Everyone loves a come-from-behind story. Here’s one: commercial nuclear power, rejected by society long ago and dying on the vine, is now coming back to life. And it’s happening in the most unlikely place — at Three Mile Island (TMI), where the wheels fell off 45 years ago.

On March 28, 1979, TMI’s Unit 2 was the scene of the world’s first-ever large-scale nuclear accident, a partial core melt — an event that triggered an abrupt change in prospects for commercial nuclear power everywhere. In short order, with public confidence in nuclear power badly shaken, new plant orders were canceled and industry plans for expansion set aside. In the decades since then, only two new nuclear units have been ordered and put into service in the United States.

But just last week, Constellation Energy, the owner of TMI-1 (the unit at that site that was not damaged in the nuclear accident and had been in operation until 2019), announced an agreement with Microsoft to refurbish and restart the plant and to provide all of the electricity it generates over the next 20 years to power Microsoft’s Artificial Intelligence (AI) projects.

It’s a blockbuster deal, another huge step in our nation’s essential effort to revive nuclear power. And for me, it’s personal: My entire professional career was in nuclear energy. I was employed by the firm that originally owned TMI, I was part of the emergency response team called in to deal with the 1979 accident, I spent five years leading the planning and engineering for the post-accident cleanup, and in subsequent years, I was heavily engaged in nuclear industry efforts to get nuclear back on track. Now that’s finally happening.

The U.S. nuclear odyssey — from feast to famine and now back — has been bizarre. In the 1960s and ‘70s, we who worked in the nascent nuclear power industry saw ourselves as pioneers of the nation’s energy future. Then, overnight, we became public enemy number one, supposedly trying to foist our frightfully dangerous technology on an unsuspecting public. Now, we’re suddenly back in good graces, offering the one and only safe, clean, reliable 24/7 electricity supply.

It’s been a wild ride, with lessons worth remembering.

First, what went wrong? The TMI accident didn’t kill the nuclear industry, but it sure helped. The accident destroyed a new billion-dollar plant in just over two hours. No one was hurt, but millions were terrified, and it gave apparent credence to the fledgling anti-nuclear movement and its predictions of doom.

That movement became an imposing force, found a welcome home with the political Left, and, over time, honed its tactics in blocking the expansion of U.S. nuclear power. Nuclear opponents learned to present legal challenges during proceedings leading to final NRC issuance of operating licenses — the point at which years and billions had been spent on the new plant, but no revenues had yet materialized. The devastating financial consequences of long delays — and possible cancellation — at that point tended to spike, once and for all, investor confidence in new nuclear construction.

In short, our nation’s unwise rejection of nuclear energy was not a conscience policy decision but the composite effect of understandable but relatively short-lived public discomfort, investor uncertainty about nuclear, and the availability of less controversial alternative energy sources (primarily natural gas). That had long-term consequences for our nation’s economy, health, and security.

And now, nuclear is back on the table. What changed? And just as importantly, what didn’t change?

Nuclear power didn’t change. With 70 years under our belts, our remaining U.S. nuclear plants are running just fine. Nuclear is still remarkably safe, reliable, and clean — just as it was when we foolishly decided to set it aside.

What has changed is simply the long-overdue recognition of the practical realities of energy supply (along with the obvious question of what took so long).

The wake-up call was an unexpected consequence (and arguably the single greatest benefit) of the furious U.S. and global efforts to combat the “existential threat” of climate change by precipitously reducing reliance on fossil fuels. The net effect has been correspondingly precipitous increases in electricity cost and decreases in availability — all compounded by the looming increases in electricity needed to support growing populations, electric vehicles, and new technologies, particularly AI.

So what now? It’s way too soon to cheer the long-awaited rebirth. The Constellation/Microsoft deal is just one of multiple investment vehicles that will help to get the ball rolling. But there are steep hills still to climb.

TMI-1 is, in effect, an antique nuclear plant designed and constructed in the 1960s. It has already operated far past its initial licensed life expectancy. Its basic reactor components are sound but will require extensive examination and, in some cases, refurbishment. Other aspects of the plant, particularly its instrumentation and controls, will need to be upgraded. And all must meet rigorous inspection and approval by the Nuclear Regulatory Commission. Constellation is projecting the restart effort to cost $1.9 billion and take about four years. It’s a very big job.

More importantly, refurbishing old plants is not the path to a viable long-term nuclear future. Our currently operating nuclear plants are growing old together and will need to be replaced by an entirely new fleet. Fortunately, the brightening landscape will encourage the investment needed to reinvigorate a U.S. nuclear industry that has been effectively moribund for decades — plant design and construction, supply chain of specialty equipment fabrication and design, training and hands-on skills of construction and operations personnel, and the like — and make that possible.

But it’s all doable because the need is compelling and the underlying technology is sound.

The takeaway lessons from all of this? I see two:

1. Our democratic processes properly call for full and transparent public engagement. But in this electronic age, the opportunity for misinformation and political manipulation of public opinion is higher than ever, and the consequences of bad decisions based on misinformation can be catastrophic.

2. Some matters should transcend politics. Energy is one. For decades, one political side has been pro-nuclear and the other anti-nuclear. Related matters are environmental protection and climate change, with similarly sharp partisan divisions. That makes no sense at all. Surely, we can stop squabbling and put our heads together on fundamentally important, apolitical issues.

With an election just six weeks away, it’s a good time to keep those perspectives in mind.