The same optic and assumed quantity carry different modeled energy costs in Sweden and Singapore.
For the same Cisco optics and quantity, five-year electricity cost is $618,047 under the Sweden profile and $1,803,607 under Singapore; tariff and PUE both change.
The modeled customer context
The whole demo estate snapshot is separate from this case. The optic match is on the SYNTHETIC DEMO London DC1 leaf; the larger regional quantity is a declared scale assumption, not a measured installation. Snapshot: October 19, 2026.
Same optic and assumed count, different regional energy cost.
The controlled comparison fixes the product, quantity, and active period. Sweden and Singapore use different sourced tariff and PUE references, so neither input alone explains the difference. These are scenario outputs, not measured bills or a relocation saving.
- Sweden, five years
- $618,046.80
- Singapore, five years
- $1,803,607.10
Every optics energy model, in one view
All 36 outputs are separate alternatives, sorted by their individual first-five-operating-year cost. Each bar stacks five annual totals from its own native operating period. First five operating years: 2026-2030 for 2026-tariff scenarios; 2030-2034 for 2030-tariff scenarios. Cross-cohort values are illustrative scenario comparisons across different calendar periods, not controlled tariff savings. Product, quantity, country, and PUE can differ across the wider grid.
Each bar is one independent output and stacks its first five operating years. First five operating years: 2026-2030 for 2026-tariff scenarios; 2030-2034 for 2030-tariff scenarios. Cross-cohort values are illustrative scenario comparisons across different calendar periods, not controlled tariff savings. Products, quantities, countries, and PUE can also differ; do not add the bars into an estate budget.
View chart data as a table
| Scenario | Operating year 1 (USD) | Operating year 2 (USD) | Operating year 3 (USD) | Operating year 4 (USD) | Operating year 5 (USD) |
|---|---|---|---|---|---|
| Sweden · Arista QSFP-100G-LR4 · 2,000 optics | 12955.95 | 12955.95 | 12955.95 | 12955.95 | 12955.95 |
| Netherlands · Arista QSFP-100G-SR4 · 2,000 optics · 2030 start | 21987 | 21987 | 21987 | 21987 | 21987 |
| United States · Cisco QSFP-100G-SR4-S · 5,000 optics · 2030 start | 27412.5 | 27412.5 | 27412.5 | 27412.5 | 27412.5 |
| Sweden · Arista QDD-400G-LR4 · 2,000 optics · 2030 start | 34549.2 | 34549.2 | 34549.2 | 34549.2 | 34549.2 |
| Ireland · Arista QSFP-100G-FR · 2,000 optics · 2030 start | 37130.4 | 37130.4 | 37130.4 | 37130.4 | 37130.4 |
| United Kingdom · Cisco QSFP-100G-LR4-S · 2,000 optics | 43842.56 | 43842.56 | 43842.56 | 43842.56 | 43842.56 |
| Sweden · Cisco QSFP-100G-DR-S · 10,000 optics · 2030 start | 61804.68 | 61804.68 | 61804.68 | 61804.68 | 61804.68 |
| United States · Cisco QSFP-100G-LR4-S · 10,000 optics · 2030 start | 62602.5 | 62602.5 | 62602.5 | 62602.5 | 62602.5 |
| United States · Arista QSFP-100G-DR · 10,000 optics · 2030 start | 70380 | 70380 | 70380 | 70380 | 70380 |
| Singapore · Arista LPO-800G-DR8 · 2,000 optics | 75429.88 | 75429.88 | 75429.88 | 75429.88 | 75429.88 |
| Netherlands · Arista QDD-400G-FR4 · 2,000 optics | 75802.8 | 75802.8 | 75802.8 | 75802.8 | 75802.8 |
| Germany · Cisco QDD-400G-FR4-S · 2,000 optics | 79212.84 | 79212.84 | 79212.84 | 79212.84 | 79212.84 |
| Singapore · Cisco QSFP-100G-DR-S · 5,000 optics | 90238.2 | 90238.2 | 90238.2 | 90238.2 | 90238.2 |
| United Kingdom · Cisco QSFP-100G-SR4-S · 5,000 optics | 96248.12 | 96248.12 | 96248.12 | 96248.12 | 96248.12 |
| Sweden · Cisco QSFP-100G-FR-S · 20,000 optics | 123609.36 | 123609.36 | 123609.36 | 123609.36 | 123609.36 |
| United Kingdom · Cisco QDD-400G-DR4-S · 2,000 optics · 2030 start | 131870.2 | 131870.2 | 131870.2 | 131870.2 | 131870.2 |
| Netherlands · Arista QSFP-100G-FR · 10,000 optics · 2030 start | 141973.2 | 141973.2 | 141973.2 | 141973.2 | 141973.2 |
| Netherlands · Arista QSFP-100G-DR · 10,000 optics · 2030 start | 141973.2 | 141973.2 | 141973.2 | 141973.2 | 141973.2 |
| Ireland · Arista QSFP-100G-SR4 · 10,000 optics | 144653.85 | 144653.85 | 144653.85 | 144653.85 | 144653.85 |
| United States · Cisco QDD-400G-SR8-S · 10,000 optics | 187680 | 187680 | 187680 | 187680 | 187680 |
| Netherlands · Arista QDD-400G-SR8 · 5,000 optics | 189297.6 | 189297.6 | 189297.6 | 189297.6 | 189297.6 |
| Sweden · Cisco OSFP-2X400G-FR4 · 10,000 optics · 2030 start | 189732.69 | 189732.69 | 189732.69 | 189732.69 | 189732.69 |
| Germany · Cisco QDD-400G-FR4-S · 5,000 optics | 197813.28 | 197813.28 | 197813.28 | 197813.28 | 197813.28 |
| Germany · Cisco OSFP-800G-DR8 · 5,000 optics | 234137.4 | 234137.4 | 234137.4 | 234137.4 | 234137.4 |
| Germany · Cisco OSFP-800G-DR8 · 5,000 optics · 2030 start | 234137.4 | 234137.4 | 234137.4 | 234137.4 | 234137.4 |
| United Kingdom · Arista QDD-400G-DR4 · 5,000 optics · 2030 start | 329504.24 | 329504.24 | 329504.24 | 329504.24 | 329504.24 |
| Singapore · Cisco QSFP-100G-FR-S · 20,000 optics | 360721.42 | 360721.42 | 360721.42 | 360721.42 | 360721.42 |
| United Kingdom · Cisco OSFP-2X400G-FR4 · 5,000 optics | 362386.16 | 362386.16 | 362386.16 | 362386.16 | 362386.16 |
| Ireland · Arista QSFP-100G-LR4 · 20,000 optics · 2030 start | 372077.55 | 372077.55 | 372077.55 | 372077.55 | 372077.55 |
| United States · Arista QDD-800G-2FR4 · 20,000 optics | 500310 | 500310 | 500310 | 500310 | 500310 |
| United Kingdom · Cisco QDD-400G-DR4-S · 10,000 optics · 2030 start | 659008.48 | 659008.48 | 659008.48 | 659008.48 | 659008.48 |
| Germany · Cisco QDD-400G-LR4-S · 20,000 optics | 791253.12 | 791253.12 | 791253.12 | 791253.12 | 791253.12 |
| Ireland · Arista QDD-400G-SR8 · 20,000 optics | 991948.95 | 991948.95 | 991948.95 | 991948.95 | 991948.95 |
| Ireland · Cisco QDD-400G-SR8-S · 20,000 optics · 2030 start | 991948.95 | 991948.95 | 991948.95 | 991948.95 | 991948.95 |
| Singapore · Arista QDD-400G-DR4 · 20,000 optics · 2030 start | 1006965.76 | 1006965.76 | 1006965.76 | 1006965.76 | 1006965.76 |
| Singapore · Cisco QDD-400G-LR4-S · 20,000 optics · 2030 start | 1006965.76 | 1006965.76 | 1006965.76 | 1006965.76 | 1006965.76 |
One bar is one scenario. Cross-cohort bar lengths are illustrative, not controlled tariff savings; do not add them into an estate energy budget.
One optic and count across two countries
The same Cisco QSFP-100G-FR-S, 20,000 optics, and active period. Tariff and PUE both change.
View chart data as a table
| Scenario | Modeled energy cost (USD) |
|---|---|
| Sweden · Cisco QSFP-100G-FR-S · 20,000 optics | 618046.8 |
| Singapore · Cisco QSFP-100G-FR-S · 20,000 optics | 1803607.1 |
The same assumed 20,000 Cisco QSFP-100G-FR-S optics cost $618,046.80 in Sweden and $1,803,607.10 in Singapore over five years. Both tariff and PUE change.
Explain how we got here
Assumptions
- Product, 20,000-optic quantity, and active period are held fixed. The quantity is SYNTHETIC DEMO scale, not observed inventory.
- The Sweden and Singapore tariffs are published reference inputs, not contracted data-center prices. PUE references are historical regional inputs. Tariff and PUE vary together, so this does not isolate a tariff effect.
- The synthetic London DC1 leaf supplies the matched product anchor. Five-year Netos outputs are USD and have zero native FX mismatch.
- The non-USD published tariff inputs were converted to USD using the dated Bank of England series recorded in the extract.
- The Sweden tariff uses the latest published annual DESNZ reference held flat. Singapore uses the latest complete EMA annual low-tension reference held flat; it is for non-contestable consumers, not a data-center contract. Historical PUE references are modeled inputs.
Sources
- Netos matched optic inventory
- Cisco QSFP-100G-FR-S power specification
- DESNZ Sweden tariff workbook
- EMA Singapore tariff workbook
- JRC Nordic PUE reference
- NEA Singapore PUE reference
- Bank of England XUDLUSS
- Bank of England XUDLSGS
- Energy case sources and assumptions dashboard
Model evidence
Equal optic count, different form factors
These form factors differ in capacity and reach; the energy difference is not a replacement saving.
View chart data as a table
| Scenario | Modeled energy cost (USD) |
|---|---|
| Cisco OSFP-2X400G-FR4 | 1811930.8 |
| Cisco QSFP-100G-SR4-S | 481240.6 |
At an assumed 5,000 optics in the United Kingdom, Cisco OSFP-2X400G-FR4 models $1,811,930.80 and QSFP-100G-SR4-S models $481,240.60 over five years. Their capacity and reach differ, so the lower energy cost is not a replacement saving.
Explain how we got here
Assumptions
- The two cases fix the country, assumed optic count, PUE, and active period. Product power differs.
- The 5,000-optic count is SYNTHETIC DEMO scale on the London DC1 matched leaf. A technical requirements review is needed before treating these optics as alternatives.
- Datasheet watts are manufacturer specifications; modeled costs depend on the declared continuous-duty and facility assumptions.
Sources
- Netos matched optic inventory
- Cisco OSFP datasheet
- Cisco QSFP datasheet
- Energy case sources and assumptions dashboard
Model evidence
What the assumed quantity changes
Same product, country, PUE, and active period; quantity is a SYNTHETIC DEMO scale assumption.
View chart data as a table
| Scenario | Modeled energy cost (USD) |
|---|---|
| 2,000 assumed optics | 396064.2 |
| 5,000 assumed optics | 989066.4 |
For Cisco QDD-400G-FR4-S in Germany, the separate 2,000-optic and 5,000-optic scenarios show $396,064.20 and $989,066.40 over five years. Those quantities are declared assumptions, not observed installations.
Explain how we got here
Assumptions
- The selected pair has the same Cisco product, Germany tariff, PUE, and active period. Only the assumed quantity changes.
- The 2,000 and 5,000 quantities are SYNTHETIC DEMO scale assumptions on the London DC1 leaf, not a matched hardware count.
- Each output has its own financial and TBM record. The native TBM cost pool classifies one output at a time; classifications and scenarios must not be added.
Sources
- Netos matched optic inventory
- Cisco 400G optic specification
- Published regional tariff input
- Energy case sources and assumptions dashboard
- Energy model drill-down dashboard