Again, to each his own. It really depends on
your definition of cheap and what you mean by being a bargain shopper. How far into the future are you accounting for? With my 360g build, I'm calculating every single cost based on a 10 year lifetime. If I can save $150 by buying a Mag 18 instead of a Reeflo Snapper, I'm not going to, because the Mag will cost a LOT more than $150 above what it costs to run the Snapper for 10 years. So, I'd buy the Snapper instead of the Mag, even though it's more money up front. Same with the LEDs. Spending an extra thousand bucks up front is PEANUTS compared to the savings over 10 years.
It varies slightly from LED to LED, but most of the HP LEDs people are using are rated to drop to 70% output at 50k hours, run at 125c junction temp and 700mA drive current. Our typical fixtures are likely running at much cooler temps, so lifetime should be far longer (following the standard "double the lifetime for every 10 degrees C drop in temp" rule of thumb used for most electronic components.)
At 10 hours a day, 50k hours is almost 14 years. And that's just to drop to 70% output - all that means is that you turn the driver up to ~25% more current, and the output is the same - you've just lost a
little of the efficiency advantage over other forms of lighting. Plus, as mentioned above, we're likely to see longer than 50k hours before we drop to 70% performance, because most of our fixtures are overcooled. 15 - 20 years is likely longer than any of our tanks are going to last, so that's plenty good enough for me.
And, given the rate of development, at that point, there will be far more efficient LEDs at dirt cheap prices, so if you do want to replace, it should be cheaper and more efficient. It's a win-win situation.
It's typically the opposite. The HP LEDs we're using are designed with long lifetime in mind, even at high operating temps. It's a blatant feature in the marketing material used by the manufacturers in many cases, so clearly it's an important feature of these LEDs. Yes, they dissipate lots of heat for such a small package, but they have been carefully engineered to support proper thermal management (thermally conductive path from the junction to a large thermal pad on the bottom of the HP LED, which is soldered to a thermally conductive MCPCB which is bolted to a big heatsink.) Meanwhile, most "average" LEDs are designed with upfront economy in mind, and have little or NO thermal management capabilities designed into the product.
Granted, there's still the possibility that some LEDs or other components on a fixture will just outright fail over time, but that's a possibility with other types of lighting. Who here that's had fish tanks for 10+ years hasn't had an MH ballast fail here or there?
And, let's please not get hung up on cost. Even if LEDs cost more (long term and short term), I would still enthusiastically use them. They give a nearly infinite degree of control over intensity, color, position, angle, directionality, and spread of light. As a hobby community, we haven't even scratched the surface in terms of creatively using LEDs on fish tanks. People are already doing things with LEDs that you could
never do with MH, and we're just getting started. I'm a control freak, and LEDs are satisfying that desire quite well.
In the end though, you have a point - DIY LEDs are certainly not for everyone. Some people build kit cars that can hit 60mph in 3 seconds. Other people drive a Honda.
To each his own.
