There's potentially a problem with LED lighting that I wasn't aware of until a friend sent me an article on the subject. It's something you probably should be aware of if thinking of switching to LEDs now that incandescents are largely out of production.
I haven't paid a lot of attention to LED lighting. Until recently it's been expensive, more useful for narrowly focused applications like task lighting, and not that much more energy efficient than the CFLs we use almost everywhere inside and outside the house (there are a few incandescents in the carport that have been there about 15 years and counting - not used much). Additionally, all LEDs are made with either the element gallium (Ga) or indium (In), and both of those are not very abundant on this planet. That leads to questions about sustainability for something produced in 100s of millions of units, like light bulbs.
The article - That 60W-equivalent LED: What you don’t know, and what no one will tell you… - is from EDN (Electronic Design News), which is a largely apolitical magazine written by and for engineers. If anything, their bias would be to get you to buy more of their advertiser's products, like LEDs.
You should read the entire article (it's non-technical), but here's the money quote:
Within an LED bulb the internal generation and distribution of heat is such that it “desperately” needs access to cool surrounding air. The fact that it has that metallic housing is irrelevant in restricted air.
That 60 watt Wal-Mart bulb, when operating base down in open air and not even using a shade, has its internal LED case at 85°C, the absolute upper end of what is considered “safe” for full life expectancy. The same deal is true for competitive bulbs. Put a shade around it... and it’s a little warmer. Put it into any kind of base-up socket and it gets a lot hotter and all life expectancy numbers are off the table. Put it into any kind of porch or post light fixture, and it can fry, with its internal power supply components at the cliff edge of failure. Put the lamp in a ceiling-mounted fully enclosed fixture and set the timer for when failure will occur.
In other words, totally unlike incandescent and substantially unlike a CFL, reliability and life expectancy go down hill sharply as soon as you install it anywhere that air is restricted. Guess what? A large percentage of places for LED best value is in those place where access is difficult and air is restricted. LEDs do not target a “table-lamp-only” marketplace.
Some related information: 85C (185F) is the maximum operating temperature for most electronic components designed for industrial use. Consumer products usually top out at 70C, military and aerospace at 125C. Per the article, full life expectancy is quoted for operation at 85C, so that's not the problem. The problem is that in an open application, like a table lamp, the LED bulb is already at 85C. When you enclose it in something, like a globe or ceiling fixture, it will get hotter - a lot hotter.
There's a rule of thumb reliability engineers use that comes from Arrhenius (the same guy who outlined the greenhouse effect around the end of the 19th century). Arrhenius said that chemical reaction rates double for every 10C rise in temperature. For estimating reliability, you turn that around and estimate that product life gets cut in half for every 10C rise in temperature. A lot of faliure mechanisms are, or act like, chemical reactions. That's the point the author of the article is actually making.
Also note that this has little to do with the actual wattage or power consumption of LEDs vs. incandescents. LEDs do use less energy, but if they don't cool as efficiently as light bulbs in an enclosed fixture, they can still get hotter. That seems to be the case here, and for all manufacturers - not just Wal-Mart suppliers. This isn't probably a safety issue - houses won't burn down - because a lot of light fixtures have built in thermal cutouts that will shut off electricity if they get too hot, or the electronics that operates the LED (LEDs operate at around 1.8 volts) will fail.
The other important factor is the backlash that this kind of situation can create. People buying LEDs bulbs are expecting lower electric bills and bulbs that last a long time. They can end up very disappointed when the hype about lifetime turns out not to be true in some applications. That can damage the credibility of people who advocated the solution. As a corollary to that, it's important to understand what Germans might call the Gestalt, others the whole enchilada, or 60s refugees the holistic nature of solutions to environmental problems. Just because something solves one problem - like reducing energy use - doesn't mean it satisfies all the other constraints it needs to, like lifetime. Completely understanding a technology is a good idea if you want to avoid unintended consequences.
As I mentioned at the outset, outside of a couple of LED flashlights - which I like a lot - I don't have any direct experience using LED light bulbs. I do have a fair amount of experience in electrical engineering, particularly reliability and temperature problems, and the guy who sent me the article link has a lot more experience. The engineering basis of the article seems correct to me.
It seems like something worth being aware of.