Moving manufacturing and infrastructure from "dirty" to "green" is a transition process. It's infeasible to modify the entire supply chain to be non-renewable in one shot.
The entire world's economy is incredibly intertwined. The entire manufacturing stack won't be purely renewable until the entire world is running on renewable energy.
In today's world, you have to make an effort to not fart with a bad carbon footprint. Most of your personal emissions will be ingested air. (Not methane, as most think.) However, the fertilizers and pesticides in the food you ate to make you fart, plus the heat to cook the food, and the industrial input it took to make the cooking implements and the kitchen the food was cooked in -- all of these have a carbon footprint.
It's precisely as you say. We are in transition. It's naive to point out that anything "green" still has a carbon footprint. I need to find a definitive article that makes the case for this idea so I can link to it every time I see the naive argument.
http://www.aljazeera.com/news/africa/2014/08/kenya-lead-pois...
Data centers are an easier target because you have more flexibility. And to provide a counterexample, Apple recently pressured Foxconn to power certain parts of iPhone manufacturing with renewables by 2018. We'll see if the two companies deliver on that goal.
https://www.apple.com/pr/library/2015/10/22Apple-Launches-Ne...
And on the flip-side, don't let good stop you from reaching perfect. This is a good step and one that should be applauded, but at the same time recognising that there is still work to be done is valuable.
The operational phase is where to look at having the largest impact. In buildings looking at a 75 year time frame you have the operational phase accounting for 94.4% of life cycle primary energy consumption. Where the manufacturing phase such as production of building materials, transportation and constructing the building accounts for 2.2% of life cycle primary energy consumption. [1]
1. http://www.sciencedirect.com/science/article/pii/S0378778803...
Modern panels in decent locations are under 4 years.
See e.g. https://www.ise.fraunhofer.de/de/downloads/pdf-files/aktuell... http://www.nrel.gov/docs/fy04osti/35489.pdf http://www.volker-quaschning.de/datserv/kev/index.php
Do we have more data now? Because I think solar panels will easily do 50 years with >50% of original output.
I believe that there are already some panels in the field that can last 50 years, and that it will become more common in the future, but 20-30 years is the safe, evidence-backed estimate for now.