Deep Retrofits: Fixing the Buildings You Already Own

Analysis · THE BUILT ENVIRONMENT'S CARBON PROBLEM · PART 5 OF 6

Most of the buildings that will exist in 2050 are already standing, and their emissions are still going up. The work now is fixing them, and there is a practical way to sequence it, pay for it, and time it to the deadlines that are already coming.

About 80% of the buildings that will exist in 2050 are already standing today, based on Cushman & Wakefield's analysis of roughly a million assets across 143 countries. That one number decides where the emissions and the risk actually sit. Most of the attention in construction goes to new buildings and cleaner materials, but the larger share of the problem is the stock that is already up and will still be in use decades from now. And the numbers on that stock are heading the wrong way. Operational emissions from buildings reached 9.9 gigatonnes of CO₂ in 2024, about a quarter of all energy-related CO₂, and they have climbed 6.5% since 2015. To stay on a net-zero path they should have fallen by close to a third over that same stretch. Closing the distance between where the existing stock is and where it needs to be is the work of the next decade, and it lands on owners, not just builders.

The work is happening far too slowly

The way you close that gap is renovation, and the rate of renovation is where the problem shows up plainly. Across the global stock, buildings are upgraded for energy performance at roughly 1% a year, and most of that work is shallow: a boiler swap, some new lighting, a bit of added insulation. Deep renovation, the kind that cuts a building's energy use by half or more, runs at about 0.2% a year in the European Union. At that pace, bringing the existing stock up to standard would take more than a century. The International Energy Agency's net-zero pathway calls for the renovation rate to at least double to around 2% a year by 2030, with about a fifth of existing buildings brought to a zero-carbon-ready level by then. The European Union's own target is a deep-renovation rate of 3% a year. Whether the benchmark is two percent or three, the work needs to happen two to three times faster than it does today.

Almost none of today's renovation is deep

Annual building renovation rate: today's rate split into deep and shallow work, vs. what net-zero pathways require

Deep renovation (cuts energy 50%+) Shallow renovation

Only about 0.2 points of today's ~1% renovation rate is deep. At that deep-renovation pace, covering the whole stock would take roughly five centuries; at the EU's 3% target, about 33 years. Sources: European Commission (A Renovation Wave for Europe); IEA (Net Zero pathway).

Why the fix stalls

The barrier is not whether deep retrofits work. A deep retrofit, one that reaches the building envelope, the heating system, and the controls together, reliably cuts a building's energy use by half or more, against the 15% to 25% a lighter upgrade delivers. The barrier is that a deep retrofit is expensive up front and, treated as a standalone project, competes for capital against uses with faster and more certain returns. This is the first-cost bias described in Part 3: the person weighing the spend sees a large cheque today against savings that arrive slowly over many years, and the money goes elsewhere. On a standalone basis those savings can take well past a decade to repay the work even with incentives, which keeps most owners in the shallow-renovation lane and the deep-renovation rate down near a fifth of a percent.

What actually works

The way through is to stop treating the retrofit as a standalone event and fold it into money that is already being spent. Every building runs through a cycle of capital events: the roof reaches end of life, the boiler fails, a major tenant turns over, the loan comes up for refinancing. Each of those is a moment when an owner is already writing a cheque and can choose the low-carbon option instead of a like-for-like replacement. RMI calls this a "zero-over-time" approach, and its analysis of Canada's federal building stock found that fully decarbonizing it added only about 4% to what maintaining that stock would have cost anyway over 2020 to 2050. The premium is small because you are not paying to rip out working equipment; you are paying the difference between the ordinary replacement and the better one, at the moment the spend was happening regardless.

That same logic can be industrialized. The Dutch Energiesprong model measures a building with laser scanners, manufactures insulated façade panels and a new roof off-site, and fits them to the existing structure with on-site work done in as little as ten days, cutting energy use by 70% to 80%. Around 5,000 homes have been retrofitted this way with roughly 100,000 more aggregated across Europe. Standardizing the product is how the rate moves from one percent toward three: when the work becomes a repeatable manufactured process rather than a bespoke construction project, the cost and the disruption both fall.

The upfront cost has answers that most owners never reach for. Commercial Property Assessed Clean Energy, or C-PACE, lends fixed-rate, non-recourse capital repaid through an assessment on the property tax bill over as long as 30 years, and the obligation transfers to the next owner on sale rather than being owed all at once. An energy savings performance contract has an energy service company install and maintain the upgrades and get paid out of the guaranteed savings, often with little or no money down. Efficiency-as-a-service puts the equipment on a third party's balance sheet rather than the owner's. In Canada, the Canada Infrastructure Bank's Building Retrofits Initiative finances the capital cost and is repaid from the energy savings, alongside the $200-million Deep Retrofit Accelerator under the Canada Green Buildings Strategy. Each of these changes the question from whether the owner can write the cheque to whether the saving covers the payment.

There is also a reason to fix rather than replace that sits underneath the energy math. Knocking a building down and rebuilding it discards the carbon already spent in its concrete and steel, the embodied carbon from Parts 1 and 2, and then spends a fresh budget of it on the new structure. The RICS Whole Life Carbon standard, in force since July 2024, now counts demolition emissions that earlier versions left out, which puts a measured penalty on knock-down-and-rebuild and pushes a retrofit-first test before demolition is considered. Some buildings are genuinely past viable reuse, but reusing a sound structure keeps the embodied carbon already paid for and starts the whole calculation from a lower base.

The clock is already running

The reason to sequence this now rather than someday is that the deadlines are arriving on their own schedule. Building performance standards set a ceiling on how much energy or carbon a building may use, and charge the owner for going over it. New York's Local Law 97 bills $268 for every tonne of CO₂ above a building's cap, and roughly 63% of covered buildings currently sit above their 2030 limits. Boston's BERDO began levying fines of $1,000 a day on non-compliant buildings in 2025, and more than 40 US jurisdictions will have standards in force by 2026. In Europe, the recast Energy Performance of Buildings Directive covered in Part 4 requires the worst-performing buildings to be renovated on a fixed timetable, and in England a commercial building already cannot be let if it falls below a minimum efficiency rating.

The pressure is real, but it is not uniform, and it is worth being straight about that. Colorado recently halved its penalties and turned its 2026 targets into non-binding goals while keeping the 2030 ones, and parts of the European and UK timelines have been softened rather than tightened. This is why the durable case for acting is the market one from Part 4, the effect of carbon performance on value, insurability, and financeability, with regulation as the accelerant rather than the whole reason. A compliance deadline can be moved by a government; the repricing of a building by lenders and insurers does not reverse.

Proof that it pencils

The clearest evidence that a deep retrofit can pay is the Empire State Building. Its upgrade was folded into a capital improvement program that was already planned, designed as a whole system rather than a list of parts, and it remanufactured the building's original windows on-site instead of buying new ones. The result was a guaranteed 38% cut in energy use, about $5.9 million a year in savings, and a payback of roughly three years on the efficiency measures, with total emissions down 54% since 2009. The building has since been certified LEED Platinum, recognition that followed the performance rather than driving it.

It worked for the same reasons the approach above works: the spend was sequenced into money that was moving anyway, and the building was treated as one connected system instead of a set of separate replacements.

A deep retrofit that paid for itself

Empire State Building: results vs. a typical building-energy retrofit

~$5.9M saved per year

~3 yrs payback on the efficiency package

−54% emissions since 2009

Energy-use reduction. The 38% cut was guaranteed and has been met or exceeded every year; a typical, largely shallow retrofit saves 15–25%. Sources: RMI (Empire State Building retrofit); Carbon Management (typical retrofit range).

How an owner actually decides

For an owner with one building or a portfolio, the answer is a sort, not a single move, because not every building should get a deep retrofit. The practical version has four lanes. A building that already sits above a performance-standard cap, or that has a lease rollover, a refinancing, or a major capital event coming, belongs in the deep-retrofit lane now, because the window to fold the work into other spending is open. A building with runway before any deadline takes the staged path: do the cheap, high-return measures today and plan the deep work for the next capital event. A building that is already compliant with low exposure gets monitored rather than touched. And a building whose retrofit cost would exceed the value it protects belongs in the exit-or-redevelop lane, sold or rebuilt with the whole-life carbon penalty counted honestly.

Every one of those lanes turns on a single figure: what it costs to do nothing. That is the value, income, insurability, and financing a building stands to lose by staying where it is against the deadlines and market pressures that actually apply to it. Once that number is on the page, the sequence mostly writes itself, because it shows which buildings are losing money by waiting and which have time. The work of a transition plan is putting that number in front of an owner and ordering the moves behind it.

The existing stock is where most of the emissions and most of the risk sit, and the tools to deal with it, sequencing into capital cycles, financing that spreads the cost, and a clear read of the deadlines, already exist. What is missing on most buildings is the plan that puts them in order. The final part of this series pulls the whole argument together: why the built environment stays underpriced, and what the people who own and finance it can do with that.

‍ ‍

Sources

Fixing the Buildings You Already Own