A few months ago, the Nuclear Regulatory Commission granted permission for initial site work to begin on new nuclear reactors in the United States for the first time since the 1970s. Georgia Power, a subsidiary of the gigantic Southern Company, plans to build the two new reactors at its Vogtle nuclear plant, near Augusta.

At first glance, I was all in favor of new nuclear construction. Among other reasons, it’s high time we stopped determining energy policy on the basis of a bad Jane Fonda movie. But as a Georgia Power customer — who’s already on the hook for part of the bill for the new facilities — I’m scratching my head a bit over both that price tag, and over the rationale for going back to the old model of massive, complex, and hugely expensive power plants.

The planned Votgle upgrade is estimated to cost around $14 billion, and each reactor will produce around 1250 megawatts of electricity (MWe). The new reactors will be added to two existing units which were completed during the 1980s.

The cost of those two original units, estimated at the time to be around $660 million, skyrocketed to nearly $9 billion in the wake of the post-Three Mile Island regulatory blizzard. That jump in costs, which was typical for the industry, effectively ended new nuclear plant construction for a generation.

As time passed and 70’s anti-nuclear hysteria ebbed, power companies around the country have petitioned the NRC for permission to build new reactors. Some 16 applications have been filed since 2007, with more anticipated.

All the current NRC applications have one thing in common: they’re for large-scale power plants, technically improved but functionally not dissimilar from the reactors of the 1970s. Today, Jane Fonda is a punchline, Real People is long since off the air, and disco is blessedly still dead, but the big electric companies remain stuck in the ‘70s as far as their strategic planning is concerned.

While political conservatives generally look favorably upon nuclear energy, the economics remain daunting. In a now-famous paper for Reason, Jerry Taylor of Cato said nuclear power “is to the Right what solar is to the Left: Religious devotion in practice, a wonderful technology in theory, but an economic white elephant in fact.” Taylor referenced industry studies showing nuclear electricity costing four to five times as much per kilowatt hour than coal or gas plants, and noted the massive subsidies and loan guarantees handed out to power companies as undermining the cost rationale for nuclear power.

All of which makes me wonder, again: this is the 21st century — why are we looking at huge, multibillion-dollar facilities in the first place? It’s not like other options don’t exist.

Take for instance the Hyperion Power Module, or HMP. Developed at, and then spun off from, the Los Alamos National Laboratory, Hyperion is marketing the diametric opposite of the power companies’ massive and complex facilities. Hyperion’s reactor is a relatively tiny device, about the size of a dinky Smart Car.

Unlike large-scale plants requiring 24/7 monitoring by a small army of engineers and technicians, an HPM contains no moving parts, and is intended to operate for years with no human interaction to speak of. Hyperion reactors are actually intended to be buried underground during their service lives, with no hands-on maintenance at all between refueling cycles, which occur every 7-10 years.

Of course, a single Hyperion unit is hardly the equivalent of a Westinghouse AP1000 reactor, two of which are planned for the Votgle facility. One HPM generates only 25 MWe, while a massive AP1000 churns out an appropriately massive 1250 MWe or so.

But nobody ever said you have to buy just one. If we assume that a single new AP1000 costs about $7 billion  for 1250 MWe (which is not entirely fair as “sticker prices” go, since the $14 billion estimate for the Votgle plant upgrade includes financing costs as well as actual production), that works out to about $5.6 million per MWe.

A single HPM currently lists for $50 million (and I should note here that this is already twice the price Hyperion promised in its initial 2008 press releases). At 25 MWe per unit, we’re looking at $2 million per MWe, a little more than a third of the unit price of power from an AP1000.

Hyperion says its reactors aren’t intended to replace large-scale generation plants, but the engineer in me wonders, why not? HPMs are built on an assembly line, and Hyperion already has over 100 orders for them. Picking up my calculator again, I figure that in order to equal the output of one AP1000 reactor, I’d need to buy 50 HPMs.

At $50 million per unit (how about a bulk discount?), that would cost $2.5 billion. Now, I don’t have that kind of cash laying around myself, but you don’t need to be an accountant to see that $2.5 billion is a lot less than $7 billion. And that doesn’t count the untold millions I’d have to spend on the aforementioned army of maintainers for the AP1000 — although either way, you’d need a sizable team of regular power plant workers to maintain the actual power turbines.

I’m sure that these simple, back-of-the-envelope numbers don’t reflect anything like every detail of big vs. small in nuclear power, but a Hyperion or similar small-scale reactor would have to get a heck of a lot more expensive to cost as much as big, traditional plants.

There would also be other benefits, in that you wouldn’t have to locate the entire power apparatus out in the middle of nowhere. Hyperion-style reactors can’t melt down, and are designed to be buried in small plots. Why not use that easy portability to distribute your power plants all over the place? Put a couple near your city’s main hospital, a couple more in your industrial zone, with single units scattered around the suburbs and residential cores, and you’ve got a redundant system that’s far less susceptible to, say, blackouts during bad weather, as opposed to running power across hundreds of miles of transmission lines.

So, Georgia Power, Nuclear Regulatory Commission, et al — why aren’t you thinking small?