Energy affordability is a priority on political agendas in many markets as energy bills have risen. The typical European household spends roughly €3,500 to €5,000 a year on energy. Since mid-2021, 17 EU member states have introduced temporary interventions to limit household electricity and gas prices.
Policy discussions often focus on the cost of electricity per kilowatt hour, while the total energy bill that households pay for power, heating and cooling, and electrified transport receives less attention. Higher bills for those things are not due exclusively to rising energy costs. Moreover, consumers must navigate the energy market alone, managing their needs across four to six providers, including new market entrants that are upending traditional models.
The retail energy industry is likely facing significant structural change. AI is compressing the cost to serve, and the duration of the customer relationship is shrinking, which could limit providers’ ability to recoup acquisition costs.
The forces and players that are redefining energy retail promise to dramatically lower consumers’ energy bills, while also achieving what has seemed impossible—simultaneously advancing the decarbonization agenda. The race has begun, and providers, especially traditional ones, need to act promptly to preserve their market base and remain competitive.
Traditional Retail Is Under Pressure
The traditional retail energy model rests on three assumptions, all of which are rapidly becoming inoperative:
- Acquisition costs are recoverable through multiyear contract margins. Recovery depends on how long a customer stays. AI-native comparison platforms and the increasingly common phenomenon of personal switching agents acting for the customer are shortening that tenure, reducing the friction associated with switching.
- Cost to serve is a sustainable differentiator. Market behemoths once benefited from economies of scale. Today, customer service based on large language models (LLMs), already deployed at scale, could cut the cost to serve in contact centers, collections, billing queries, and back-office processing by 40% to 60% over the next three to five years.
- Customer inertia is a reliable retention mechanism. This is the most important but least discussed assumption. Switching rates are not yet rising: household switching across the EU averaged just 9% in 2024, and Australian rates have held steady for years. However, those numbers reflect friction, not loyalty—and the friction is disappearing from both ends of the supplier-customer relationship. By year-end 2026, new EU rules will take effect requiring suppliers to make any requested switch within 24 hours. Automation is also at play: an AI agent that performs a price comparison reads price and little else, while customers largely remain disengaged. As a result, their inertia will not hold when switching costs nothing and when marketing does little to disincentivize their exit.
A pure-play energy retailer has little left that is structurally defensible, except for one key capability that commoditization does not erode: buying energy skillfully and carrying the price risk for the customer. Volatile prices, flexible demand, and assets that shift load throughout the day make the purchasing book more complex, not less.
Nevertheless those same forces carry tangible benefits for customers. The AI that weakens traditional retailers’ economics is creating the largest potential household savings in two decades. The falling cost to serve, the elimination of acquisition costs, and the optimization gains that arise from managing the three dimensions together represent surplus value moving from the retailer to the household. Volatility makes the last of these possible, because households can take advantage of the spread between hourly prices rather than being locked into the average price. The question is how much reaches the household, and how quickly?
The New Downstream Model
The new downstream, as we call it, involves integrated management of all three dimensions of the household energy relationship:
- Power. The kilowatt-hour does not become obsolete. What changes is where the margin sits. Instead of earning on the spread between purchase and sale price, the retailer earns on when the kilowatt-hour is consumed—managing consumption against hourly prices, whether the customer sees them or not.
- Heating and Cooling (HVAC). Rather than offer an unmanaged thermal load—heating in colder markets, air conditioning in warmer ones—the retailer offers service scheduled against price, with the heat pump or air conditioning unit as the controllable asset.
- Electrified Transport. The retailer turns a separate mobility cost into an integrated energy-and-mobility offer, with the vehicle battery acting as a grid asset.
The €3,500 to €5,000 that an average European household pays for these three dimensions each year is the cost before the household takes any steps toward optimization. The corresponding assets behave differently:
- Solar is the outlier. It reduces the bill simply by existing, since every self-generated kilowatt-hour is a kilowatt-hour not bought. For an appropriate-size system with storage, that amounts to several times the savings that flexibility can deliver.
- The battery consumes nothing on net; it simply moves energy in time.
- A heat pump and an EV primarily substitute for natural gas and petrol.
- Air conditioning in markets that have not used it in the past is the exception, adding energy consumption rather than replacing it.
Where a household falls within the annual cost range depends partly on how electrified it already is. A relatively electrified household will be at the lower end of the range. (Our recent UK article makes that case in detail.) Integrated management works on each dimension—the battery on the power bill, the heat pump on heating and cooling, and the EV on transport—while solar cuts across all three.
Our figures exclude the annual cost of the assets to the customer. A heat pump and an EV carry a premium over the boiler and the internal combustion engine car that they replace, but the household would buy one or the other in any case. Solar and a battery are additional investments that displace nothing (and cost €25,000 or more for the pair).
Moreover, the figures apply to a single-family home, where the case is clearest. Apartments are not excluded, but they are somewhat different: air conditioning, balcony solar, and shared charging are controllable loads, and the economics improve when tenant-supply models or energy communities aggregate what individual housing units cannot. The mechanism is the same; the difference lies in who owns the asset and who captures the savings.
Doing the Math
What would a fully integrated AI-first approach look like? A typical European household with the relevant assets can lower its total energy bill by roughly 60% today, and by up to 80% once it can trade those assets in intraday and ancillary-services markets rather than only against day-ahead prices (“full market access”). However, the upper end of projected savings assumes bidirectional charging, which is not yet commercially available at scale. A €3,500 annual bill would drop to around €1,400, or under optimum conditions to €700. Most of the savings would come from self-generation: solar and a battery account for around half the bill alone. Flexibility would contribute an additional 9% today and a further 20% in the future. (See Exhibit 1.)
How the reduction splits between self-generation and flexibility matters more for the retailer than for the household. Self-generation provides the greater savings, but it accrues entirely to the customer. Flexibility provides less and is the only part that a provider can price. A 20% to 30% fee on flexibility is worth €55 to €80 per household per year today—less than the €80 to €150 that it costs a retailer to acquire that household. The recurring business resides in financing, installing, and operating the assets. Flexibility preserves the relationship, but the assets pay for it.
The new model requires someone to be exposed to hourly prices: the customer, through a dynamic tariff; or the retailer, carrying the price risk and passing back a share of the gain as either a lower fixed price or a flat bonus. Determining who carries price risk is a commercial choice, not a technical one. For mass-market customers who do not want exposure to volatility, the retailer-carries-the-risk version is more sellable and harder to displace. A customer operating on a dynamic tariff can seek optimization elsewhere, whereas one who has handed over the price risk has ceded the relationship. Many households with higher energy literacy or with assets they already manage will prefer the former and keep the full upside.
Our estimates assume hourly settlement, which necessitates using a smart meter, but they do not assume that the household is on a dynamic tariff.
Benefits to the Energy System
Household flexibility serves the greater good. A battery that charges when prices are low absorbs renewable output that might otherwise be curtailed. A heat pump that avoids evening peak usage defers network reinforcement. An EV that charges during off-peak hours reduces balancing costs.
In Belgium, a shift from uncoordinated EV charging to coordinated vehicle-to-grid would cut annual power system operating costs by anywhere from €470 million to €520 million, according to the Elia Group. The household savings and system savings stem from the same behavior: what a household does to lower its own bill is what the system needs it to do. That alignment is unusual, and it’s the main reason why we expect flexibility markets to open further rather than tighten.
Retailers can also trade flexibility in day-ahead and intraday wholesale markets. Where regulation permits, they can offer it in ancillary services, which is what makes a dynamic end-customer tariff optional.
The Challengers
Five types of players are extending into retail energy, challenging utilities for the household energy relationship (see Exhibit 2):
- The Oil Majors. Examples include Repsol and TotalEnergies in Europe, and ENEOS in Japan. Their position is strong, thanks to customer bases in the tens of millions, fuel station networks repurposed for EV charging, and consumer brand equity.
- Automotive OEMs. Examples include VW, Tesla, Stellantis, General Motors, Nissan, BYD, and Hyundai. The EV is the entry point into home energy management, since the manufacturer controlling the vehicle battery also controls the vehicle-to-grid transfer and, ultimately, the household flexibility asset.
- Tech Players and Neo-Retailers. Examples include Octopus/Kraken, Tibber, Australia's Amber Electric, Japan's Mitsubishi Motors/MC Retail Energy/Kaluza consortium, and Enpal. AI-first from day one and unencumbered by legacy systems, they are converging on the same position from opposite ends: some started with the tariff and the software and moved into solar, heat pumps, and batteries; others started by financing and installing hardware and added supply afterward. What they share is strong tech capability at the core.
- Telcos and Multiutility Bundlers. Examples include Telstra in Australia and TIM in Italy. These companies already bill tens of millions of households monthly, and they own the billing and service infrastructure that energy retail requires.
- Real Estate Companies. Examples include Vonovia in Europe, Greystar in the US, and Mitsui Fudosan in Japan. Large housing companies are penetrating the rental market, where utilities currently have almost no leverage. Private-equity portfolios often control the meter, too, but holding periods and the absence of an operating platform make them more likely to be partners than challengers.
Why Total Bill Reduction Is the Smarter Business Model
The new downstream is still in its early stages, so a body of proven outcomes does not yet exist. But our observations and analysis point to five forces that make focusing on total energy bill reduction not just a compelling strategy but also an imperative for energy retailers:
- The Size of the Asset Pie. Self-generation creates the single biggest reduction in a household’s bill, and a retailer captures none of it. The customer buys solar panels from an installer and consumes the panels’ output; the retailer supplies the residual energy that the customer requires. That residual will shrink every year, independent of AI’s effects. By financing, installing, leasing, or operating assets, a retailer can participate in the larger half of the total energy bill.
- Volume over Margin. As adjacent pools—flexibility, heating and cooling services, EV charging, and data-enabled services—expand, per-kWh margin will shrink along with the overall bill. Total revenue per customer, however, can grow.
- The Rise of AI Natives. If an incumbent does nothing to reduce a household’s total energy bill, an AI-native competitor will.
- Retention over Churn. A commodity contract is easy for an agent to leave; the product is identical everywhere, and only the price differs. In deregulated markets, churn is expensive. In Europe, the cost of reacquiring a commodity customer runs between €80 and €150. An integrated relationship is harder to leave because there is no comparable price, installed assets do not transfer, and optimization resets on switching. The retailer can gain lifetime value at a multiple of the commodity customer’s.
- A New Value Pool. In deregulated markets, the flexibility layer represents recurring, unregulated revenue. The household keeps 70% to 80% of the gain; the provider’s 20% to 30% service fee could double once intraday and ancillary-services markets open, but take rates will compress as the market matures. This is new—not redistributed—value. In price-capped markets (most of the US and parts of Australia and Asia-Pacific), however, the mechanism may not be available.
Who Gets to the Household Most Cheaply
When acquiring a customer no longer pays for itself, the question becomes, how can you reach the household without paying for it? Three routes are emerging:
- Contract-Base Led. Utilities, telcos, landlords, and others with a recurring billing relationship can cross-sell from an existing customer, materially lowering the marginal acquisition cost.
- Installation Led. The neo-retailers enter by financing and installing home-energy hardware. Then they extend that asset relationship into tariffs, optimization, and services.
- Anchor-Asset Led. EV and heat-pump OEMs start with the household’s largest flexible loads and use control of the asset and its data to expand into whole-home energy management.
Platform providers sit outside this map. Rather than competing for the household’s business, they are selling the capability to those that are competing. These routes can be combined; and whether the technology is built, licensed, or assembled through partners is a separate operating-model choice. Technology ownership alone does not create customer access.
No route can work without AI at the core. The critical asset is household-level energy intelligence: knowing a household’s energy needs, flexibility potential, and grid-service opportunities across all three dimensions at once. But intelligence alone is not enough. The harder capability is operational: delivering across installed assets, metering, settlement, billing, and VPP dispatch, reliably and at scale. Although many players can model the savings, few can run the delivery.
A third capability must precede both energy intelligence and operational capability: making an integrated proposition simple enough to buy. The provider is asking a household to finance hardware, accept an algorithm controlling it, and trust a single provider across three parts of its life. Explaining that in a way that people will say yes to is not trivial, and it is where most integrated offers have failed so far. A company that has all three capabilities will own the customer relationship of the next decade. Today almost no utility can claim all three, and within five years having them will define market leadership.
Several conditions must remain in place: data-access rules that permit third parties to reach the meter, the ability to run a services P&L with a different capital and margin profile, and flexibility markets that fairly remunerate aggregated household assets. The biggest unknown is whether customers will cede control of a battery, a heat pump, or their car to an algorithm.
Three Decisions That Retailers Need to Make
Utilities’ traditional retail model is unlikely to survive the next AI wave unchanged. Indeed, within three to five years, the window for establishing an integrated customer interface will largely have closed. So sitting behind a tech platform or OEM interface as a commodity infrastructure provider may no longer be enough. Taking three actions over the next twelve months is a good starting point:
- Model the total bill. Size it across your customer base to determine the share you could realistically influence.
- Decide whether to play. Owning the customer interface is not the only viable position. The alternative is to be the settlement and delivery partner behind the interface entity. Adopting that role means having a lower margin, but also lower risk—and it is a deliberate choice rather than a default. If you do play, your route is the contract base you already have.
- Secure the assets and partners. The installers, OEMs, and platform providers worth having are a finite pool, and the value proposition depends on them. Also, the asset pie is the larger half of the prize.
Optimizing for a lower customer bill is not a concession. In fact, it may well be the stronger strategy.
Where the new downstream has not yet taken off—as in much of North America, Latin America, and Asia-Pacific—a sound near-term move is to apply AI to the existing book. Compress acquisition costs and the cost to serve, and use agentic AI for switching and retention.
For oil majors, OEMs, telcos, real estate companies, and tech platforms, the prize will be large: in Europe alone, it amounts to some €3,500 to €5,000 a year across more than 200 million households—more than most adjacent service markets that they already compete in. Their path mirrors the one facing utilities: use the time remaining to build across more than one of the three dimensions. A one-dimensional offer, such as charging alone or heating alone, may be easier to launch, but it’s also easier to dislodge.
The role of energy retailer is being redefined. Its most likely successor is a household energy manager that optimizes the total bill across power, thermal control, and mobility, using household flexibility as the lever.
Retailers that get this balancing act right can measurably lower what households pay and can contribute to a socially sustainable energy system. For the first time in two decades, decarbonization and affordability are set to move in the same direction.
The question is not whether to act, but whether to lead—or surrender the customer to someone who does.
