Summary of Key Points
With the dual impacts of market-oriented electricity reforms (new energy no longer enjoying guaranteed quantity and price, with market-based pricing becoming the norm) and low-carbon constraints (carbon emissions included in assessments, and carbon prices becoming tangible costs), the profit model for the new energy industry has shifted from easy profits to more challenging earnings. To cope with these pressures, companies are increasingly investing in AI technology to enhance their power generation forecasting, dispatching efficiency, and trading capabilities in order to gain market premiums. However, they also face practical challenges such as data barriers and lack of transparency in algorithms. The competitive landscape within the industry is shifting from a focus on installed capacity to a focus on the intelligence of these systems.
1. Changes in the Rules of Making Money: From Easy Profits to Relying on the Market
In the past, new energy companies had a much easier time—the state guaranteed that all the electricity they generated could be sold at a fixed, favorable price, essentially allowing them to make money effortlessly. But with the implementation of various policies by 2025, this model has come to an end:
- All new energy must be traded on the market: Wind and solar power cannot rely on state support; companies now have to sell their electricity in the electricity market, where prices fluctuate based on supply and demand.
- Full coverage of spot markets: Electricity prices change in real-time, similar to the price of fresh vegetables; companies must adapt to market conditions.
- Stringent carbon constraints: Carbon emissions are now part of party and government assessments, and carbon prices are no longer a vague “environmental cost” but a tangible financial factor.
As a result, company revenues have declined (the revenue of the five major state-owned power generation companies all decreased in 2025), and their debt ratios have soared (many exceeding the 70% threshold set by the State-owned Assets Supervision and Administration Commission), making it increasingly difficult to make profits.
2. Companies Investing in Intelligence: AI as a Lifeline for New Energy
To survive under the new rules, companies are equipping their energy systems with intelligent capabilities to boost profitability:
- Long-term vision: Using large-scale weather models and energy models to integrate solar and storage technologies for smart power generation and storage management.
- Tianhe Fujia: Developed an AI model for carbon management to help companies calculate and manage their carbon emissions.
- Luming Starlight: Creating an AI system for electricity trading, aiming to assist customers in trading 80 billion kWh of electricity by 2026—helping them negotiate prices more effectively and sell electricity at higher prices in the market.
The core goal of these AI tools is to transform the traditional fixed-income model (based on installed capacity) into a dynamic one that relies on intelligent operations to generate additional profits.
3. How AI Solves Three Major Challenges
What practical problems can AI solve for companies? Mainly three key issues:
1. Accurate power generation forecasting: Unstable wind and solar power generation (e.g., clouds covering solar panels, sudden gusts of wind) can be accurately predicted using weather data, preventing fines due to inaccurate forecasts.
2. More flexible dispatching: The electricity system includes various distributed resources such as charging stations, energy storage, and virtual power plants; AI can coordinate these resources in real-time, much more efficiently than manual processes.
3. More profitable trading: Selling electricity involves considering multiple types of transactions (spot, medium- to long-term, auxiliary services, etc.); AI can analyze vast amounts of data to help companies choose the best timing for bidding and maximize their profits.
4. Hindrances to the Integration of AI: Data Barriers and Algorithmic Transparency
AI is not omnipotent and still faces several significant challenges:
- Data barriers: Data from power grids, users, and equipment is often siloed; sharing this data while protecting privacy is a major issue.
- Algorithmic opacity: The logic behind AI decisions is often unclear, leading to distrust from regulators and users.
- Security risks: AI systems can be attacked (e.g., with fake data causing errors in predictions), resulting in trading losses.
Without resolving these issues, it will be difficult to widely adopt AI in the energy industry.
5. A Major Shift in Competitive Logic: From Competing on Installed Capacity to Competing on Intelligence
Previously, new energy companies competed based on the efficiency of their components and the size of their installations. Now, the focus has shifted to who has the most intelligent energy systems.
As an expert from Tianhe Fujia said, “In the past, we talked about components and batteries; now, we need to discuss the ‘intelligence of the energy system.’” The future of new energy is not about installing more equipment but about making the existing equipment more intelligent and maximizing the value of every kilowatt-hour of electricity generated.
In summary, the “good times” for the new energy industry are over, and companies must now rely on intelligence to thrive. AI is the key to breaking this deadlock, but there is still a long way to go. Understanding this trend helps us understand why many companies are advocating for “energy + AI”—it’s not just a gimmick; it’s a necessity for survival.