Introduction
Australia’s electricity market is changing quickly, driven by the rapid growth of distributed energy resources (DER). New technologies are turning rooftops into miniature power stations and allowing batteries to supply backup power that helps keep the grid stable, flexible, and resilient.
Virtual power plants (VPPs) play a key role in this transition by linking thousands of small energy devices so they can operate as a single, coordinated participant in the market. Over the past year, VPPs have become firmly established in Australia, supported by new funding, pilot projects, and regulatory initiatives across several states.[1]
Since our article on Virtual Power Plants: Five things you need to know in August 2024, the VPP market in Australia has shifted.[2] One of the most significant evolutions is the new dispatch mode that will enable VPPs to directly compete in the energy market.[3] The new dispatch mode will commence on 23 May 2027,[4] meaning market participants have approximately 18 months to prepare for the upcoming change.
This article delves beyond the current VPP market to explore the future of VPPs, specifically how emerging technologies such as AI will pair with VPPs to provide optimised efficiency, reliability and market participation in the National Electricity Market (NEM).
The current VPP challenge
A VPP is a scheduling system that brings together a variety of energy resources,[5] often including DER such as rooftop solar panels and energy storage systems.[6] A VPP orchestrates these resources by determining how and when energy is used.[7] VPPs can also directly interact with the NEM by participating in market bidding and providing essential services, such as maintaining frequency stability through Frequency Control Ancillary Services (FCAS).[8]
Traditionally, VPPs rely on algorithms, predictive models and historical data but do not include real-time data due to limitations in processing speeds.[9] This poses the following challenges:
- Energy resource analysis: Perfecting real-time analysis and model fusion.
- Load and power output production prediction: Reacting to unexpected events and extreme weather conditions.
- Data security: Characterising and scaling to large-scale systems.
- Resource aggregation: Improving speed, robustness and system stability under extreme conditions.
- Scheduling: Adjusting to changing regulatory environment and data insufficiency.
- Market trading: Integrating algorithm with system and balancing efficiency.
The AI solution
AI broadly refers to machines designed to replicate human intelligence, including reasoning, calculating, and learning.[10] In application, AI can be used to analyse data, make decisions, manage large datasets and operate autonomously.[11] In fields that emphasise data analysis and predictive capabilities, AI is increasingly replacing traditional methods by leveraging its capacity to process complex datasets and deliver clear outputs quickly.[12] When applied to electricity grids, these capabilities have the potential to enhance efficiency, resilience and overall performance.[13]
For instance, VPPs could, with the help of AI, monitor environmental conditions such as weather (including sunlight), anticipate energy demand with greater precision, and autonomously participate and optimise performance in real-time electricity trading within the NEM. Already, AI is being applied to VPPs in the following ways:
- In the USA, grid intelligence company LineVision has combined power line sensors and machine learning to model transmission line capacity up to 10 days ahead,[14] cross-checking AI outputs with physics-based methods for reliability.[15] Also in the USA, software company Amperon has used AI-powered forecasting for supply, demand and price predictions relevant to VPP operations,[16] while the National Renewable Energy Laboratory has developed eGridGPT to assess grid conditions and recommend operational improvements.[17]
- In China, State Grid Xintong Company has applied the User-Side Load model to improve forecasting algorithm stability and precision.[18]
- Studies have also analysed how VPPs can apply machine learning to improve grid control during extreme events, enhancing power management efficiency and resilience.[19]
One specific application of AI in VPPs is Digital Twin Technology (DTT) – a virtual simulation that replicates physical systems in real-time using sensor data.[20] When applied to VPPs, DTT enables operators to analyse performance patterns, predict maintenance needs and test system changes in a virtual environment before implementation.[21]
DTT addresses several VPP challenges identified above, particularly improving real-time analysis and system integration.[22] Digital twins can reduce unexpected maintenance events, enhance grid stability planning, and allow operators to simulate future grid changes such as adding renewable sources or retiring aging plants.[23] Australian energy companies are already implementing these AI technologies, with Essential Energy deploying digital twins covering 95 per cent of NSW,[24] and Endeavour Energy pioneering an ‘engineering grade digital twin’.[25]
As VPPs prepare for the new ‘dispatch mode’, establishing robust data management frameworks has become essential for operational success. Firms developing VPP projects must implement unified data systems that provide real-time visibility across distributed energy resources, enabling accurate forecasting and optimal resource allocation. This project intelligence approach – integrating pipeline data, resource availability and financial metrics into a single source of truth – directly addresses VPP challenges, particularly improving real-time analysis capabilities and enhancing market participation accuracy.
Navigating challenges in the AI era of VPPs
Cyber security
While an AI-powered VPP future is promising, it also brings a new set of challenges. The integration of distributed energy resources and the use of machine learning focused on real-time optimisation exposes VPPs to various cyber threats, including data breaches.
Specific risks include:
- Vulnerability: The interconnected nature of VPPs makes them susceptible to cyberattacks. Unauthorised control over VPPs can lead to critical functions such as FCAS being interfered with.[26]
- Device Security: Devices can be targeted by cyberattacks that may impede the functionality and health of essential VPP elements, such as inverters, energy storage systems and smart meters.[27]
- Data Security: Data security is essential to protect against unauthorised access and consequent data breaches.[28] Tampered data directly impacts the reputation of VPP operators and the wellbeing of their customers, as well as the authenticity of the information.[29]
- Personal Applications: The use of mobile applications for remote monitoring and dispatch, as seen in South Australian VPP implementations,[30] introduces additional privacy considerations around user data collection, location tracking, and secure authentication protocols.
In light of the above emerging challenges, VPP operators should ensure clear contractual risk allocation between themselves and AI service providers.
Security of Critical Infrastructure Act 2018 (Cth)
The Security of Critical Infrastructure Act 2018 (Cth) (SOCI Act) is highly relevant to the operation of VPPs.[31] While VPPs are not explicitly mentioned as an asset class under the SOCI Act, they may fall under existing critical infrastructure categories depending on their specific characteristics and operations.[32] For example, if a VPP has a nameplate generation capacity greater than or equal to 30 megawatts and is connected to a wholesale electricity market it may be classified under the Act as a critical electricity asset.[33]
VPP operators must avoid compliance gaps and ensure robust security frameworks. It is critical for VPP operators and related businesses to build robust data security and reporting provisions into their contracts with subcontractors and service providers, ensuring clear allocation of responsibilities, obligations and risks to safeguard against any potential cyber incident or data breach.
Market conduct
As VPPs increasingly rely on AI for automated market participation, operators should also consider emerging regulatory compliance risks around market conduct. The recently released Staff Working Paper on AI and Energy (Staff Working Paper) by the Australian Energy Market Commission (AEMC) highlights concerns about algorithmic collusion, where AI systems may inadvertently learn to coordinate pricing strategies in ways that could be considered anti-competitive.[34] The Staff Working Paper highlighted that AI algorithms can systematically learn to collude and charge supracompetitive prices, potentially leading to higher prices for consumers.[35]
The NEM’s characteristics make it particularly susceptible to such algorithmic coordination.[36] VPP developers who will bid VPP resources into the market must maintain appropriate safeguards. These should include maintaining human oversight of algorithmic decisions, ensuring transparency in AI decision-making processes, and establishing clear protocols to prevent pricing strategies that could be interpreted as anti-competitive. The AEMC has recommended establishing a working group with the Australian Competition and Consumer Commission and the Australian Energy Regulator to develop appropriate regulatory responses, indicating this will be an area of ongoing regulatory focus as AI adoption in energy markets accelerates.[37]
Conclusion
The integration of AI into VPPs is a step towards a more responsive and efficient energy system. While many of these developments are nascent, early applications suggest that these technologies could help address some of the key challenges faced by VPPs.
The Australian regulatory framework is evolving to facilitate the future of a grid with DER. AI is well placed to address persistent VPP pain points – from forecasting to operation control and maintenance – within a maturing Australian regulatory framework.
However, a critical question emerges: is the current regulatory process adequate given the rapid pace of technological change occurring in the VPP sector? With AI capabilities advancing at an unprecedented rate, and the new dispatch mode not commencing until May 2027, there is a risk that regulatory frameworks may lag behind technological developments, potentially creating compliance gaps or missed opportunities for market optimisation. However, to date, VPP operators have demonstrated a willingness to embrace these challenges and opportunities.
The Hamilton Locke team advises across the energy project life cycle – from project development, grid connection, financing, and construction, including the buying and selling of development and operating projects. For more information, please contact Matt Baumgurtel.
[1] Australian Renewable Energy Agency, Project Jupiter (Web Page, 23 May 2025) <https://arenalla.gov.au/projects/project-jupiter/>; Government of South Australia (Energy and Mining), South Australia’s Virtual Power Plant (Web Page) <https://www.energymining.sa.gov.au/consumers/solar-and-batteries/south-australias-virtual-power-plant#about>; Rule Determination, National Electricity Amendment (Integrating price-responsive resources into the NEM) Rule 2024 (Cth) [1] (‘Rule Determination’).
[2] Australian Energy Market Operator, Integrating price responsive resources into the National Electricity Market: High level implementation assessment (Preliminary view, March 2025) 5.
[3] Australian Energy Market Commission, Energy market gets clear vision: Reform opens door for all to benefit from virtual power plants (Media release, 19 December 2024) <https://www.aemc.gov.au/news-centre/media-releases/energy-market-gets-clear-vision-reform-opens-door-all-benefit-virtual-power-plants>.
[4] Ibid.
[5] Xinxing Liu and Ciwei Gao, ‘Review and Prospects of Artificial Intelligence Technology in Virtual Power Plants’ (2025) 18(3) Energies 3325, 3327 (‘Prospects of AI in VPPs’).
[6] Ibid.
[7] Ibid.
[8] Australian Energy Market Operator, Virtual Power Plant (VPP) Demonstrations (Web Page) <https://aemo.com.au/initiatives/major-programs/nem-distributed-energy-resources-der-program/der-demonstrations/virtual-power-plant-vpp-demonstrations>.
[9] Prospects of AI in VPPs (no 5) 3327.
[10] Yongjun Xu et al, ‘Artificial intelligence: A powerful paradigm for scientific research’ (2021) 2 (4) Innovation (Camb) 1.
[11] Prospects of AI in VPPs (no 5) 3327.
[12] Ibid.
[13] Nuraini Diah Noviati, Sondang Deri Maulina and Sarah Smith, ‘Smart Grids: Integrating AI for Efficient Renewable Energy Utilisation’ (2024) 3(1) International Transactions on Artificial Intelligence 2.
[14] Maeve Allsup and Laura Weinstein, Seven ways utilities are exploring AI for the grid (Web Page, 16 October 2023) <https://www.latitudemedia.com/news/seven-ways-utilities-are-exploring-ai-for-the-grid/> (‘Seven ways utilities are exploring AI for the grid’).
[15] LineVision, How LineVision’s Non-Contact Sensors Power Dynamic Line Ratings (Web Page) <https://www.linevisioninc.com/methodology>.
[16] Ibid.
[17] Danhao Wang, Daogang Peng and Dongmei Huang, ‘Application and prospects of large AI models in virtual power plants’ (2025) 241 Electric Power Systems Research 111403, 111410; 11409 (‘Application and prospects of large AI models in VPP’).
[18] Ibid.
[19] Md Romyull Islam et al, ‘Building a Resilient and Sustainable Grid: A Study of Challenges and Opportunities in AI for Smart Virtual Power Plants’ (Conference Paper, ACM Southeast Conference, 27 April 2024) 101.
[20] Abdelali Abdessadak et al, ‘Digital twin technology and artificial intelligence in energy transition: A comprehensive systematic review of applications’ (2025) 13 Energy Reports 5196, 5200 (‘Digital twin technology and artificial intelligence in energy transition’).
[21] Orhan Korhan, Digital Twin Technology – Fundamentals and Applications (IntechOpen, 2023) Introductory Chapter.
[22] Ildar Idrisov et al, ‘Microgrid Digital Twin Application for Future Virtual Power Plants’ (Conference Paper, 49th Annual Conference of the IEEE Industrial Electronics Society, October 2023) 1.
[23] Digital twin technology and artificial intelligence in energy transition (no 20) 5201.
[24] Essential Energy, Digital Twin increases opportunities for renewable energy (Web Page, 22 May 2023) <https://www.essentialenergy.com.au/media-centre/media-release/news-7-digital-twin>.
[25] Endeavour Energy, Endeavour Energy pioneers Neara digital twin in transition to modern grid (Web Page, 13 December 2021) <https://www.endeavourenergy.com.au/news/media-releases/endeavour-energy-pioneers-neara-digital-twin-in-transition-to-modern-grid>.
[26] Sriram Prabhakara Rao, Ranganathan Prakash and Shree Ram Abayankar Balaji, ‘Virtual Power Plants Security Challenges, Solutions, and Emerging Trends: A Review’ (Conference Paper, Cyber Awareness and Research Symposium, October 2024) 3.
[27] Ibid 4.
[28] Ibid.
[29] Ibid.
[30] Tesla, Tesla App for Energy (Web Page) <https://www.tesla.com/en_au/support/energy/powerwall/mobile-app/tesla-app-for-energy>.
[31] See the types of critical infrastructure energy assets at: Critical Infrastructure Security Centre, Australian Government Department of Home Affairs, SOCI Act 2018 for energy (Web Page, 5 December 2023) <https://www.cisc.gov.au/information-for-your-industry/energy/legislation-regulation-and-compliance/soci-act-2018>.
[32] Ibid.
[33] Ibid.
[34] AEMC, ‘Addressing the risk of algorithmic collusion’ (Staff Working Paper, July 2024) <https://www.aemc.gov.au/sites/default/files/2025-09/AI%20and%20the%20future%20of%20energy%20regulation%20-%20Algorithmic%20Collusion.pdf> 13.
[35] Ibid.
[36] Ibid 14.
[37] Ibid 9.
