15GWh Energy Storage, 30 Billion kWh Transactions: Huanghe Digital Energy Raises Tens of Millions of Yuan in Pre-A Round Financing, the Most Costly Link in the Energy Storage Sector Has Finally Been Separately Priced
Energy Insight learned that Huanghe Digital Energy recently completed a Pre-A round of financing of tens of millions of yuan, jointly led by the North China Electric Power University Education Foundation and a leading enterprise in the energy storage industry chain. The use of funds is very clear: the engineering implementation of the vertical power industry model, the large-scale replication of energy storage asset operation capabilities, and the expansion of the R&D team.
In the same period, as the only new-type market participant in the power sector, Huanghe Digital Energy was invited to join the "Education Community" jointly established by North China Electric Power University and State Grid. This seems to reflect that in the era when AI is reshaping all industries, power, as the "life source" of AI, is also in urgent need of compound talents who understand the market, power and AI. The field that this company is deeply engaged in provides a fertile test ground for universities and the power grid side.
The company's model is quite unique — as an investor that holds energy storage power stations by itself, it also entrusts the operation of multiple third-party power stations. Up to now, its total accumulated operation scale has exceeded 15GWh, and the total accumulated traded electricity volume has exceeded 30 billion kWh.
Objectively speaking, this financing scale is not prominent in the current energy storage track, but the direction it bets on is worth examining: when the revenue of energy storage power stations increasingly depends on who makes the "charge and discharge" decisions for them, whether the operation capability itself can be separated from specific people and specific provinces and precipitated into a set of standardized capabilities that can be delivered in batches, is becoming a key variable that determines the industry pattern.
01 Operators of energy storage power stations are becoming a variable in the asset pricing chain
The asset attribute of energy storage power stations has undergone an easily overlooked change in 2026. In February 2025, Document No.136 promoted the full access of grid-connected new energy power to the market; in January 2026, Document No.114 included independent energy storage into the capacity tariff mechanism. After the rules are unified, national power assets begin to face the same set of price signals, and the decision-making power of revenue shifts from "generating more power" to "selling power at better prices".
On April 28, 2026, the "Caitong Asset Management - Canadian Solar Held Real Estate Asset-backed Special Plan (Carbon Neutrality)" was established, becoming the first inter-institutional REITs with energy storage power stations as the underlying assets in China, and was listed on the Shanghai Stock Exchange in June. In August, the 304 million yuan energy storage REITs project of Huanghe Tianwei was accepted by the Shanghai Stock Exchange. What REITs value is predictable cash flow, and whether the cash flow of energy storage power stations can be predicted depends on operation. This is not an inference, but a question that financial institutions must answer during due diligence.
However, an energy storage power station is by no means a machine that automatically generates revenue. China Post Securities has made a set of calculations based on a 100MW/200MWh system: under the assumptions of EPC at 1.6 yuan/Wh, cells at 0.5 yuan/Wh, and 260 days of operation per year, the after-tax IRR of independent energy storage in Shanxi is 5.9%, and that in Shandong is -1.8%.
The source of the gap is not the battery, but the revenue structure. The revenue in Shanxi includes primary frequency regulation, contributing about 63,000 yuan per day on average; the benchmark spot price spread assumption in Shandong is only 0.4 yuan/kWh, 0.1 yuan lower than that in Shanxi, and the capacity compensation is limited.
With the same set of equipment, crossing a provincial boundary, the account book is completely different.
This is why operation is pushed to the forefront. Asset holders can be industrial capital, financial institutions, and local state-owned enterprises, but making a power station complete every declaration, every charge and discharge without errors in the rules of its own province is becoming a matter that requires special capabilities. The particularity of energy storage power stations is that they play two roles at the same time: they are power sources when discharging, and loads when charging. The revenue needs to be superimposed from three directions: capacity tariff, spot arbitrage and auxiliary services, and the weight of each layer varies from province to province and continues to drift — wind power and photovoltaic only need to manage the source side, while energy storage needs to put role switching, market declaration and battery boundaries into the same decision-making, which is the new energy asset with the longest decision chain among the three types.
Hengyi Group entrusted Xinghe Microgrid with the full life-cycle operation of its 113.5MW/226.72MWh energy storage power station cluster, and the contract covers real-time monitoring, fault response and power trading assistance. The bid price for the operation and maintenance service of the 200MW/400MWh shared energy storage power station of Jiyang Green Energy Storage is 48 million yuan (including tax). Guoneng Rixin provides full trusteeship service of power trading for a 100MW/400MWh energy storage power station in Yongning County, Ningxia, with a trusteeship fee of 1.5 million yuan per year. Third-party operation is no longer a concept, it already has quotations, contracts and settlements.
Huanghe Digital Energy has a specific position in this sequence: its own power stations and entrusted power stations run on the same set of models. The 15GWh operation scale accounts for a limited proportion of the total installed new energy storage capacity in China, but the problem behind it is becoming increasingly prominent: who operates the power station, like the province where the power station is located and the configured duration, has begun to have a substantial impact on asset pricing.
What supports this closed loop is the self-developed vertical power industry model: taking the electrochemical and thermodynamic mechanism model as the prior boundary, and using the operation data of the power station for online calibration, so that the conclusion of the model is upgraded from statistical correlation to physically verifiable.
02 Why this company, and why now
Energy storage operation has been obscured by an overly smooth assumption in the past two years: The more accurate the AI makes the electricity price forecast, the higher the revenue of the power station.
This assumption omits all frictions between forecasting and settlement. The clearing cycle of Shanxi spot market has been shortened from 15 minutes to 5 minutes, generating 288 price signals every day, and the price spread fluctuates sharply within a day. Forecasting is only the starting point. From declaration to clearing, and from clearing to settlement, every link is faced with deviation assessment and physical constraints.
The statement of Huanghe Digital Energy is to connect forecasting, optimization, declaration and settlement into a closed loop, and all model outputs must go through the hard constraint verification of the safety rule engine before being converted into issued instructions. This statement sounds like technical language, but the reality it corresponds to is very simple: The barrier of energy storage operation does not lie in the decimal point of prediction accuracy, but in the series of error-prone links from the nominal price spread to the net cash flow.
It is no accident that Huanghe Digital Energy started from Shanxi. The company was initiated and established by Huanghe Industry, and its shareholders include leading private energy enterprises in Shanxi such as Dongyi, Xinfei, Yaxin, Meijin, Kaijia and Huaxin. According to public data, the total annual raw coal production capacity of the above six enterprises alone has exceeded 26 million tons, with nearly 1.3 billion tons of retained coal resource reserves, and the total annual metallurgical coke production capacity has exceeded 7 million tons. A group of enterprises that have completed primitive accumulation through coal and coking have invested their capital and organizational capabilities in energy storage operation — this scale determines that it can afford the long-cycle investment in the initial stage of operation capability precipitation. Shanxi is also the first officially operated power spot market in China, with 5-minute clearing and 288 price signals a day, which provides a high-intensity training environment for this set of models. Not every energy storage operator has such a training ground.
The scarcity of talents is another underestimated variable. According to the report of Boss Human Resources Group, the talent gap in the energy storage industry in 2026 is 78,000, with a gap rate of 24%. But what is really scarce is not general energy storage talents, but people who understand both the changes of power market rules and the physical boundaries of energy storage assets. The scarcity of such people, in turn, explains why "entrusted operation" has become an independent market: not every power station holder can afford and retain a trading team that can continuously iterate with the rules of each province. The trusteeship project of Guoneng Rixin and the goal of Haibosichuang to operate 35-40GWh energy storage assets by the end of 2026 point to the same thing: the supply side of third-party operation is being pushed forward by both capital and organizational capabilities.
03 The reference frame of "Chinese version of Octopus" and what it really wants to verify
Calling Huanghe Digital Energy the "Chinese version of Octopus" is a convenient analogy, but the problems faced by the two companies are different.
UK's Octopus Energy completed the spin-off of its Kraken platform at the end of 2025, with an independent valuation of 8.65 billion US dollars. In January 2026, it entered the Chinese market by establishing a joint venture Bitong Energy with Bicheng Energy. The capability of Kraken is built on the aggregation of millions of household batteries, heat pumps and charging piles, with small single-point scale and fine decision granularity, but the decision complexity of a single device is low. What Huanghe Digital Energy faces are independent energy storage power stations and wind-solar power stations, with a single station measured in hundreds of MWh, and the cost of one declaration error is not at the same level as that of a household battery.
There is no doubt that the premium space of power generation side operation is larger. The larger the asset and the longer the decision chain, the more obvious the revenue improvement that professional operation can leverage. However, the replication of power generation side operation is more difficult. The model calibration of a power station requires real working condition data, and the strategy launch in a new province requires the verification of local settlement rules. On the other hand, the capability boundary of energy storage power stations is naturally wider: it has both source side and load side roles, while wind-solar power stations only need source side capabilities — forecasting output, reporting quantity and price, and undertaking assessment; virtual power plants only need load side capabilities — aggregating resources, responding to dispatching, and settling incentives. Both of these two types of capabilities are subsets of the capabilities of energy storage power stations. This gives "large-scale replication" a larger space than imagined: managing the most difficult type of assets means that the capabilities of the source side and the load side can be migrated downward at the same time.
Huanghe Digital Energy is currently running L2-level automatic declaration, and the agent completes forecasting, quotation and declaration in limited scenarios with on-duty manual supervision. The L3 level has only verified the closed loop of limited scenarios in the sandbox environment. This means that its operation capability still has the trace of "human", and there is still a distance from completely breaking away from the specific team and becoming a standardized product.
Among the uses of Pre-A round funds, "large-scale replication" is the most noteworthy item. This sum of money cannot buy scale, but can only buy time. What it wants to verify is not whether 15GWh can be turned into 30GWh, but whether a set of operation capabilities developed in the Shanxi spot market can reproduce the same settlement performance after leaving Shanxi.
In 2026, third-party energy storage operation is no longer a problem that needs to demonstrate "whether there is demand". Hengyi's contract, the winning bid price of Jiyang Green Energy Storage, and the trusteeship fee of Guoneng Rixin are all evidence of the existence of demand. The real unanswered question is whether this capability can be delivered in batches, and how much premium remains when it is delivered in batches. Huanghe Digital Energy's Pre-A round is testing exactly this. When the top-down electricity price reform meets the bottom-up accumulation of operation capabilities, this "local octopus" rooted in the soil of China's energy storage power stations has just stretched out its first tentacle.
This article is from the WeChat official account "Energy Insight", and is published by 36Kr with authorization.