1001 Smart Cities
🇩🇰 Copenhagen

AI-Driven Demand Response for Copenhagen's District Heating

🤖 AI-generated summary of the sources below · 🔎 Last verified · ◆ 14 sourced claims, none individually verified yet

This overview was generated with AI from the public sources listed below and checked against them before publication. It has not had a complete human fact-check — treat it as a starting point and follow the sources.

How this atlas is made · What “last verified” means · What “source-checked” means · The full record behind this page

✎ Corrected 21 Aug 2026 2 claims added. What changed →

HOFOR and Danfoss use AI-based predictive control (Leanheat) to shift heat demand in Copenhagen buildings. A 2021/22 pilot on 17 municipal buildings cut the maximum peak load up to 22%, and the morning peak by 14%. A 2023-launched rollout now covers 140 buildings from 20 owners, activating about 8 MW of flexible capacity and averaging 5% energy savings.

Live 📅 since 2023 🤖 Uses AI
Outcome🎯 Succeeded
Evidence🏛️ Official
what does this mean?The strongest source is the operator's or the city's own reporting — an official report, dataset or press release. It grades who did the looking — not how well the project went. All four grades →
Run byHOFOR & Danfoss (Leanheat)
Scale🧪 Pilot

📊 Impact

2021/22 pilot cut the maximum peak heat load up to 22% across 17 municipal buildings — the morning peak specifically by 14%; the 2023 rollout covers 140 buildings/20 owners as reported in October 2025, activating ~8 MW of flexible capacity and averaging 5% energy savings.

⚡ EnergyHeat & coolingArtificial intelligenceSmart grid & energySensors & IoT

Sources

↗ Danfoss — Danfoss and HOFOR partner to maximize building energy efficiency with AI-powered Leanheat solutionOfficial Danfoss press release, 2023-11-20 — rollout announcement, target scope.
↗ DBDH Hot|Cool — From Consumer to FlexumerDanish District Heating Association trade journal, 2025-10-21 — current status: 140 buildings, 20 owners, ~8 MW flexibility, 5% savings.
↗ DBDH — Exploiting the Power of Digitalization and AI for End-to-End Optimization of the DH SystemTrade press, 2023-07-07 — 2021/22 municipal-building pilot results (14%/22% peak reduction).
Show your work What this page rests on and what it can’t tell you · all 14 claims with their sources · the forensic record

Evidence Passport · public beta 14 addressable claims, 3 source dossiers, one permanent baseline. The passport is unreviewed — an AI-assisted structural migration, not a human or expert endorsement. Every revision keeps a permanent copy under revision history.

Inspect the claims, one by one — verbatim text, sources and caveats for all 14

The text below is copied verbatim from the published record. Each claim names the source dossiers that carry it; caveats stay attached to the claim they qualify.

Claim 1 clm-ac673c56a1a19d4b descriptive
HOFOR and Danfoss use AI-based predictive control (Leanheat) to shift heat demand in Copenhagen buildings
Published field
snippet · verbatim
Mapping
mapped · no interpretation or interpolation added by this passport
What the mapping says
The DBDH trade article names all three actors and the mechanism: "HOFOR (district heating utility), Copenhagen City Properties & Procurement (Municipality's building department), and Danfoss are currently testing the potential of minimizing peak heating demand", and describes the control layer as "Leanheat's AI learns how the building's thermal mass reacts to ambient conditions and evaluates the flexibility potential based on the forecasted weather and set comfort requirements". Demand shifting is stated as "charging and discharging the thermal energy in the building mass". The Danfoss press release confirms the corporate partnership and calls the product a solution that "combines cutting-edge technology, data analytics, and artificial intelligence to optimize energy consumption".
Caveat
The Danfoss release describes the software only as combining data analytics and artificial intelligence; the words "predictive control" come from the DBDH sources, not from Danfoss.

Sources carrying this claim

  1. DBDH — Exploiting the Power of Digitalization and AI for End-to-End Optimization of the DH System
  2. Danfoss — Danfoss and HOFOR partner to maximize building energy efficiency with AI-powered Leanheat solution
Claim 2 clm-1daae3fc3ba24f4f quantitative
A 2021/22 pilot on 17 municipal buildings
Published field
snippet · verbatim
Mapping
mapped · no interpretation or interpolation added by this passport
What the mapping says
Verbatim: "The first part of the demonstration took place during the heating season 2021/2022 and included 17 municipal buildings (mainly daycare centers)." Both the season and the building count match the entry exactly.
Caveat
The source calls it the first part of a demonstration rather than a pilot, and specifies the buildings were mainly daycare centres.

Sources carrying this claim

  1. DBDH — Exploiting the Power of Digitalization and AI for End-to-End Optimization of the DH System
Claim 3 clm-875a45574140f43f quantitative
cut the maximum peak load up to 22%, and the morning peak by 14%
Published field
snippet · verbatim
Mapping
mapped · no interpretation or interpolation added by this passport
What the mapping says
The source reports two distinct reduction figures for the pilot: "the average morning peak decreased by 14%, compared with the average peak consumption before the implementation of the smart control" and "The maximum peak power has decreased from 27.5 kW/building to 21.5 kW/building (-22%)." The 22% is the second of these. The source states the basis of the comparison: "The calculation compares the highest peak during load shifting to the highest measured peak in the previous heating season at the same outdoor conditions." It also labels these "the initial results of the demonstrations for the cluster of 17 buildings".
Caveat
The 22% belongs to maximum peak power (27.5 to 21.5 kW per building); the source gives 14% for the morning peak. The entry previously merged the two.

Sources carrying this claim

  1. DBDH — Exploiting the Power of Digitalization and AI for End-to-End Optimization of the DH System
Claim 4 clm-392337653ded5929 quantitative
A 2023-launched rollout
Published field
snippet · verbatim
Mapping
mapped · no interpretation or interpolation added by this passport
What the mapping says
The press release is dated 20 November 2023 and states: "This first project is scheduled to commence during the heating season 2023/24 and run until 2026." It sets the scope as reaching "up to 700 buildings within HOFOR's network".
Caveat
The release announces a project scheduled to commence in the 2023/24 heating season; it does not report a rollout already under way in 2023, and its stated target is up to 700 buildings, not the 140 the entry reports.

Sources carrying this claim

  1. Danfoss — Danfoss and HOFOR partner to maximize building energy efficiency with AI-powered Leanheat solution
Claim 5 clm-8c4e19c8d3cc5c39 quantitative
now covers 140 buildings from 20 owners
Published field
snippet · verbatim
Mapping
mapped · no interpretation or interpolation added by this passport
What the mapping says
Verbatim: "So far, the project has proven quite successful, being able to activate up to around 8 MW of flexibility through voluntary agreements with 20 large building owners and 140 buildings." Both counts match the entry.
Caveat
This 2025 article describes HOFOR's flexumer proof-of-concept; neither it nor the Danfoss release states that this is the same project the 2023 release announced.

Sources carrying this claim

  1. DBDH Hot|Cool — From Consumer to Flexumer
Claim 6 clm-c8b1fee4ae8bf46b quantitative
activating about 8 MW of flexible capacity
Published field
snippet · verbatim
Mapping
mapped · no interpretation or interpolation added by this passport
What the mapping says
Verbatim: "being able to activate up to around 8 MW of flexibility through voluntary agreements". The article adds that the system benefits are to be tested "during the heating season of 2025/2026".
Caveat
The source says up to around 8 MW can be activated — a ceiling of available flexibility, not a measured average of capacity actually dispatched.

Sources carrying this claim

  1. DBDH Hot|Cool — From Consumer to Flexumer
Claim 7 clm-272b7bbb3caabeeb quantitative
averaging 5% energy savings
Published field
snippet · verbatim
Mapping
mapped · no interpretation or interpolation added by this passport
What the mapping says
The 2025 trade article reports a realised figure: "they have, on average, realized a 5 % energy savings since the beginning of 2025 by accessing an online predictive control system for their district heating substations, utilizing weather forecasts as part of the POC." The 2023 Danfoss release had forecast the same number in advance, as "anticipated savings of an average of 5 %".
Caveat
Both figures are the operators' own reporting; no independent evaluation is cited. The Danfoss number is an expectation stated before the project began, the DBDH number a self-reported result covering only 2025.

Sources carrying this claim

  1. DBDH Hot|Cool — From Consumer to Flexumer
  2. Danfoss — Danfoss and HOFOR partner to maximize building energy efficiency with AI-powered Leanheat solution
Claim 8 clm-40c1b72698a3e06b quantitative
2021/22 pilot cut the maximum peak heat load up to 22%
Published field
impact · verbatim
Mapping
mapped · no interpretation or interpolation added by this passport
What the mapping says
The source dates the demonstration to "the heating season 2021/2022" and reports "The maximum peak power has decreased from 27.5 kW/building to 21.5 kW/building (-22%)", alongside a separate figure that "the average morning peak decreased by 14%". The source states the basis of the comparison: "The calculation compares the highest peak during load shifting to the highest measured peak in the previous heating season at the same outdoor conditions." It also labels these "the initial results of the demonstrations for the cluster of 17 buildings".
Caveat
The 22% belongs to maximum peak power (27.5 to 21.5 kW per building); the entry no longer attributes it to the morning peak.

Sources carrying this claim

  1. DBDH — Exploiting the Power of Digitalization and AI for End-to-End Optimization of the DH System
Claim 9 clm-2143e5e25f29451b quantitative
across 17 municipal buildings
Published field
impact · verbatim
Mapping
mapped · no interpretation or interpolation added by this passport
What the mapping says
Verbatim: "The first part of the demonstration took place during the heating season 2021/2022 and included 17 municipal buildings (mainly daycare centers)."
Caveat
No claim-specific caveat recorded

Sources carrying this claim

  1. DBDH — Exploiting the Power of Digitalization and AI for End-to-End Optimization of the DH System
Claim 10 clm-e4ff103b3b55c2b4 quantitative
the 2023 rollout
Published field
impact · verbatim
Mapping
mapped · no interpretation or interpolation added by this passport
What the mapping says
Danfoss announced the HOFOR partnership on 20 November 2023, stating that "This first project is scheduled to commence during the heating season 2023/24 and run until 2026" and would reach "up to 700 buildings within HOFOR's network".
Caveat
2023 is the year of the announcement; the release schedules the start for the 2023/24 heating season and names a target of up to 700 buildings.

Sources carrying this claim

  1. Danfoss — Danfoss and HOFOR partner to maximize building energy efficiency with AI-powered Leanheat solution
Claim 11 clm-1efa0f1e06909adf quantitative
covers 140 buildings/20 owners as reported in October 2025
Published field
impact · verbatim
Mapping
mapped · no interpretation or interpolation added by this passport
What the mapping says
The article, published 21/10/2025, states: "So far, the project has proven quite successful, being able to activate up to around 8 MW of flexibility through voluntary agreements with 20 large building owners and 140 buildings."
Caveat
October 2025 is the article's publication date, not an as-of date stated in the text; the article says only "so far". It also describes the arrangement as a proof-of-concept, with system benefits still to be tested in the 2025/2026 heating season.

Sources carrying this claim

  1. DBDH Hot|Cool — From Consumer to Flexumer
Claim 12 clm-26e8e7017876f064 quantitative
activating ~8 MW of flexible capacity
Published field
impact · verbatim
Mapping
mapped · no interpretation or interpolation added by this passport
What the mapping says
Verbatim: "being able to activate up to around 8 MW of flexibility through voluntary agreements with 20 large building owners and 140 buildings."
Caveat
The source states an upper bound of activatable flexibility, not a measured quantity dispatched.

Sources carrying this claim

  1. DBDH Hot|Cool — From Consumer to Flexumer
Claim 13 clm-8b1386632baf03e6 quantitative
averaging 5% energy savings
Published field
impact · verbatim
Mapping
mapped · no interpretation or interpolation added by this passport
What the mapping says
Verbatim from the 2025 article: "they have, on average, realized a 5 % energy savings since the beginning of 2025 by accessing an online predictive control system for their district heating substations, utilizing weather forecasts as part of the POC." The 2023 Danfoss release names the same figure prospectively: "anticipated savings of an average of 5 %".
Caveat
Operator-reported in both cases, with no independent evaluation cited; the DBDH figure covers only the period since the beginning of 2025, and the Danfoss figure is an expectation stated before the project started.

Sources carrying this claim

  1. DBDH Hot|Cool — From Consumer to Flexumer
  2. Danfoss — Danfoss and HOFOR partner to maximize building energy efficiency with AI-powered Leanheat solution
Claim 14 clm-aeacbd15407fa867 quantitative
the morning peak specifically by 14%
Published field
impact · verbatim
Mapping
mapped · no interpretation or interpolation added by this passport
What the mapping says
The same source gives 14% for the morning peak specifically, next to the 22% for maximum peak power — the two figures the entry previously conflated.
Caveat
No claim-specific caveat recorded

Sources carrying this claim

  1. DBDH — Exploiting the Power of Digitalization and AI for End-to-End Optimization of the DH System
Forensic record — source dossiers, gaps, hashes and the exact revision this page renders

This layer records what is known about each source, what is still missing, and the exact snapshot this page is rendering. A missing date, archive or hash stays visible as missing.

unreviewed Revision 2026-08-21-unsourced-text-correction effective 2026-07-23 · created 2026-08-21

Awaiting operator and domain-expert review; publication of this revision implies neither.

sha256:c7b507017d8b8df8b6ee08d3430d01073ed8ff44eb2a0da5f1af772774c572c6
Revision history → Permanent version link → Machine-readable record → Evidence Constitution →
14
of 14 claims mapped
3
source dossiers
0
publication dates recorded
0
archives linked
0
exact locators
0
content captures hashed

Governance facts carried by the record

Operator
HOFOR & Danfoss (Leanheat)
Actor types
utility, private
Funding model
Not recorded
Funding description
Not recorded
Cost
Not recorded
Ownership
Not recorded
Decision rights
Not recorded
Exit conditions
Not recorded

3 source dossiers

Danfoss — Danfoss and HOFOR partner to maximize building energy efficiency with AI-powered Leanheat solution danfoss.com · 5 linked claims

Open original source ↗

Stable source ID
src-e3bdeb3de90cf2c0
Publisher
Name not classified · danfoss.com
Type / language
web-page · language not classified
Published
Not recorded
Updated
Not recorded
Source last checked
No source-level check date
Record last verified
2026-07-23
Dossier created
2026-08-21
Retrieved copy
No local capture recorded
Archive
No archived copy linked
Content hash
No captured content hash
Locator
Open the recorded source note

Official Danfoss press release, 2023-11-20 — rollout announcement, target scope.

Descriptive only — not page/paragraph exact
Claims linked
DBDH Hot|Cool — From Consumer to Flexumer dbdh.org · 6 linked claims

Open original source ↗

Stable source ID
src-9b8d2fb4ad1ee716
Publisher
Name not classified · dbdh.org
Type / language
web-page · language not classified
Published
Not recorded
Updated
Not recorded
Source last checked
No source-level check date
Record last verified
2026-07-23
Dossier created
2026-08-21
Retrieved copy
No local capture recorded
Archive
No archived copy linked
Content hash
No captured content hash
Locator
Open the recorded source note

Danish District Heating Association trade journal, 2025-10-21 — current status: 140 buildings, 20 owners, ~8 MW flexibility, 5% savings.

Descriptive only — not page/paragraph exact
Claims linked
DBDH — Exploiting the Power of Digitalization and AI for End-to-End Optimization of the DH System dbdh.org · 6 linked claims

Open original source ↗

Stable source ID
src-de805229232d7a15
Publisher
Name not classified · dbdh.org
Type / language
web-page · language not classified
Published
Not recorded
Updated
Not recorded
Source last checked
No source-level check date
Record last verified
2026-07-23
Dossier created
2026-08-21
Retrieved copy
No local capture recorded
Archive
No archived copy linked
Content hash
No captured content hash
Locator
Open the recorded source note

Trade press, 2023-07-07 — 2021/22 municipal-building pilot results (14%/22% peak reduction).

Descriptive only — not page/paragraph exact
Claims linked
Known limits of this passport
  • Publisher names and source languages have not yet been classified; domains are recorded without guessing.
  • Existing source notes are descriptive locators, not page- or paragraph-exact citations.
  • A missing archive, publication date or content hash is shown as null rather than inferred.
  • The project-level evidence grade ranks the strongest source in the entry; it is not a truth score for every claim.

Continue exploring

Compare this project

🇫🇮 Katri Vala Heat Pump Plant · Helsinki Leanheat's AI shifts when Copenhagen buildings draw heat; Katri Vala instead recovers heat from wastewater since 2006. 🇮🇹 Milan District Heating & Smart Network Monitoring (A2A) · Milan Adding data to district heating: Copenhagen's Leanheat AI shifts building demand; Milan maps its 390 km network by aerial thermography. 🇨🇳 Shanghai's Virtual Power Plant demand response · Shanghai Demand flexibility in two energies: Leanheat shifts when buildings draw heat; Shanghai's virtual plant shifts when they draw power.

Explore the wider context

📟 A Reading Is Not a Decision🌡️ Heat & cooling guide🏙️ Copenhagen city profile⚡ Energy projects🤖 AI in Action📖 District heating📖 Adaptive traffic signals📖 Prepaid metering
📎 Cite this project

1001 Smart Cities (2026). “AI-Driven Demand Response for Copenhagen's District Heating” — Copenhagen, Denmark. The Smart City Atlas. https://1001smartcities.org/projects/copenhagen-leanheat-ai-heating/ (last verified 2026-07-23). Data: CC BY 4.0.

The underlying data is free to reuse with attribution — see the open data page.

🚩 Report an error · 🔔 Follow changes to this entry (RSS)

✉️ Get the next source-checked smart-city project in your inbox

One short email a week: the Landmark, one Insight, the new cities. No spam, one-click unsubscribe. About the newsletter · RSS