Data story: Real-world energy savings from smart building deployments — what the data shows

Why It Matters

Vendors routinely promise 20 to 40 percent energy savings from smart building technologies, yet a 2025 meta-analysis by Lawrence Berkeley National Laboratory found that the median verified savings from building automation and control system (BACS) deployments across 312 commercial buildings was just 18 percent, with a wide interquartile range of 9 to 28 percent (LBNL, 2025). That gap between marketing claims and measured outcomes has real consequences: building owners who budget for 30 percent savings and achieve 12 percent face payback periods that stretch from five years to twelve, undermining the business case for decarbonization. Buildings consume approximately 30 percent of global final energy and produce 26 percent of energy-related CO2 emissions (IEA, 2025). Improving operational efficiency through automation, analytics, and controls is one of the fastest, most cost-effective decarbonization pathways available. But the data reveals that success depends less on the technology selected and more on how it is deployed, commissioned, and maintained. Understanding what separates high-performing smart building deployments from underperforming ones is essential for facility managers, real estate investors, and policymakers setting performance standards.

Key Concepts

Building automation and control systems (BACS). BACS encompass hardware and software that monitor and control HVAC, lighting, plug loads, and other building systems. The European standard EN 15232 classifies BACS into four efficiency classes: Class D (no automation), Class C (standard), Class B (advanced), and Class A (high-performance). Moving from Class C to Class A typically yields 20 to 30 percent HVAC energy savings in commercial buildings according to the standard's reference calculations, though real-world results vary significantly (CEN, 2024).

Fault detection and diagnostics (FDD). FDD software continuously monitors building system performance data to identify operational faults such as stuck dampers, simultaneous heating and cooling, and sensor drift. A 2024 study by the Pacific Northwest National Laboratory analyzed FDD deployments across 45 U.S. commercial buildings and found that addressing identified faults saved an average of 14 percent of HVAC energy, with the top quartile of buildings achieving 22 percent savings (PNNL, 2024). Critically, 60 percent of identified faults in the study had been present for more than six months before detection, indicating that manual inspections miss persistent inefficiencies.

AI-driven optimization. Machine learning models that predict thermal loads, occupancy patterns, and weather impacts can optimize setpoints and equipment staging beyond what rule-based controls achieve. Google DeepMind's collaboration with JLL reported 15 to 20 percent cooling energy reductions in data center and office environments using reinforcement learning algorithms (DeepMind, 2025). However, these results come from buildings with high-quality sensor networks and clean data pipelines, conditions not present in most existing commercial stock.

The performance gap. The difference between designed (or vendor-claimed) energy performance and actual measured performance is well documented in building science. The UK's Building Performance Evaluation programme found that operational energy use in new non-domestic buildings exceeded design predictions by an average of 3.5 times (CIBSE, 2024). Smart building technologies can close this gap, but only when paired with robust commissioning, ongoing monitoring, and operational adjustments. Technology alone does not guarantee results.

Measurement and verification (M&V). The International Performance Measurement and Verification Protocol (IPMVP) provides standardized methods for quantifying energy savings. Option C (whole-building metering with regression analysis) is most commonly used for smart building projects, requiring at least 12 months of baseline and post-intervention data. Without rigorous M&V, reported savings often conflate weather normalization errors, occupancy changes, and technology effects.

What's Working and What Isn't

Lighting controls deliver consistent results. Among all smart building subsystems, automated lighting controls show the most reliable savings. A 2025 analysis of 180 commercial buildings by the U.S. Department of Energy found that occupancy-sensor and daylight-harvesting systems reduced lighting energy by a median of 35 percent, with relatively tight variance (28 to 42 percent across the interquartile range) (DOE, 2025). Lighting controls benefit from straightforward logic (occupied vs. unoccupied), minimal interaction effects with other systems, and simple verification through sub-metering.

HVAC optimization savings are real but variable. HVAC systems account for 40 to 60 percent of commercial building energy use, making them the primary target for smart controls. The LBNL meta-analysis (2025) found median HVAC savings of 16 percent from advanced BACS deployments, but the range extended from 3 percent (poorly commissioned systems) to 38 percent (fully integrated, continuously optimized systems). The strongest predictor of savings was not the technology brand but the presence of a dedicated analytics team reviewing system performance at least monthly.

Vendor claims systematically overstate savings. A 2024 analysis by the Rocky Mountain Institute compared vendor-projected savings with independently verified results across 85 smart building retrofit projects in North America. Vendor projections averaged 28 percent whole-building savings; independently verified savings averaged 17 percent, a systematic overstatement of roughly 40 percent (RMI, 2024). The primary sources of overstatement were failure to account for interactive effects between systems, optimistic occupancy assumptions, and exclusion of plug loads from savings calculations.

Commissioning quality is the strongest predictor of success. Buildings that underwent enhanced commissioning (a systematic process of verifying that all systems perform according to design intent) achieved 25 percent higher savings than those with standard commissioning, and 60 percent higher savings than those with no formal commissioning process (PNNL, 2024). Retro-commissioning of existing buildings with newly installed smart systems is particularly effective, as it identifies and corrects pre-existing faults that would otherwise mask automation benefits.

Data quality and integration challenges persist. Many buildings deploy smart sensors and analytics platforms on top of legacy building management systems with proprietary protocols, inconsistent naming conventions, and missing metadata. A 2025 survey by Verdantix found that 47 percent of facility managers cited data integration as their primary barrier to realizing smart building value, ahead of cost (32 percent) and organizational resistance (21 percent) (Verdantix, 2025). Buildings with open protocols such as BACnet, Haystack-standardized tagging, and unified data models achieved measurably better outcomes.

Persistence of savings declines without ongoing attention. Initial savings from smart building deployments often erode over time as setpoints drift, overrides accumulate, and equipment degrades. The LBNL study found that buildings without ongoing monitoring-based commissioning lost an average of 30 percent of their first-year savings within three years. By contrast, buildings with continuous commissioning programs maintained 85 to 95 percent of initial savings over five years (LBNL, 2025).

Key Players

Established Leaders

• Siemens Smart Infrastructure — Building automation and digital twin platform (Desigo CC) deployed across 10,000+ commercial buildings globally, with integrated FDD and energy optimization.
• Johnson Controls — OpenBlue digital platform combining AI-driven HVAC optimization, FDD, and sustainability reporting; serves over 200,000 buildings worldwide.
• Honeywell Building Technologies — Forge enterprise performance management platform offering AI-based optimization, cybersecurity, and predictive maintenance for large portfolios.
• Schneider Electric — EcoStruxure Building platform integrating IoT sensors, advanced analytics, and power management across commercial and industrial portfolios.

Emerging Startups

• Brainbox AI — Autonomous HVAC optimization using deep reinforcement learning; independently verified 15 to 25 percent HVAC savings across 200+ buildings in North America and Europe.
• Passiv Energy — Cloud-based HVAC optimization for commercial real estate, specializing in portfolio-scale deployment with standardized M&V reporting.
• Switch Automation — Smart building analytics platform providing FDD, energy benchmarking, and ESG reporting for institutional real estate portfolios.
• 75F — IoT-native building automation system combining wireless sensors with cloud-based optimization, targeting mid-market commercial buildings underserved by legacy BMS providers.

Key Investors/Funders

• Building Technologies Office, U.S. DOE — Federal funding for smart building research, M&V protocols, and grid-interactive efficient building pilots.
• Breakthrough Energy Ventures — Investing in next-generation building automation and AI optimization startups.
• JLL Spark — Corporate venture arm of JLL investing in proptech and smart building technologies including Brainbox AI and other analytics platforms.

Examples

Empire State Building Realty Trust, New York (Johnson Controls / JLL). The Empire State Building's ongoing energy optimization program, which combines a Johnson Controls Metasys BMS with JLL's analytics platform, has reduced energy consumption by 40 percent from a 2009 baseline through 2025. The smart building layer, added in 2021, contributes approximately 12 percentage points of that total through AI-optimized chiller staging, demand-controlled ventilation, and predictive scheduling. Critically, the building maintains a full-time energy management team that reviews analytics dashboards weekly and acts on FDD alerts within 48 hours. Independently verified by IPMVP Option C, the smart systems deliver annual savings of $4.4 million against a $2.8 million cumulative technology investment (JLL, 2025). The project demonstrates that mature buildings with engaged operations teams can achieve and sustain savings at the high end of industry benchmarks.

British Land Portfolio, United Kingdom (Siemens / Demand Logic). British Land, one of the UK's largest commercial REITs, deployed Demand Logic's analytics platform across 25 million square feet of office and retail space starting in 2022. By 2025, the platform had identified over 8,000 operational faults across the portfolio, with an average resolution time of 11 days. Verified savings across the portfolio averaged 19 percent of HVAC energy, with individual buildings ranging from 8 to 31 percent. The strongest performers were buildings where Demand Logic data was reviewed in weekly operations meetings with a clear escalation protocol. British Land reported that the analytics deployment contributed to a portfolio-level energy intensity reduction from 142 to 108 kWh/m² between 2022 and 2025, supporting its 2030 net-zero pathway (British Land, 2025).

Brookfield Properties, Global Portfolio (Brainbox AI). Brookfield Properties piloted Brainbox AI's autonomous HVAC optimization across 12 commercial buildings in Toronto, New York, and London between 2023 and 2025. The system uses a deep reinforcement learning model trained on each building's unique thermal behavior, weather patterns, and occupancy data to adjust HVAC setpoints every five minutes without human intervention. An independent M&V assessment conducted by WSP using IPMVP Option C found median HVAC savings of 21 percent across the 12 buildings, with a range of 14 to 29 percent. Tenant comfort complaints decreased by 18 percent during the pilot period. Brookfield has since expanded the deployment to 150 buildings globally, making it one of the largest verified AI-driven HVAC optimization rollouts in commercial real estate (Brainbox AI, 2025).

Singapore Government Office Buildings (GovTech / Schneider Electric). Singapore's Government Technology Agency deployed Schneider Electric's EcoStruxure platform across 28 government office buildings as part of the national Green Government Buildings programme. Between 2023 and 2025, the platform reduced portfolio energy use intensity from 195 to 152 kWh/m², a 22 percent reduction verified through the Building and Construction Authority's Green Mark framework. The deployment integrated IoT sensors, AI-driven chiller plant optimization, and demand-controlled ventilation with a centralized operations dashboard monitored by GovTech's smart facilities team. The program's success was attributed to standardized deployment protocols, centralized monitoring, and contractual performance guarantees tied to verified energy savings (BCA Singapore, 2025).

Action Checklist

• Establish a minimum 12-month energy baseline with calibrated sub-meters before deploying smart building technologies, following IPMVP Option C methodology.
• Require vendors to commit to independently verified savings targets rather than accepting marketing claims; include performance guarantees in procurement contracts.
• Invest in enhanced commissioning for all new smart building deployments and retro-commissioning for retrofit projects, budgeting 2 to 5 percent of project cost for commissioning activities.
• Standardize building data using open protocols (BACnet, Project Haystack tagging) and ensure clean metadata before deploying analytics platforms.
• Assign dedicated personnel or contracted services for ongoing monitoring-based commissioning, with weekly review cycles and clear fault escalation protocols.
• Deploy FDD software as a minimum viable smart building intervention; it offers the highest verified savings-to-cost ratio and builds the operational discipline needed for more advanced optimization.
• Track and report savings persistence over multi-year periods, not just first-year results, to ensure technology investments deliver sustained returns.

FAQ

What is a realistic energy savings expectation from a smart building deployment? Based on the LBNL 2025 meta-analysis of 312 buildings, a well-executed smart building deployment (advanced BACS with FDD and ongoing commissioning) typically delivers 15 to 25 percent whole-building energy savings. Lighting controls tend to deliver 28 to 42 percent reductions in lighting energy, while HVAC optimization yields 10 to 25 percent of HVAC energy. Whole-building savings are lower because plug loads and process energy are generally unaffected by building automation. Achieving the upper end of these ranges requires high-quality sensor data, enhanced commissioning, and dedicated operational review.

Why do vendor claims often overstate savings? The Rocky Mountain Institute's 2024 analysis identified three primary drivers. First, vendors frequently calculate savings against a hypothetical worst-case baseline rather than actual pre-intervention performance. Second, projections often ignore interactive effects between systems (for example, lighting retrofits reduce cooling loads, so HVAC savings from automation are lower than if calculated independently). Third, vendor demonstrations often occur in ideal conditions with clean data and attentive operations staff, conditions that do not persist in typical commercial operations.

How important is commissioning to smart building success? Commissioning is the single strongest predictor of smart building performance. PNNL's 2024 study found that buildings with enhanced commissioning achieved 60 percent higher savings than those without formal commissioning. Commissioning identifies sensor errors, control logic conflicts, and equipment issues that would otherwise cause the smart system to optimize against incorrect data. Ongoing monitoring-based commissioning is equally critical: without it, the LBNL study found that 30 percent of first-year savings erode within three years due to setpoint drift, overrides, and equipment degradation.

What is the typical payback period for smart building technology? Payback periods vary widely depending on building size, energy costs, and deployment scope. For FDD software deployed on existing BMS infrastructure, payback is typically 6 to 18 months due to low capital cost and immediate fault-correction savings. Full BACS upgrades from EN 15232 Class C to Class A typically have payback periods of 3 to 7 years. AI-driven HVAC optimization platforms (SaaS model) often achieve payback within 12 to 24 months due to subscription-based pricing and rapid deployment. The critical variable is whether savings are sustained: projects that lose 30 percent of savings by year three have effective payback periods nearly double the initial projection.

How do smart building savings compare across building types? Commercial offices and retail spaces show the most consistent savings from smart building technologies because they have predictable occupancy patterns, significant HVAC loads, and established controls infrastructure. Hospitals and laboratories show high absolute savings potential but greater implementation complexity due to strict environmental requirements. Industrial facilities benefit most from process-level optimization rather than traditional BACS. Residential smart home technologies show lower savings (typically 8 to 15 percent) because individual dwellings have simpler systems and lower baseline energy waste compared to large commercial buildings (DOE, 2025).

Sources

• Lawrence Berkeley National Laboratory. (2025). Meta-Analysis of Verified Energy Savings from Building Automation and Control Systems: 312 Commercial Buildings. LBNL.
• International Energy Agency. (2025). Buildings Sector Energy and Emissions Overview 2025. IEA.
• Pacific Northwest National Laboratory. (2024). Fault Detection and Diagnostics in Commercial Buildings: Energy Savings, Fault Persistence, and Commissioning Effects. PNNL.
• Rocky Mountain Institute. (2024). Bridging the Performance Gap: Vendor Claims vs. Verified Savings in Smart Building Retrofits. RMI.
• DeepMind. (2025). AI-Driven Building Energy Optimization: Results from Commercial Office and Data Center Deployments. Google DeepMind.
• U.S. Department of Energy, Building Technologies Office. (2025). Commercial Building Energy Savings from Advanced Lighting and HVAC Controls. DOE.
• Verdantix. (2025). Global Survey: Smart Building Technology Adoption Barriers and Success Factors. Verdantix.
• CIBSE. (2024). Building Performance Evaluation: Closing the Gap Between Design and Operation in Non-Domestic Buildings. Chartered Institution of Building Services Engineers.
• CEN. (2024). EN 15232-1: Energy Performance of Buildings - Impact of Building Automation, Controls, and Building Management. European Committee for Standardization.
• BCA Singapore. (2025). Green Government Buildings Programme: Performance Report 2023-2025. Building and Construction Authority.