Introduction
Global data-center growth is accelerating at an unprecedented pace, largely driven by the surge in AI workloads. According to the International Energy Agency (IEA), data centers consumed about 415 TWh of electricity in 2024 (≈1.5% of global demand), and this figure is projected to more than double to 945 TWh by 2030, roughly equivalent to Japan’s current annual electricity consumption.
This rapid expansion is creating severe pressure on power grids. In many countries, grid connection backlogs for large loads can stretch from 3 to 7 years, particularly in major data-center hubs in the US and parts of Europe. Utilities face unprecedented interconnection queues as transmission and generation capacity struggle to keep pace. Furthermore, beyond infrastructure challenges, there is growing concern that this surge in demand will drive up electricity prices for consumers.
Under these conditions, securing reliable grid access and power supply has become a critical success factor for data-center projects. This is where Behind-the-Meter (BTM) configurations offer a compelling solution.
BTM setups involve on-site or near-site generation and storage assets directly connected to the data-center. This setup alleviates the load on the grid as part of the datacenter power demand is produced locally. It also helps mitigate the impact on consumer power price where large data-centers are concentrated.
The Key Advantages of BTM for data-centers
Faster deployment
By reducing the stress on the grid, it can avoid multi-year grid interconnection delays.
Lower CAPEX on grid infrastructure
Shared grid connection between data-center and generating assets, reducing the need for costly transmission upgrades and electrical infrastructure.
Reduced grid dependency
On-site generation significantly decreases reliance on external grid.
Enhanced resilience
On-site generation and BESS provide backup and supports N-1 contingency scenarios.
Efficiency gains
Many electricity markets incorporate transmission losses into pricing, often through locational marginal pricing or similar mechanisms. The principle being that losses increase with distance and congestion. Incorporating it into pricing encourages generation closer to demand and reduces system costs. Behind-the-Meter setup avoids locational marginal pricing.
Lower exposure to grid tariffs
and potential to monetize surplus energy generated on-site.
Hedging against high electricity prices
with BESS displacing excess renewable power to high priced hours.
Permitting & Community Acceptance
Smaller grid impact and higher sustainability profile improve local support, especially when renewables are integrated.
Job creation
Renewable generation assets support long-term local job creation and stimulate broader regional economic growth.
Increased renewable penetration
in the data-center power supply.
Many of the world’s largest data-center operators claim their facilities are fully or nearly powered by renewable energy. However, these claims often rely on energy credits or offsets rather than the direct use of renewable electricity.
Looking ahead to 2030 and beyond, several hyper-scalers have set ambitious goals to match their data-center power consumption with renewable generation on an hourly basis, moving beyond traditional annual matching. To achieve this, mechanisms like 24/7 clean power purchase agreements are being developed. Yet, with surging data-center demand and increasingly challenging grid connection bottlenecks, there is a growing incentive to invest in behind-the-meter renewable generation. This approach not only supports hourly matching for part of the facility power consumption but also eases pressure on the grid, helping accelerate the connection process.
Wind, Solar and BESS as building blocks
Wind and solar generation are the primary renewable assets to supply green power to data-centers. Both technologies are mature, have progressed significantly down their cost curves, and continue to improve. Other renewable sources, such as hydro or geothermal, can be valuable where locally available, but wind and solar remain the most scalable and widely deployable options globally.
The cost and performance of BESS technologies have improved significantly in recent years, driven by advancements in lithium-ion chemistry, manufacturing scale, and supply chain optimization. As a result, BESS has become increasingly competitive for short-duration storage applications. However, current battery technologies remain economically limited in their ability to store large amounts of energy over extended periods (multi-day to seasonal storage).
Solar Assets
- Low-cost energy
- Relatively predictable resource,
- Strong diurnal and seasonal patterns.
- Inherent day-night intermittency
- Maximum achievable capacity factors 30–35% in the best regions
Wind Assets
- Generally higher levelized cost than solar
- Deliver more consistent generation profile throughout the day and year
- Modern onshore wind farms can reach capacity factors of 40–60%
- Inherently more variable and less predictable than solar, with year-to-year fluctuations driven by broader climate patterns.
The addition of Battery Energy Storage Systems (BESS) enables
- Load shifting by storing excess renewable generation and discharging it during periods of lower production and high price.
- More stable and reliable power supply to data-centers.
- Energy arbitrage and participation in ancillary service markets, when grid-connected, creating additional revenue streams
Location matters: Combining attractive wind and solar resources
In some regions, wind and solar can provide very complementary generation profiles. With solar peaking during daylight hours and wind often performing better at night or during different seasons, making them a strong foundation for a diversified renewable supply strategy for data-centers.
Combining, on a single map, regions with strong solar and wind resources, while excluding protected areas, high-slope terrain, and other unsuitable land, provides a solid foundation for identifying potential sites that can maximize renewable power penetration in data-center operations. This approach helps pinpoint locations where high-quality renewable resources can meet a large share of the data center’s electricity demand without requiring disproportionate oversizing of generation assets relative to the data-center capacity.
However, achieving full grid (or fossil fuel) independence can become very costly, even in a location with strong combined resources and requires a high willingness to pay for BTM green power generation. Nevertheless, data-centers can still benefit from having some BTM renewable generation capacity, even without achieving full grid independence, as it can lower the net cost of power supply and provide a hedge against electricity price volatility.
Figure 1: Renewable resource attractiveness based on a score of 1-5 with 1 representing the most attractive locations. The map is based on a weighted overlay of the wind/solar resource, terrain constructability and exclusion of build-up and environmentally protected areas.
The cost of being free from grid: Understanding the “Hockey Stick” supply cost curve
Ensuring a constant and reliable power supply from intermittent generation sources is challenging. It often requires significant overbuilt of generation capacity and the integration of storage systems to shift surplus production, typically occurring during the day, to periods of low generation, when the sun doesn’t shine or the wind doesn’t blow.
Figure 2 illustrates the power supply cost to a data center as a function of the share of BTM renewable energy in its consumption, i.e. the proportion directly supplied by wind, solar, and BESS, excluding any renewable contribution from the grid.
Up to a 70% BTM renewable energy (RE) share for the data center, grid power is displaced by low-cost renewable generation, while overall curtailment remains below 15%. No BESS capacity is required, and the renewable mix is predominantly wind-based. As a result, the total power supply cost continues to decline, reaching its minimum at approximately 70% BTM RE penetration, well below the average cost of sourcing 100% of power from the grid, on a spot basis.
Figure 2: Total power supply cost to data-center and optimal capacity mix to achieve various RE% levels. Considering a data center with a constant power demand of 300 MW operating 24/7. In this example, the data center is connected to the grid and covers any generation shortfall from renewable assets with grid power. No power export is allowed.
Beyond that point, significantly more solar capacity is added to the mix to reach higher RE penetrations, increasing excess solar capacity that is captured and redistributed by added BESS capacity. The total power supply cost starts increasing until reaching, in this example, grid parity around 95% RE share. Note that the grid parity level depends on the structure of grid tariffs and the shape of the spot price curve, making it inherently location specific.
The final tranche of renewable energy share is the most challenging to achieve with renewable assets. It requires significant oversizing of both generation and storage capacities. This drives a sharp increase.
In total supply cost and curtailment level, giving the cost curve its characteristic “hockey stick” shape. This is where alternative storage technologies could play a role, offering a more cost-effective solution than BESS for achieving the last increments of renewable penetration.
The shape of the supply cost curve is highly project and location specific. It is strongly influenced by the availability and complementarity of natural resources, as well as the CAPEX levels and development complexity of the different assets.
Figure 3: Avg. daily power supply stack. Resource profiles require massive overbuild and storage to reach high BTM renewable penetration
To summarize, high BTM renewable penetration (>80%) can be achieved at a competitive cost, though reaching near-100% remains costly. Ultimately, the target renewable share and total power supply cost must align with the data center’s ESG commitments, its willingness to pay for green energy and its power sourcing alternatives.
That said, not all data centers can be co-located with large behind-the-meter renewable capacities. Hyperscale data-centers, supporting large-scale cloud services, big data analytics or AI training, have moderate latency requirements and therefore can be located in remote areas with abundant land and strong natural resources.
On the other-hand, data-centers with stringent latency requirements, such as edge facilities handling real-time data or AI inference, must be located near urban centers, where combining land availability with good renewable resources is highly challenging.
For these facilities, achieving near 100% BTM renewable penetration with conventional technologies is unlikely. Still, even a smaller BTM renewable capacity can deliver competitive advantages in terms of cost and risk mitigation to a data-center.
Cut costs, Hedge risks with BTM solar + BESS
In this section, we focus on a data center co-located only with solar and BESS behind the meter (BTM) to demonstrate the economic benefit on the net power supply cost as well as on the market price exposure. Note that while some level of wind generation through front-of-the-meter (FTM) contracts is possible, it would not reduce grid dependence nor alleviate grid load where the data-center is connected. This is why no wind capacity is considered in this example.
Solar production is prioritized to the Data-Center. The excess production is first sent to the battery, then sold to the grid. In this setup, approx. 20% of solar production in sold on merchant market. BESS charges from solar power and from the grid during low priced hours. This project configuration enables 44% BTM renewable energy penetration at the data-center, with approximately 15% of its energy demand covered by battery storage, hedging high priced hours.
Figure 4: Power distribution and power flow in a Solar + BESS behind the meter configuration. Solar capacity 2.5x the data-center capacity and a BESS capable of sustaining five hours of data center operation. BESS grid arbitrage is enabled.
Figure 5: Illustrative daily operation power-supply from BESS and Solar.
When accounting for revenues from surplus power sold to the grid and from BESS-enabled arbitrage, the net cost of power consumed by the data-center is reduced by roughly 15% compared with a standalone setup, under Central price projection scenario (Figure 6).
There is, of course, a risk that merchant revenues do not materialize and that future market dynamics lead to a very different price environment. This is why it is essential to test the robustness of the case across multiple price scenarios. Figure 7 compares both setups (standalone and BTM solar plus BESS) under High, Central, and Low price scenarios.
Figure 6: Breakdown of data-center power supply cost per annum and comparison of net cost per MWh consumed in data-center between a stand alone set up (grid connected data-center 100% powered from the grid) and a setup with solar and BESS BTM. Commercial structure based on a 15-year PPA with full pass-through of revenue share/firming cost where the solar power sent either to the data-center or the BESS is sold at a fixed price to the data-center for 15 years tenure. Grid charges are covered in full by the data-center during the PPA tenure.
Figure 7: Comparison of power supply cost per MWh under three price scenarios.
As shown in Figure 7, the net power cost is both stabilized and reduced relative to the Central and High market forecasts. Under the Low-price forecast, the net power cost is similar across both configurations. Overall, the analysis indicates that a BTM solar and BESS setup significantly reduces market exposure and mitigates the risk associated with high-price scenarios
Takeaways
1
The global data-center growth is creating severe pressure on power grids, leading the unprecedented interconnection queue as transmission and generation capacity struggles to keep pace
2
Behind-the-meter (BTM) renewable assets offer a strategic advantage, easing grid stress and avoiding multi-year interconnection delays, incentivizing data centers to co-locate with BTM renewables.
3
Sites with strong natural resources can achieve >80% BTM renewable penetration at competitive cost. However, reaching near-100% renewable penetration remains costly with conventional technologies, as it dramatically increases the need for over-sizing and storing power
4
Not all data-centers can be co-located with large BTM renewable assets, particularly those with strict latency requirements near urban areas. Yet, even a modest BTM renewable capacity can remain cost competitive against wholesale price forecasts, while reducing market exposure and mitigating risks from high-price scenarios.
5
Combining data centers with BTM renewable assets creates opportunities to structure strategic framework between renewable project developers, investors, and data center operators, aligning energy supply with demand and investment objectives.
Intrested in learning more? We’re one mail away.
Xavier Mauclere
Senior Consultant – Business Case & Strategy
xma@bluepp.dk
