Apparently, it’s time to ban Edison’s venerable, now vilified, light bulb. European leaders, green pundits and the widely reported light bulb provisions of the U.S. Energy Independence and Security Act of 2007 all urgently push the abandonment of incandescent bulbs.

The plan appears to be to convince everyone to switch to compact fluorescent lights (CFL), a technology that was introduced in the 1930s and perfected when rock was young and computers used vacuum tubes.

What irony. Here we are, in the 21st century, with a semiconductor revolution in lighting around the corner. Semiconductor light emitting diodes (LEDs) are finally on the verge of having the capability to radically alter the entire lighting landscape with staggering improvements in both lighting efficiency and efficacy.

The high-power LED emerges directly from the same semiconductor, digital infrastructure and intellectual property that brought us iPhones, laptops, hybrid cars and Xboxes. Though the word “revolution” has been devalued by overuse, the LED is only the third revolution in illumination technology since the dawn of fire.

Nevermind, though. Even the technorati at Yahoo! have jumped on the CFL-mania bandwagon, launching a Web site called www.18seconds.org to encourage and calculate the benefits of taking 18 seconds to replace a bulb with the now-vaunted CFL. (There’s got to be a how-many-bloggers-to-change-a-light-bulb joke in there somewhere.) The retail might at Wal-Mart has pledged to push CFLs. And, in desperation to cut demand, the ever-faithful purveyors of yesterday’s solution, electric utilities, will promote and even give them away.

So why the rush to push them now? The answer: energy.

About 12 billion electric lights on the planet use Edison bulbs; a third are in the U.S. So, lighting up the world consumes about 2 trillion kilowatt-hours annually, or one-eighth of all electric power. This takes a lot of fuel: the equivalent of nearly a billion tons of coal annually. In the U.S., half of that is in fact coal. Or, in oil-equivalent terms, U.S. lighting uses the equivalent of 50% of the energy used by all cars on American roads.

The reason we’ve arrived at this state? People crave illumination, have for centuries, and the real economic cost of illumination has utterly collapsed. It is 1,000 times cheaper to light up a room today than at the dawn of the 20th century, and 10,000 times cheaper than in 1850. Consequently, appetites have soared for the central product of illumination technology, lumens. Humans consume 40,000 trillion lumen-hours per year. Focus on the lumen, not the bulb, to understand the future.

Until now, there have been only three illumination technologies: fire, from, first, torches, tallow and wax candles and, later, whale and kerosene oil lamps; electric incandescence, famously perfected in 1879; and electric-induced fluorescence, introduced in 1938 by General Electric. In 1962, Nick Holonyak, a GE engineer, invented the LED, marking the next pivot in illumination’s millennia-long march toward ever more ubiquitous, cheaper lumens. Holonyak may eventually become more famous than Edison.

LEDs are fabricated from various concoctions of “compound” semiconductors, mainly gallium and indium. All older forms of illumination use brute force to heat something till it glows. The central difference with an LED is that photons are emitted by designing exquisitely precise semiconductor junctions, tuned just so to employ elegantly efficient quantum phenomena that emit photons. To crudely analogize: It’s the quantum equivalent of using a butane lighter instead of rubbing sticks to start a fire.

Commercial LEDs were quickly introduced by 1968. Early ones dribbled out pathetically small flows of lumens and were ridiculously inefficient, converting fewer watts to lumens than even the notoriously inefficient incandescent bulb. Still, they were enough to start the revolution. The tiny, solid, robust and very long-lived LED was ideal for instrument and indicator lights (and cheesy digital watches).

With time, engineers rapidly refined the machinery of photonic semiconductors. Today’s LEDs are 1,000 times more efficient than they were in 1968. Comparable efficiency gains can be found only in the semiconductor world, such as the oft-cited Moore’s Law for microprocessors–named after Intel co-founder Gordon E. Moore–which states that information processors get twice as fast (or twice as cheap) every 18 months. LEDs have followed a lesser-known photonic equivalent, called Haitz’s Law, named after the LED researcher Roland Haitz. The law says LED brightness doubles every 18 to 24 months.

By the eve of the 20th century, the first high-brightness LEDs had emerged. Measured by the key economic and enviro-metric lumens produced per electric watt, LEDs are now about to blow past all antediluvian illuminators. LEDs have already achieved 100 lumens per watt, with 200 visible. Incandescent bulbs yield some 15 lumens a watt, CFLs about 80. LED pioneer Cree passed a related milestone last September, introducing a single, tiny LED that produced 1,000 lumens, the output of a standard bulb but at almost five times the efficiency.

Applications for LEDs have followed their rising power and declining cost, from instrument displays in the early days, to traffic signals and automobile brake lights and interiors, to back-lighting mobile phones and LCD screens ( Apple’s brilliant new Macbook Air), signature decorations (the 2007 White House and Rockefeller Center Christmas trees) and chandeliers in Buckingham Palace.

Annual global sales of high-brightness LEDs already total $4 billion, even before invading the 40,000 tera-lumen market for general illumination. General illumination is currently supplied about equally by incandescent bulbs (dominant in residential markets) and fluorescents (dominant in commercial buildings).

With the lumen power in hand and costs declining, there is one last hurdle for LEDs to overcome in order to conquer general illumination. The reason most people have resisted and will resist using CFLs, and that also currently impedes LED adoption (particularly for indoor lighting) is the quality of the light. If you hadn’t noticed, it’s lousy. CFLs must achieve the light quality of incandescence to enjoy wider adoption.

All lumens are no more equal than all calories are. The eye, like the palate, is exquisitely sensitive to quality. One thousand calories at McDonald’s doesn’t taste the same as 1,000-calorie foie gras. One thousand lumens from a CFL is visually harsh and makes food (and people) look ugly, compared with the same 1,000 from Edison’s incandescence.

Optical scientists cleverly named the visible mix rendered by the colors in light the Color Rendering Index. The CRI gold standard is numerically 100, the color of pleasing Edison incandescence. A CRI below 90 is noticeably nasty to the eye. CFLs produce and are essentially stuck, for physical chemistry reasons, in the CRI 70 range.

But LED lights have been inching up the CRI curve. LED Lighting Fixtures, for example, just started selling a 12-watt product with a pleasing 92 CRI that looks like a standard 60-watt indoor floodlight and uses 50% less energy than a CFL. Others will follow.

Add awesomely long lifespans to the benefits ledger for LEDs. An Edison bulb burns 1,000 hours before the hammered metal filament flakes out. CFLs achieve 10,000 hours. LEDs last 50,000, and soon 100,000, hours. This feature alone is enough to give LEDs an enormous advantage in lower maintenance costs for many applications, especially in outdoor and industrial environments.

The small size of the LED light chip allows another advantage arising from the physics of optics: Engineers can employ collaterally small lenses making it possible to efficiently extract, direct and focus the lumens, achieving useful illumination with far fewer lumens–and far less energy–wasted.

There are even more benefits from the semiconductor character of LEDs. Lumens can be controlled and modulated for both color and spatial distribution to create smart lights that adjust to the environment or need. Headlights or room lights could be spectrally tuned to match the innate differences in day, night and peripheral vision, providing more comfort and safety.

LEDs can also communicate by electronically modulating the lumens. Brake lights could talk with a trailing car’s cruise control; head lights talk with road signs to activate in-car audio/video warnings. Opportunities will be limited not by imagination but by how quickly costs collapse.

Lighting cognoscenti will note no mention thus far of an LED advantage over a CFL liability. CFLs contain tiny amounts of mercury. Individually not much, but billions add up, and no one really believes recycling will keep all that mercury out of landfills. Also, don’t count out more efficient incandescent bulbs emerging. While they may never be able to beat LEDs, these new alternatives could match CFLs and will enjoy, for much lower low capital cost, a huge advantage when lumens are needed only occasionally.

So who are the LED players? Opportunities abound in a market this vibrant and rapidly growing. Cree and GE are leaders, as previously noted, and though both are pioneers, the former is a pure play, the latter obviously not. Siemens is a major non-pure play, owning LED powerhouses Osram Opto Semiconductors and Osram Sylvania.

The other major player is Royal Philips Electronics , with its no. 1 position in global lighting. While not a pure play, it is moving there with nearly one-fourth of its sales from lighting. Philips has been aggressive in acquiring LED properties, buying out the Hewlett-Packard/Agilent piece of their Lumileds Lighting in 2005. Philips also acquired LED light pioneer Color Kinetics in 2006 (for $791 million) and TIR Systems for its white LED technology. If Philips shed its low-margin consumer electronics business, it could become the undisputed globally dominant near-pure play for 21st century illumination.

Count among the movers into LED-based fixtures Austria’s Zumtobel, with its long history in efficient lighting, or smaller pure-plays like England’s Dialight. On the LED chip-supply side there’s also Seoul Semiconductor, and there are LED semiconductor tool makers, like Germany’s Aixtron and Emcore Corp. and Veeco Instruments in the U.S.

IPOs are sure to emerge too, though many aren’t on anyone’s radar yet. Perhaps the U.K.-based Bespoke Lighting (responsible for the Buckingham Palace PR move), which recently merged with Los Angeles-based LEDtronics, will be the first.

Forecasts for a $12 billion LED market within four years probably understate what will happen. As Haitz’s Law continues to drive costs down, applications and markets will emerge in ways impossible to anticipate.

Think for a moment about the world’s 1.6 billion people who have neither modern lighting nor electricity. A modern wired electric grid and lights are a long way off for developing nations. Their lumens come from fuel-burning lamps, which collectively consume a surprising 1.3 million barrels of oil per day (equal to Qatar’s total production). But efficient cheap LEDs, combined with (soon cheaper) solar arrays and batteries, will permit a leap past wired lights, just as cellphones permitted a leap past landline phones in the unwired part of the world. A solar-battery LED light can already produce lumens at a fraction of the cost of a kerosene lamp.

Around the corner: a wireless lighting revolution, a phrase coined by my daughter on her return from working in rural Haiti, where she saw wireless phones were easier to come by than wired lights.

Investment opportunities aside, there are countless social benefits from the pending tsunami of pleasant electric incandescence from LEDs. Government regulators and policy wonks might try to claim they’re enabling a faster LED revolution with their CFL-centric ban-the-bulb movements. But the vastly better technology will take over when it’s ready.

And while that may not be literally tomorrow, the LED-centric future is now foreseeable. In any case, why would the market squander precious capital on yesterday’s technology on the eve of tomorrow’s? It won’t.

Mark P. Mills is a physicist and a co-founding partner of Digital Power Capital, an energy tech venture fund. Mills is also the co-author of The Bottomless Well: The Twilight of Fuel, the Virtue of Waste and Why We Will Never Run Out of Energy (Basic Books, 2005). Mills may hold positions in companies discussed in this column and may provide technology assessment services for firms that have interests in the companies. He can be contacted at inquiries@digitalpower.com.