California’s three investor-owned utilities have released the results of a study called for by the California Public Utilities Commission to assess the “effective load carrying capability” of various energy technologies—in other words, how much of a wind or solar plant’s total theoretical capacity can be counted on when the grid needs it most.

The findings are striking and significant for California’s future electricity mix, and really for the future of any grid aiming for, say, 100% carbon-free electricity by 2035. The study determined that wind energy will have a 19% effective load carrying capability in 2022, a wonky way of understanding how much wind power can get from plant to the grid in moments when the grid needs it most. A 100-megawatt wind power plant, in other words, would be capable of supplying 19 megawatts of energy.

That value is fairly stable out to 2030. Add four hours’ worth of energy storage in the form of a battery and it performs better. But that’s still far from a natural gas-fired power plant that can almost always be relied on to generate its peak capacity if needed.

Solar is not wind. Industrial-scale installations are close to useless at managing electrical load when the risk of blackouts is highest, and it gets even worse at residential-scale, with an effective load carrying capability of just 4% in 2022.

But add four hours of energy storage and solar’s capabilities change dramatically. A solar project with four hours of storage in 2022 will have a 99.8% effective load carrying capability—essentially a total ability to ensure the grid remains reliably supplied with power. That value declines slightly over time, as more solar comes on line, but by not much.