This fall, the wind stopped blowing in much of Europe. Hundreds of onshore and offshore wind generators stopped turning. The resulting energy shortfall has helped (combined with other causes) create a crisis that is bringing months of hardship and a massive spike in the price of gas and energy. As world leaders congregated in Glasgow, they got a first-hand look at the downside of Europe’s growing dependence on energy sources that depend on weather.

America is in danger of making a similar mistake. Since World War II, America has been powered by a fleet of coal, gas, and nuclear power plants that deliver 24/7-reliable, what is called “baseload,” electric power. That power is always on call and dispatchable as power needs rise and fall with demand. But in the last few years, we’ve begun closing those plants with no plan for similarly reliable replacements. Our key competitors, China and Russia, would get great benefit if we continue to close our baseload plants before we have deployed clean replacements.

What’s the problem? Public discourse and much policy have overlooked the critical difference between essential and dispatchable baseload power and helpful supplemental power. Baseload-power is by definition reliable—something that’s dependable and always there when you need it, 24 hours a day, 365 days a year. Supplemental power, on the other hand, is supplemental precisely because it is not reliable, not always available. Wind and solar power, for example, generate very inexpensive power that can supplement baseload generators, reduce fuel costs, and cut emissions. But obviously they are dependent on the weather and thus can only be a supplement to, and not a replacement for, baseload power. As Europe is discovering, they are supplemental precisely because they don’t always produce power when needed.

For many reasons, America’s efforts to clean up its electricity generation have been focusing almost entirely on supplemental power—intermittent renewables. And for good reason. Adding supplemental wind and solar power is one of the easiest and quickest things we can do to reduce emissions.

But we also need power when the weather doesn’t cooperate—clean, reliable, baseload power. Is it possible to store enough energy to tide things over when the weather is bad? Unfortunately, no. At this time, there is no battery or other storage technology practicably capable of storing the massive amount of energy it would take to keep an economy running throughout days or weeks of widespread, uncooperative weather. The world’s largest battery, Tesla’s $66 million Australian battery, for example, would power just one state, Illinois, for only two minutes. New innovation might substantially improve that performance, but it is inconceivable that it could happen at the scale needed for the economy and in the timeframe needed to prevent disastrous climate change.

Realistically, wind and solar renewables will play an increasingly important part in our electricity systems, but we must realize that they can only supplement baseload generators and not replace them. This is the often-unrecognized issue. If we fail to deploy clean and dispatchable baseload power, the need for toxic but reliable coal and gas baseload generation will not go away. Germany is a real-world example. Despite over half a trillion dollars spent on its energiewende and supplemental power, Germany has had to open a new coal mine and build more fossil fueled baseload plants. Its emissions are predicted to hit an all-time high this year. The climate will not tolerate duplication of Germany’s experience.

What’s the solution? At present there are three promising technologies capable of the scale necessary to provide economy-wide clean baseload power: fossil generation where the carbon is captured and stored underground (carbon capture and storage, or CCS); Allam cycle generation, an improved version of gas CCS; and some form of nuclear generation. A major advantage of these technologies is that they can often use the infrastructure at existing power plant sites and thus deliver huge cost savings with a minimum disruption to the environment or people’s way of life. This is particularly true of nuclear’s up-and-coming advanced small modular reactors (SMRs). The new SMRs will be super-safe and many are self-extinguishing—when the fuel gets too hot, it’s thermal expansion tamps down the reaction. Moreover, many designs make little or no waste, they can be built in factories quickly and less expensively and can be trucked to existing sites.

Although we’re doing some great work on SMRs and other technologies, such as deep geothermal, it’s not nearly enough. To replace our fossil-fueled plants and get truly clean and reliable power by mid-century, we need to massively scale up and accelerate the development and deployment of clean, dispatchable, baseload power—as we did by providing huge subsidies and incentives for wind and solar. If we got intentional, the first SMRs could be online before the end of 2025. And, well before 2050, we could have great new jobs and fleets of affordable, clean, baseload generators that would give us the reliable, dependable, clean power that will be essential to a modern, competitive, clean, and vibrant economy. It’s a fantastic opportunity.

But like anything else these days, there are serious hurdles faced by each of the contenders. CCS is currently very expensive and takes lots of energy. Also, both Allam cycle and gas CCS will require new pipelines for the transport and disposal of the carbon dioxide. Geothermal is very expensive and its scaling potential is unknown. Nuclear power has a serious PR problem. It suffers from years of misinformation that has much of the public believing it is dangerous with an unsolvable waste problem. Neither of these things is true. All the data say that per KWh dispatched, nuclear is safer than any other form of energy. The waste from civilian power reactors is small compared to many other toxic waste streams, including from solar cell mining and decommissioning. Nonetheless, fear and politics has produced regulations that drive the cost of new nuclear into the stratosphere. But the future can be different. The small size and self-extinguishing feature of most of the advanced SMR technologies should allow a new class of regulations that will make advanced nuclear power plants much cheaper and faster to build. That will have to happen before the United States can get the new reactors quickly online, deeply clean emissions, and stay competitive.

The takeaway: When it comes to power, reliability matters. Supplemental power like wind and solar certainly helps cut emissions. But supplemental power does not have the reliability required to run an electrified economy. As Germany is discovering, an industrial economy needs 24/7 reliable, dispatchable, baseload electricity. As America electrifies more and more of its economy, it will need clean baseload electric power. If we are to come even close to net zero, clean baseload power must get the same effort we’ve been giving supplemental power. Best of all, if we seriously embrace clean, reliable baseload power, we can usher in a new era of wonderful prosperity and environmental stewardship.