Deep dive: China’s “AI + Energy” plan

The energy sector has emerged as one of the first big focuses of China's high-level “AI+” policy drive – China’s most comprehensive blueprint yet for how it plans to develop and deploy AI domestically...and compete internationally.

On Monday, the macro planner (NDRC) and energy regulator (NEA) issued guidelines to advance the use of AI across China’s energy sector.

• The plan comes hot on the heels of a State Council blueprint for implementing the “AI+” initiative, issued late August.

Long story short, Beijing’s goal is to integrate AI throughout the power system.

• Targeted applications include everything from demand forecasting and power market management to renewables dispatching, predictive maintenance, smart grid development, and even fusion experiments.

But why is energy such a high priority within the “AI+” initiative?

China’s power system – the backbone of both the industrial and digital economies – is undergoing a sweeping transformation.

• And Beijing wants AI to help manage the growing challenges.

The core issue: Once dominated by a relatively small fleet of large, stable hydro and coal-fired plants, China’s power system is rapidly incorporating vast numbers of decentralized, intermittent renewable energy supplies.

• While cleaner and often cheaper, renewables and distributed energy are making the broader grid increasingly complex and difficult to manage.
• Meanwhile, it also needs to make enduring fossil assets more efficient, both to meet carbon targets and to shore up energy security.

Regulators now believe AI can help optimize and manage the rapidly evolving domestic energy system. To those ends, the plan calls for:

• By 2027, the development and deployment of five sector-specific large language models (LLMs) to support the grid, power generation, coal mining, and the oil and gas industry – along with 10+ “AI+energy” pilot projects that could be replicated at the national level.
• By 2030, the establishment of “world-leading energy-specific AI capabilities,” (to be defined), including breakthroughs in AI-enabled smart grid dispatch, energy resource exploration, and renewable generation forecasting – all primary focuses within the broader range of grid management challenges.

Naturally, we were keen to see what the plan augurs for climate.

• Here's what we've found.

Implicitly, the plan positions AI as an indispensable tool for climate mitigation through grid management optimization.

First: To increase total consumption of intermittent clean energy, regulators are integrating new flexible resources into the grid, including virtual power plants, energy storage projects, and vehicle-to-grid (V2G)-enabled electric vehicles (which effectively serve as batteries for the grid when plugged in).

• These projects will greatly increase the scale and frequency of bidirectional power flows and make grid dispatch more complex.

Second: To make the energy transition work, regulators are gradually developing a unified national power market that will allow electricity to move across vast distances – e.g., from the sun-rich desert to the power-hungry coastline – and be traded in increasingly complex ways.

• This will make system dispatch far more challenging for human operators, but a prime candidate for AI-supported management.

Third: Regulators are promoting renewables-powered zero-emissions industrial parks and self-sustaining renewables-centric microgrids to accelerate industrial decarbonization.

• AI-driven smart systems could play a critical role in tackling the immense challenge of coordinating renewables, storage, and flexible loads to maintain stable and reliable power supply.

Meanwhile, it also sets AI up to help the energy system adapt to the impacts of climate change.

First: As climate change impacts intensify, the (increasingly renewables-powered) power system will be increasingly exposed to erratic and extreme weather conditions.

• That creates a pressing need for far more advanced weather forecasting tools – an obvious opportunity for AI-supported R&D.

Second: China is building massive renewables bases and long-distance transmission lines in remote, harsh environments, e.g., in deserts and out at sea – which will be made harsher over time by climate-driven weather extremes.

• Automated maintenance and repair – ultimately including embodied AI, including in the form of drones and humanoid robots – will be key to keeping systems reliable and costs in check.

Of course, there's much more to explore here – not least what the plan suggests for the longevity of various fossil fuel assets.

• Plus plans for hydropower, hydrogen, nuclear fission...even fusion.
• If you're interested in those areas or other implications, email us!

But getting back to the bigger picture: All in all, Beijing’s “AI + Energy” vision is highly ambitious, promising to drive:

• Development of novel LLMs tailored to energy sector applications
• Breakthroughs in AI-enabled smart grid dispatch, energy resource exploration, and forecasting of renewable energy generation
• Rollouts of pilot projects to test and scale up promising applications

So what will this all mean in practice?

Among the (many) specific instances we'll be watching for are:

• Development of advanced AI-driven VPP management systems capable of coordinating dispatch across vast numbers of distributed demand and generation-side energy resources
• Integration of LLMs into renewable energy management, optimization, and site selection tools to optimize project site choice and automate power generation planning and operational strategies
• Development of smart microgrid management systems for industrial parks that can coordinate and adjust on-site power generation and demand, enabling the system to operate self-sufficiently
• Creation of next-generation LLM-enabled weather forecasting tools that offer significantly improved forecast accuracy and more precise projections of renewable energy output
• Buildout of AI-powered systems that leverage remote sensors, drones, and unmanned ships and vehicles to automate maintenance and repairs of projects in remote regions

But for now: We'll keep you updated as implementation efforts are designed...

• ...and as they're rolled out.

Get smart: The energy sector is a smart early focus for China's “AI+” policy drive.

• The vast scale of China’s power system and the massive volume of high-quality data it generates make it a major potential beneficiary of AI.
• In turn, it’s also a promising testing ground for AI deployment, with significant potential to help practical applications scale and mature.

Get smarter: But the plan is long on ambition and short on time.

• China’s ability to develop next-gen, AI-driven grid dispatch and load management systems could make or break its energy transition over the coming decade.
• Yet the broad scope of the plan’s ambitions risks excessively fragmenting R&D efforts – leaving many half-baked solutions in lieu of fewer actionable outcomes.
• An "everything above" approach may work for policy to incentivize commercial product development, where the market can pick winners.
• But it’s a risky proposition when it comes to R&D that'll reshape critical national infrastructure.