Google's 'Smart Grid' idea? Get the govt to pay for it

Then there's the other direction to be considered. Consumers might also program their houses and cars to sell juice back to the grid: If you didn't want your car for a couple of days you might program it to charge up at cheap times and flog its juice back to the company during price spikes. This would tend to smooth out the price graph, making the market nicer for renewables. Likewise you might sell home-generated solar or wind 'leccy, though you'd probably struggle to make money doing that - you would generally be trying to sell when the local spot price was low.

At present, this sort of capability is pretty difficult to deliver. Power companies seldom have any idea when an individual consumer is drawing what amount of power, for a start. Even if they did, it's still not clear that they would want to simply take a small commission on the spot price rather than charge a flat rate as they do now. In the end, if consumers save money they get less revenue.

In any case, spot energy price is a very complicated thing - especially in America where there isn't any central clearing house like the UK National Grid. Again, the government would probably need to weigh in heavily to the wild and woolly US energy market to get this kind of thing to work. Someone would also have to pay for the parallel information network to run alongside the grid, reading every house's meter back to the company and the company's prices out to the house - in near real time, very securely and without errors.

Even if Google's push in Washington all pans out and a splendid new, nimble, open and well-informed US 'leccy market (or smart grid) appears, the game may not be worth the candle. Most people use their cars on a daily cycle most of the time - they can't shift their charging up periods to deal with days-long calms or cloudy spells. The existence of extra, swappable vehicle batteries to deal with inconveniently long charging-up times could help, but this may already be an obsolete idea.

Cambridge physics and computing prof David MacKay, considering this, reckons that in any low-carbon UK future where the proportion of renewables gets really serious, even the universal use of smartly-gridded electric vehicles simply can't smooth out the supply-demand swings. He suggests enormous pumped-storage facilities to fill the gap, though he doesn't say who would pay for these - perhaps, again, Google plan to hit up the taxpayers.

Consulting engineer Jim Oswald, working for the Renewable Energy Foundation, says even this won't work. Pumped storage could deal with a couple of days' calm or clouds at the outside, but he says the system must be able to cope with five-day renewables droughts across western Europe annually - and really it would be better if it could handle ten-day dropoffs, as these can be expected every twenty years or so.

So your grid can be as smart, as agile as you like, but it still won't cope above a certain proportion of ordinary renewables - solar, wind, water etc. There will be a need for dependable, reliable energy.

Google does seem to concede this, despite a natural tendency to think that IT can solve any problem; but it still can't bring itself to go with the more normal nuclear or carbon-sequestration baseload ploys. Instead it reckons the answer is geothermal. Ordinary geothermal is something of a busted flush, except in rare cases - thus Google will put perhaps a thousandth of a year's ad revenues into developing "enhanced geothermal", and will - again - call upon Washington to dig still deeper.

Google believes that smart grid + lots of ordinary solar / wind unreliable renewables + reliable geothermal = a viable plan to power western civilisation at a cost less than using coal (the so-called RE<c scheme). It appears that GE will do the necessary hardware, and the feds will cover most of the costs.

At one level, it's a surefire scheme. As coal eventually runs out, it will become very expensive. Even a combo wind/solar/geothermal/smartgrid solution will be cheaper one day. RE really will be < C.

Still, though. Apart from maybe pitching in on the software and the shouting for taxpayer money - though it's hard to see why it would naturally be good at either - what does Google bring to the RE<C party?

Well, not much. But it's Google's party - or anyway Google says it is. ®