Here is what happened. On August 3, China’s top economic planner released the 15th Five-Year Plan for building a new power system. The headline: 300 GW of new-type storage in operation by 2030. The plan adds roughly 160 GW of new storage across the five-year window, including 80 GW of independent grid-side storage dedicated to peak supply.
Three numbers, one message. In short, the ambition became a schedule. The signal is not the size of the target; the signal is what the target forces next.
The number, split into its parts
That is not a small point, and I want to slow down on it, because it is the part most readers will skim. A storage asset with three revenue lines can be valued the way a power plant is valued: discounted cash flow, a hurdle rate, a repayment period. That changes who is willing to own it. Pension funds do not buy compliance obligations; they buy cash flows. The moment a storage project has a cash flow statement, it stops being a policy line item and becomes an asset class competing for real money.
There is also the question of what happens to the weakest assets. When every project must carry storage as a mandate, the marginal cost is hidden, spread across a portfolio, and nobody notices it. When storage must earn its keep, the projects that cannot clear the market simply do not get built, or get built with less storage. That is not a loss of ambition; it is the market doing its job of pricing a resource. The cheap megawatt-hour wins, and the system pays less overall for the same flexibility.
Here is what the next twelve months will test. The first provinces to adopt the autonomy model will produce the first real data on how much storage developers actually choose when the choice is theirs. That data will be worth more than any plan. It will tell the industry whether storage was always a good deal hiding behind a mandate, or whether the mandate was doing all the work. Both answers are useful, and neither is the answer the policy text itself provides.
Start with the arithmetic. 300 GW by 2030, against a fleet that entered the plan period far below that level. The gap is the opportunity, and the gap is the work.
Of the 160 GW of new additions, 80 GW is carved out for independent storage built to hold the grid steady at peak load. That carve-out is the detail worth pausing on. It tells you where the planner thinks the system will strain hardest: the minutes around the evening peak, when solar fades and the grid must find power from somewhere else.
I have covered storage policy since the days when a project’s storage was measured in a few megawatt-hours attached to a wind farm. The vocabulary has changed completely. The old question was how much must each project bolt on. The new question is how much the system needs, and who gets paid to provide it.
That shift in vocabulary is the story. Numbers change slowly; the frame around the numbers changed in a single document. A plan that names a target is a statement. A plan that names a target, a breakdown, and a revenue model is a contract with the industry. This document is the second kind, and the industry is already responding to it as if it were one.
The grid companies answered in three weeks
State Grid responded on August 25 with twenty measures. Its service territory will add 140 GW of new-type storage by 2030. It also plans to hold 120 GW of pumped hydro, in operation or under construction, by the same date.
China Southern Power Grid set its own line the same day: 60 GW of new-type storage by 2030.
Let me think about what those two answers mean together. The two largest grid operators in the world now carry separate, compatible, published schedules. One covers the north and center of the country; the other covers the south. Their numbers do not overlap and do not conflict. That is the difference between a target and a plan.
Let me reconsider the scale for a moment. 140 GW plus 60 GW is 200 GW of new-type storage across the two grid territories. Add the central plan’s 160 GW of additions, and you see how the pieces lock together: the grid companies are carrying the biggest share of the build-out, not the renewable developers. That is a structural change.
The pumped-hydro line deserves its own reading. 120 GW in operation or under construction by 2030 is the long-duration anchor underneath the whole battery fleet. Batteries give the grid four to six hours of flexibility; pumped hydro gives it days. The two are not competitors; they are different rungs on the same storage ladder. State Grid booking both at once is the sign of a system that expects to lean on storage every single day, not just at crisis moments.
No time to linger on the mechanics. The point is the timing: three weeks after the central plan, both grid operators published concrete numbers. In the policy world, that speed is itself a signal. It means the target was not written in a vacuum; the operators knew their numbers before the plan landed.
Shandong rewrote the rulebook
Then came the province-level signal, and this is the one worth the most time. On August 28, Shandong published new rules for storage attached to renewable projects. The key word is autonomous. New, operating, and grid-connected projects can now determine their own storage ratios.
There are limits. New solar projects can allocate storage up to 50 percent of installed capacity; new wind projects, up to 30 percent.
I almost framed this as another capacity mandate from above. Let me correct that framing, because it is wrong. The mandate framing is the old model: a regulator tells each project how much storage to install, and the cost is quietly folded into the project. Shandong flipped it. The plant owner decides the ratio and bears the cost. That is a different sentence, and a different incentive structure.
Why does autonomy matter? Because a developer who chooses the storage ratio is a developer who has to justify it. The storage must pay for itself, in the developer’s own spreadsheets, against the developer’s own revenue assumptions. When a regulator chooses for you, nobody has to run that calculation. When you choose, you run it every day.
The caps matter too. Solar at 50 percent, wind at 30 percent. Those are ceilings, not floors. A ceiling tells the market: storage beyond this point is not required, so build the amount that earns money. It converts a compliance problem into an economic one.
The payout model is the real news
The revenue side is where the accounting actually changes. Storage under the new arrangements can earn through three channels: capacity compensation, energy-market revenue, and ancillary services.
Capacity compensation pays the asset for being available, regardless of whether it charges or discharges. Energy-market revenue pays for the electricity the storage actually sells. Ancillary services pay for grid support at moments of stress, the voltage and frequency actions a battery does well.
Three channels, three different products. That is how an asset class stops being a cost center and becomes an investment case. A single revenue line is fragile; three lines hedge against the failure of any one.
I keep coming back to one observation. The hard constraint on storage was never the chemistry of the battery; it was the certainty of the revenue under the asset. A battery that earns from three sources can be priced, compared, and financed. A battery that earns from a single subsidy line can only be hoped for.
The difference between those two is the difference between a market and a mandate. Shandong did not just adjust ratios; it assembled the revenue stack that makes storage a bankable project. That is the sentence to underline in this whole story.
The arithmetic of the next five years
Now the practical read. The window runs 2026 through 2030. To add 160 GW in five years, the industry must install roughly 32 GW a year, every year, on average.
Let me check that arithmetic once more, because the pace is the point. 160 divided by 5 is 32. That is a steady annual rhythm, not a startup surge. The last cycle grew storage fast, but it grew in bursts. A steady 32 GW a year is a different operating rhythm for the whole supply chain: cells, containers, inverters, control systems, and the engineering queue at every point of interconnection.
There is a picture that stays with me from a grid dispatch center I visited years ago: the operators watching the load curve climb into the evening window, and the screens lighting up for the storage units to step in. That was a pilot then. The plan assumes that picture is routine by 2030, with 80 GW of independent storage built specifically for those minutes.
The plan is, in effect, a bet that the peak-supply problem can be solved with batteries and pumped hydro instead of new thermal plants. The bet may hold or may not; what is certain is that it now has a schedule. And schedules have a way of concentrating minds, manufacturers, and capital.
What to watch next
Three things, briefly. First, whether other provinces copy Shandong’s autonomy model. Second, whether the capacity-compensation price holds as the fleet grows and more assets compete for the same payments. Third, whether the grid companies hit their own published numbers.
The third is the one I watch most carefully. A grid operator’s storage schedule is a promise about its own territory. Those promises get tested twice a year, at the winter and summer peaks, when the load curve rises and the question is whether the storage actually shows up.
No time to linger on the risks, though. The direction is set. The storage industry now has a volume target, a build-out path, and a payout model, all within the same month. That combination has not happened before in a single policy window.
What’s next matters more than the headline. The headline is 300 GW. The next chapter is whether each gigawatt earns its keep, and whether the revenue model survives contact with a real winter peak.
In short: the plan sets the size of the room; the grid companies set the furniture order; the payout model decides whether anyone wants to sit down. Watch the last one.
And that is the signal worth taking away today: the numbers got a schedule, and the schedule got an accounting. The build-out is no longer a slogan. It is a ledger, and ledgers get audited at peak hour.