------- | -------- | ---------- | | Initial capex | 8.5 | 11.0 | | Energy (10y × $5.6M or $5.3M/yr, disc) | 37.6 | 35.6 | | Maintenance (10y × $0.5M or $0.4M/yr, disc) | 3.4 | 2.7 | | Insurance (10y × $0.25M or $0.20M/yr, disc) | 1.7 | 1.3 | | Temporary power during retrofit | 3.0 | 0.0 | | Operational disruption / continuity | 5.0 | 0.0 | | Stranded / early-life existing assets | 3.0 | 0.0 | | Residual value at year 10 (PV) | (2.3) | (6.6) | | **10-y PV TCO** | **60.0** | **44.0** | The retrofit continuity cost stack (temporary power, operational disruption, stranded early-life assets) is what inverts the day-one capex comparison. Once execution and stranded-asset costs are counted, the economic day-one advantage disappears. Retrofit's cumulative TCO exceeds greenfield's throughout the hold period in the base scenario. Cross-over as previously stated (year 3-4 on capex-and-operating-only comparison) is superseded by this bridge; where the retrofit continuity stack is smaller (light or medium retrofit depth, see Figure 7.1), the crossover reappears. Retrofit day-one capex per MW is typically lower than greenfield day-one capex (author scenario range: roughly 60-80 percent when comparing site-development-through-commissioning scope on both sides, excluding IT load and land), and this appears to be the reason some operators evaluate the two on capex alone and choose retrofit. That evaluation is incomplete. Adding opex. Energy losses from the retrofit's inherent inefficiencies (extra conversion stages, undersized cooling), maintenance premium on hybrid AC/DC operation, insurance premium on a facility that holds both architectures during phased retrofit. Retrofit TCO per MW over ten years lands 30-50 percent above greenfield TCO. On a capex-and-operating-only comparison excluding retrofit continuity costs, the $2.5M/MW capex gap divided by the $0.35M/MW-yr operating gap yields a \~7-year cumulative crossover. Once retrofit continuity costs are counted (see the TCO bridge above), greenfield is lower TCO from day one; retrofit does not cross over in the base scenario; day-one crossover stays unavailable. #### Chart 3\. TCO retrofit vs greenfield over ten-year hold at 800 VDC target Retrofit shows favourable day-one capex but in the base scenario, greenfield stays lower on cumulative TCO throughout the ten-year hold; retrofit's day-one capex advantage is neutralised by the continuity-cost stack (temporary power, disruption, stranded assets). The crossover can reappear in lighter retrofit scenarios where those continuity costs are materially smaller (see Fig 7.1). By year 10 greenfield TCO is 30-50 percent lower than retrofit TCO for equivalent 800 VDC capacity. The specific delta depends on the retrofit's efficiency penalty, its insurance premium, and its operational-continuity cost. Illustrative TCO model. Base case assumptions: 200 MW facility, 10-year hold, retrofit efficiency penalty 6 pp, retrofit insurance premium uplift of roughly 25 percent above equivalent greenfield (author estimate; carrier-specific). ## 05Break-even sensitivity: when retrofit makes sense The conditions under which retrofit becomes economically viable are: the facility is under 5 years old (so remaining useful life justifies the retrofit capital); the cooling plant is already liquid-cooling-ready (avoiding the largest single retrofit cost); the real-estate is constrained (greenfield alternative stays physically unavailable). When facility age, cooling readiness and land constraint align, retrofit TCO can beat greenfield-plus-decommissioning. If only two hold, the case is marginal. If only one, retrofit is almost certainly the wrong choice. #### Chart 4\. Retrofit break-even sensitivity: facility age at retrofit decision A facility retrofit at age 2 delivers positive NPV over the remaining useful life. At age 5, the calculation is marginal. At age 8, retrofit TCO exceeds greenfield-plus-decommissioning under most reasonable assumptions. The exception is where real-estate scarcity dominates, such as permit-constrained urban markets where greenfield stays physically unavailable. NPV sensitivity model with 10-year hold, 10 percent discount rate, and typical retrofit efficiency penalty. ## 06Phased retrofit: the approach that sometimes works A phased retrofit executes the upgrade one rack cluster (or row, or hall) at a time while the rest of the facility continues to operate. The advantage is operational continuity, which avoids a full shutdown and the associated revenue interruption. Phasing typically brings a timeline premium and a capex premium relative to a shutdown-and-retrofit approach (author scenario ranges: 25-40 percent timeline, 15-25 percent capex; project-specific), plus the complexity of running two architectures side-by-side during the transition period. The specific cases where phased retrofit works well are hyperscaler campuses with multiple halls (retrofit one hall at a time while others handle the load), colocation facilities with staged tenant migrations (align retrofit with tenant renewals or moves), and facilities where the retrofit is targeting only a specific rack-cluster type (AI training racks, for example, while the rest of the facility remains AC). ## 07Publicly disclosed retrofit evidence Public disclosure on data centre retrofits is thin because operators have limited incentive to publicise operational disruption. Public disclosure on 800 VDC data-centre retrofit is thin. The commentary below distinguishes documented facts (issuer press, filings) from PRIVATE DD OBSERVATIONS across evaluations 2024-2026\. Every anonymised row in the table below is a PRIVATE DD OBSERVATION. | Operator / facility | Retrofit scope | Reported outcome | Observed lesson | | ----------------------------------------------------- | ----------------------------------------------------------------------------------- | -------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------- | | **Anonymised hyperscale operator (US East Coast)** | Partial DC-native rack rows retrofit into existing AC halls | \~18-month project, on budget | Cooling plant compatibility was the enabling factor; would not have been feasible without prior liquid-cooling investment | | **Anonymised colocation operator (US Mountain West)** | Retrofit of colocation racks to 400 VDC (bipolar) for a specific hyperscaler tenant | \~24-month project, publicly reported delays and cost overruns | Phased execution added 6+ months; tenant-specific spec reduced retrofit ROI to marginal | | **Anonymised European colo operator** | Discussed retrofit; ultimately chose greenfield satellite instead | Retrofit business case did not close | Facility age (12+ years) and cooling constraint drove the greenfield decision | | **Hyperscaler-A campus (undisclosed)** | Phased 800 VDC retrofit of a 3-hall campus, one hall per year | In progress, first hall commissioned Q1 2026 | Success dependent on other halls taking on load during commissioning windows | ## 08Decision framework for operators evaluating both A useful decision framework runs three sequential screens; the final call remains a continuous TCO/NPV calculation. Gate 1: has the facility recently been built (loosely, under about five years)? If yes, proceed. If no, greenfield unless real-estate scarcity is binding. Gate 2: is the cooling plant already liquid-cooling-ready or easily converted? If yes, proceed. If no, greenfield unless the retrofit budget can take on a substantial cooling-plant rebuild. Gate 3: is there real-estate for a greenfield satellite within acceptable distance of the retrofit target? If yes, greenfield-plus-decommission-old. If no, retrofit becomes the default. FIGURE 7.2 · DECISION VARIABLES · CONTINUOUS INPUTS adikumar.co · essay 7/13 Three variables. Continuous inputs to the TCO model, not binary gates. Where the three variables align on the left, retrofit becomes defensible; where they diverge, the comparison typically inverts. FACILITY AGERETROFIT-FAVOURABLEGREENFIELD-FAVOURABLEyoung / recently builtolder / near refresh cycleCOOLING-PLANT READINESSRETROFIT-FAVOURABLEGREENFIELD-FAVOURABLEliquid-cooling capableair-cooled, needs full rebuildLAND AVAILABILITYRETROFIT-FAVOURABLEGREENFIELD-FAVOURABLEconstrained, no satellite landavailable satellite land Retrofit vs greenfield is a portfolio of continuous trade-offs, not a three-gate binary. SOURCES · author DD scenarios 2024-2026 · Schneider Electric retrofit vs new build framework · adikumar.co Figure 7.2\. Decision variables as continuous inputs. Retrofit becomes defensible where the three variables align on the retrofit-favourable side. Figure 7.1 The three-gate decision framework. The three variables are screening indicators, not gates; the final decision is the TCO/NPV calculation. Where the variables align on the retrofit side the a-priori case is stronger; each is more usefully treated as a continuous input to the TCO model than as a binary gate Gate 1: Facility age Facility under 5 years old? Remaining useful life must justify retrofit capital Gate 2: Cooling ready Liquid-cooling in place? Or convertible for <40% cooling capex Gate 3: No greenfield Real-estate constraint binding? Greenfield satellite not available within range RETROFIT defensible All three gates cleared GREENFIELD (default) Any gate fails ⇒ greenfield satellite with staged decommissioning of legacy Break-even TCO analysis favours greenfield by 30-50% over ten-year hold in most cases The three narrow retrofit-defensible cases: Young facility (<5 yrs remaining life) Liquid-cooling ready (largest single retrofit cost avoided) Real-estate constrained (no greenfield alternative) Framework synthesised from author DD observations 2024-2026 (anonymised) and DD-lens application to 12+ evaluations 2024-2026\. Author framework. Three-gate decision framework: retrofit has the strongest a-priori case when facility age is relatively recent, cooling is liquid-cooling-ready, and greenfield stays physically unavailable. Where a single condition falls short, the decision typically tips toward greenfield satellite unless the value of existing power, land, shell or operating revenue is sufficiently high with staged decommissioning of the legacy facility. #### Chart 5\. Retrofit vs greenfield decision flowchart, three gates Gate 1: facility age. Gate 2: cooling readiness. Gate 3: real-estate availability. Projects clearing all three screening conditions have the strongest a-priori case for retrofit; where one condition fails, the decision reverts to the TCO/NPV calculation with the value of existing power, land, shell and operating revenue as continuous inputs. Where the three variables diverge, the TCO comparison typically inverts toward greenfield. Framework synthesised from public retrofit case studies and DD-lens application to 12+ evaluations 2024-2026. #### Chart 6\. Illustrative per-MW all-in installed capex range: retrofit vs greenfield (some observed, some modelled) All-in installed retrofit projects cluster around $12-18M per MW (including retained/rebuilt infrastructure; wider scope than Chart 3's incremental DC-conversion base case of $8.5M/MW). Greenfield 800 VDC projects cluster around $8-12M per MW (the Chart 3 base case of $11M/MW falls inside this range). In the overlap band ($11-13M per MW), retrofit-well-executed and greenfield-poorly-executed produce similar all-in cost, and the decision-framework detail matters. Named projects from public disclosure plus author's DD engagements 2024-2026\. Specific project data anonymised where not publicly disclosed. ## 09What the model says Retrofit costs less than greenfield on capex terms, but the TCO comparison inverts across a ten-year hold in the observed cases. Operators evaluating an 800 VDC upgrade should build the ten-year TCO model. The retrofit case survives that scrutiny under specific facility conditions: relatively young age, cooling-plant compatibility with liquid-cooling and land-scarcity that rules out a satellite greenfield. Across 12+ evaluations in author DD work, greenfield satellite is the lower-TCO outcome in most of the modelled cases with staged decommissioning of the legacy facility. Part 8/13 moves into cooling, the co-emerging technology that determines whether the retrofit or greenfield decision even matters. Above 100 kW per rack, 800 VDC and direct-to-chip liquid cooling become linked architecture choices. ### Glossary of terms used 800 VDC 800 Volts Direct Current. An emerging high-voltage DC distribution class for AI data centres (one of several LVDC configurations under active OCP work: bipolar ±350/400 V, unipolar 700/800 V, TN-S and IT grounding variants) for AI-scale data centre power distribution. DD Due Diligence. The workstream discipline of testing a target's claims before committing to a transaction. LV Low Voltage. Typically below 1 kV, the class historically used for data centre power distribution below the MV substation. MV Medium Voltage. Typically 1 kV to 35 kV, the voltage class connecting data centre power distribution to the utility grid. NPV Net Present Value. Discounted cash-flow metric used to compare projects with different timing and duration. TCO Total Cost of Ownership. Capex plus opex over a defined hold period, discounted to present value. For the full corpus glossary of acronyms used across all essays, see [adikumar.co/glossary](https://adikumar.co/glossary/). Related topic hubs - [800V Data Centre Power](https://adikumar.co/800v-data-centre-power/) - [AI Power Semiconductors](https://adikumar.co/ai-power-semiconductors/) - [Browse all topic hubs →](https://adikumar.co/topics/) **Method and sources.** This essay is compiled from public information plus author DD observations across evaluations 2024-2026, distinguished throughout as PUBLIC FACT vs AUTHOR SCENARIO vs PRIVATE DD OBSERVATION. Retrofit case studies from author DD observations across evaluations 2024-2026, distinguished as PRIVATE DD OBSERVATION where specific project details stay undisclosed. Capex composition estimates from vendor engineering quotes (Vertiv, Schneider Electric, Eaton, Delta Electronics public engineering references) and Modular Datacenter Stack essay work. TCO model assumptions are the author's; specific projects require project-specific modelling. No advisory relationship with any named party. **Series footer.** Part 7/13 in *The DC-DC Transition*Related reading: Part 3/13 on the AI power chain architecture map (retrofit inventory), Part 8/13 on cooling co-emergence (cooling is the largest retrofit constraint), Part 4/13 on arc-fault protection (retrofit interacts with insurance premium), Part 11/13 on vendor economics (retrofit vs greenfield vendor mix differs). Companion context: *[The AI Power Chain](https://adikumar.co/tag/ai-power-chain/)* Part 6/13 on Modular Datacenter Stack economics, *[Due Diligence for the AI Buildout](https://adikumar.co/tag/due-diligence-for-the-ai-buildout/)* Part 8/13 on capex + growth stress-testing. Written in a personal capacity. No advisory conflict on any named party. Nothing here is investment advice. The DC-DC Transition: start with the foundation 1. [The real reason data centres are going DC (four forces)](https://adikumar.co/the-real-reason-data-centres-are-going-dc-and-its-not-capacity-alone/) 2. [Two architectures wearing the same name (800V DC vs OCP Mt Diablo)](https://adikumar.co/two-architectures-wearing-the-same-name/) 3. [The AI power chain architecture map](https://adikumar.co/the-architecture-map/) ## Frequently asked When does 800 VDC retrofit make economic sense over greenfield? Retrofit typically works when the existing facility has usable land, existing utility interconnection, and remaining structural life. Break-even depends on the depth of the retrofit (electrical distribution vs cooling vs both), the remaining useful life of the existing infrastructure and the workload mix. A phased retrofit is often more defensible than a full one-shot conversion. What proportion of 800 VDC retrofit capex is stranded from the existing facility? The stranded fraction depends on how much of the existing 480 VAC distribution, transformer capacity and BESS/UPS infrastructure can be reused. When the existing infrastructure has substantial remaining life, stranded capex can be significant; when the retrofit is timed to a natural refresh cycle, the stranded fraction can be much smaller. Can retrofit and greenfield workloads coexist on the same campus? Yes, and often should. A hall-by-hall or row-by-row hybrid architecture (per the framework map) can retain 480 VAC for legacy workloads while a new hall runs native 800 VDC. Supplement B provides the quantitative scenario case. What is the typical ten-year TCO difference between retrofit and greenfield? The TCO comparison is site- and workload-specific. Greenfield typically wins on efficiency, footprint and future flexibility; retrofit wins on land, interconnection, permitting and time-to-market when the site is suitable. The relative magnitude depends on the assumed capex ranges, ride-through duration, chemistry choice and the workload rack density. When does a phased retrofit strategy dominate a one-shot conversion? Phased retrofit is generally preferable when the operator has a mixed tenant base, a rolling refresh cycle, or protection/grounding architecture that cannot be redesigned in one shot. It also lets the operator learn from the first phase before committing capex to subsequent halls. What are the biggest hidden costs in an 800 VDC retrofit? Protection redesign (arc-fault, grounding, insulation monitoring), commissioning coordination across mixed architectures, revised insurance underwriting for the new fault topology, and the operator-side skills gap to operate a hybrid facility. The DC-DC Transition · Navigate the series - **← PREVIOUS** · [Essay 6 · Battery integration at 800 VDC](https://adikumar.co/dc-dc-transition-06-battery-integration/) - **→ NEXT** · [Essay 8 · 800 VDC and liquid cooling](https://adikumar.co/dc-dc-transition-08-cooling-co-emergence/) - **FRAMEWORK MAP** · [The AI Power Chain framework](https://adikumar.co/framework-ai-power-chain/) - **DEEP DIVE** · [Supplement A · The Transition Tax (capex decomposition)](https://adikumar.co/dc-dc-supplement-a-transition-tax/) - **DEEP DIVE** · [Supplement B · Hybrid forever (quantitative scenario)](https://adikumar.co/dc-dc-supplement-b-hybrid-forever/)