The economics of renewable energy are evolving. Historically, developers’ main challenge was building affordable wind and solar capacity. Now, the focus is shifting to how much revenue each megawatt can generate after connection.

GlobalData expects global battery energy storage system capacity to increase sixfold between 2025 and 2030, at a compound annual growth rate of 42%, partly as operators respond to the variability, imbalances and curtailment that come with more wind and solar.[i]  The International Energy Agency (IEA) also expects renewables to overtake coal in global electricity generation in 2026.[ii]

How AI is turning renewable flexibility into asset value

A solar plant that produces heavily in a low-price midday market may have excellent resource quality but still have weaker realised revenues. Storage can shift some output to higher-value periods, but the commercial outcome depends on continuous decisions: when to charge, when to discharge, how much capacity to reserve for grid services, and how aggressively to cycle the battery.

This is where artificial intelligence (AI) starts to play a role in asset economics. At the 2026 Global Low-Carbon Industry Forum, co-hosted by the Global Solar Council, China Energy Research Society, World Alliance for Low Carbon Cities, Tsinghua Shenzhen International Graduate School and Huawei Digital Power, industry leaders discussed how AI, energy management and grid-forming technologies can enable more stable and dispatchable renewable energy systems.

Huawei outlined an approach that combines AI-based forecasting, energy management and grid-forming control across solar, wind and storage. Its energy management system uses weather forecasts and machine-learning models to anticipate generation and load, then co-ordinates storage around scheduled output, peak shaving and energy shifting.

According to Huawei, its AI foundation model can co-ordinate solar, wind and storage as a single resource, making output more controllable and dispatchable, rather than optimising each asset in isolation. In this model, AI helps determine how the assets should be operated, while grid-forming control helps the combined renewable and storage system deliver more stable, dispatchable output. The company also envisages AI agents supporting energy management, trading assistance and operations and maintenance (O&M), potentially enabling storage to serve multiple commercial purposes rather than functioning primarily as a grid-support asset.

Why battery intelligence is becoming a financial metric

For storage owners, state-of-charge (SOC) accuracy is one of the less visible factors affecting returns. An operator that cannot confidently determine how much usable energy remains may need to maintain a wider operating margin, reducing the capacity available for dispatch or trading.

Huawei says its combination of high-precision SOC management, round-trip efficiency and high availability can improve lifecycle returns by more than 10%, and its end-to-end AI agents can reduce O&M costs while supporting availability above 99.9%. Its Smart String Grid-Forming energy storage platform has also received TÜV Rheinland Level 3 SOC characteristics certification, covering areas including SOC accuracy, calibration and balancing. The company also described FusionSolar Agent as a four-layer AI architecture spanning computing infrastructure, power models, intelligent equipment and software agents. The broader significance is that forecasting, battery management, maintenance, dispatch and eventually trading decisions can increasingly be co-ordinated within the same operational system.

MTerra Solar shows dispatchable renewables at utility scale

The Philippines offers a useful test case. GlobalData forecasts that the country’s renewable capacity will rise from about 7.1GW in 2025 to around 30GW by 2035, with solar photovoltaic capacity increasing from 3.6GW to 18.7GW.[iii]  MTerra Solar provides one answer at an unusually large scale.

The project is planned to combine 3.5GWp of solar with 4.5GWh of battery storage. Phase 1 began commercial operations on 26 August 2026, initially supplying 600MWac of mid-merit capacity under a power supply agreement with Meralco. By the time of its July inauguration, 1.373GWac of solar and 3.3GWh of storage had been energised. Huawei is supplying inverters and smart string grid-forming ESS.

The completed project is backed by a 20-year power supply agreement to provide 850MW for an average of 13 hours a day. Huawei says its Smart String Grid-Forming ESS controls are designed to help the plant respond to grid dispatch instructions and smooth fluctuations in solar generation. Once completed, the project is also expected to supply clean electricity to around 2.4 million households and avoid approximately 4.3 million tonnes of carbon emissions each year. The project was frequently highlighted during forum discussions as an example of how large-scale solar, storage and grid-forming capabilities can support the evolution of renewables from supplemental resources into dispatchable power assets.

As renewable portfolios expand, returns will rely less on installed capacity and more on how effectively that capacity is managed. While AI cannot eliminate risks related to market design, grid limitations or weather conditions, the shift from forecasting to optimisation, dispatch and trading support indicates that software and control systems are playing an increasingly vital role in renewable investments.

To find out more, visit www.digitalpower.huawei.com/en/news/activity/low-carbon-industry-forum

[i] https://www.globaldata.com/media/power/global-battery-energy-storage-installed-capacity-to-rise-sixfold-by-2030-forecasts-globaldata/

[ii] https://www.iea.org/reports/electricity-mid-year-update-2026/executive-summary

[iii] https://www.globaldata.com/media/power/philippines-renewable-power-capacity-to-reach-30gw-by-2035-forecasts-globaldata