One of the more interesting things I learned while working for Google was the intricate way in which the "grid" is managed by a separate but co-operating set of entities. Google disrupted that happy bunch by creating a wholly owned subsidiary that was a licensed power company. That gave it standing to buy from and sell energy to the grid and it completely short circuited a lot of crazy negotiations that were going on between Google and various regional power companies. Now instead of having the substation outside the data center owned by the local power company it could be owned by "Google Energy Inc." And Google Energy could buy energy from any vendor connected to that grid.
Most people are familiar with the 'last mile' problem where the connection to the Internet from your house has to go through the local monopoly utility. The same is true when buying power, and this move on Google's part completely side stepped it.
It's fantastic that Google plans to switch to renewables since they are one of the biggest players. Definitely an exaggeration to say that they are leading the charge. Apple was among the first to make renewables an issue and achieved 100% in data centers back in 2013, with little fanfare.
And creating a subsidiary that's a power company is table stakes:
Google is a much much larger energy consumer, has been carbon neutral since 2007, and is going to use 100% renewables globally within the next year.
Also, Google made its energy company play back in 2010. Apple made theirs in 2016.
From Google's 2016 energy report:
"And in 2017 Google will reach 100% renewable energy for our global operations..."
That seems implausible. The electrical requirements of Aluminum production are extreme. Obviously there's a lot of issues about where one draws the line between the company and its subcontractors and so on, but it seems hard to believe Apple's global energy footprint is not higher than Google's
When we measure our carbon footprint, we include hundreds of suppliers, millions of customers, and hundreds of millions of devices. And we’re always looking for ways to make the biggest difference in five major areas: manufacturing, product use, facilities, transportation, and recycling.
Then lets do a back of the envelope calculation for energy consumption for Apple's aluminum use.
Very roughly Apple sells ~25 million macs and ~200 million iphones per year. Again lets assume that it needs 2kg of aluminum for macbooks and 100g for phones (not all iphones are aluminum and this might be lower), that makes ~70 thousand metric tonnes of aluminum.
Modern smelters use 13KW/h to produce 1 kilogram of aluminum, but their production involves recycled aluminum (and possibly shavings from milling?) so number should be lower, lets say it is 10KWH per kilo.
This makes 70M*10KWh = 700GW/h year
Google's datacenter energy consumption, is difficult to guess, if there are 1 million servers, and assuming each server consumes 10KWH per day, then it makes ~3.5 TWH per year.
This is apples to oranges of course, aluminum vs server running energy costs, but still kinda implies that Google uses more energy, at least directly.
FWIW, the linked article mentions
> The 5.7 terawatt-hours of electricity Google consumed in 2015 “is equal to the output of two 500 megawatt coal plants,” said Jonathan Koomey
though they don't mention the source
Likewise, you couldn't just keep buying big batteries from third parties to power all your needs, and claim to have no carbon footprint because you didn't make the batteries.
USGS has online reports.
http://minerals.usgs.gov/minerals/pubs/commodity/recycle/rec... 40% ~2001
For 2015, it looks as of 30% of supply came via recycling:
But the energy used, carbon emitted, is fairly small for the 1000 or so flights they take relative to the aggregate benefit they're making by buying so much "cleaner" power.
You really ought consider the relative perspective, I don't think their 757 carbon impact compares to the output of a 5MW data center running 24/7.
"A Boeing 747 jet has an average power consumption of 140 megawatts."
You can use this link to see San Francisco's power usage in 2015: http://ecdms.energy.ca.gov/elecbycounty.aspx
It was almost exactly 140 megawatts/hour * 24 hours * 365 days.
I highly doubt Google's worldwide data centers use more power than all of SF.
So, not silly!
MW/h doesn't make any sense. A unit of power would be MW, a unit of energy would be MWh (megawatt times hour, not divided by hour). MW/h would be a power increase over time, which is surely not what the parent means.
 Unless they're in the US, Liberia or Myanmar. :)
Google is purchasing 2600GW of renewable energy as of Nov 2016, while Apple is purchasing around 150GW.
For example, Apple's global power consumption definitely includes power used in manufacturing, a problem that Google externalizes from this sort of computation.
Every other phone or tablet sold in the E.U. has a sensible power adaptor, usually the mini USB flavour. Apple though, they don't think the law applies to them, or at least not the spirit of the law, so they keep churning out these Apple only power adaptors complete with the cables that fray and break a lot easier than normal cables.
Sure the data centres may be green but the landfill is not. Meanwhile, Google (who have far larger and more useful data centres) have been providing E.U. compliant power adaptors (or lack of) since the first Nexus phones, if not a long time before.
Nonetheless, framing ones' opinion of Apple's sustainability practices on this one pet issue seems quite myopic. The actual sustainability complaint stems from the alleged poor durability, not the choice of connection; but even if we accept that criticism, replacing charging cables occasionally seems like small drop in the bucket vs. the company's environmental impact as a whole.
That's different than switching to a new energy source.
A Google datacenter near the Columbia River may derive 100% of it's energy from hydro. Google pioneered that practice and many others followed. But their overall real estate footprint relies heavily on local energy suppliers, which may be coal.
Any company making a claim of being 100% renewable, is basically just using offsets.
And offsets are a way the government allows companies to (metaphorically) continue driving down the road in gas-guzzling SUVs, but make investments in renewable energy to offset their footprint.
You argue that by being physically near the Columbia river, a hydro electric plant can be constructed which generates power and sends it directly to a Google data center. But their other properties, some of which are nearer power plants which generate power using fossil fuels can not be considered to be using renewable energy. Is that about right?
If so, I see it a bit differently, if Google builds a concentrated solar plant in the Mohave Desert and that plant puts adds 1200MWhrs of energy to the Western Grid every month, and Google's real estate in California consume 1000MWhrs of energy from the grid every month, I can see that being claimed as using 100% renewable energy. There isn't a specific path of the electrons to the Google buildings but in that scenario Google will have added more energy through renewable sources into the grid than they extracted from the grid
I've got Solar panels on my house and they generate more power during the day than I can use, they generate no power at night. I have them connected to the grid. When I'm generating excess power its probably being used by my neighbors but somewhere nearby a fossil fuel plant is using less fuel because of the power I am feeding into the grid.
Because is is set up this way, PG&E bills me on an annual basis and on that basis the 'net' amount of energy my house uses is quite small, in terms of ratios, over 90% of my total electrical energy usage is accounted for by the energy I generate on my rooftop. I think of that as "90%" of my energy comes from renewable sources, even though every night, at that moment, I'm using fossil fuel sources (or simply the grid). Once I have a couple of PowerWall's install it will get a bit more explicit but I consider both interpretations valid. I'm guessing from your argument that you do not.
 You can also argue that the 2nd law of thermodynamics tells us that adding it and then removing it will leave a bit behind so perhaps a draw of 1MWhr might need an input of 1.2MWr in order to offset grid losses to radiation and heat.
Frankly I don't think anyone should. It's mental gymnastics to be able to say you aren't using dirty energy when you really are.
Let me illustrate with an example of why you can't get by with 'offsets'.
Let's say Arizona is running on 80 GW natural gas and 20 GW solar. Google wants to build in Virginia, which relies heavily on coal.
So they bring online a 2 GW datacenter in Virginia which is powered by a 2 GW increase in coal burning. They "offset" it by building 2 GW of solar in Arizona, which scales back natural gas by 2 GW.
The result of their 'clean offset' has been the conversion of 2 GW natural gas to 2 GW of coal, which is significantly worse for the environment.
Would it help if the renewable source had to be connected to the same grid-tie (there are 4 I think in the US) ? Then you're putting renewable energy into the same grid you are pulling it out of.
What if your solar plant in Arizona use the Fischer–Tropsch to make fuel, then you drove it in a tanker across the country and poured it into generators in Virginia?
The way I see it is that watts are watts (just like dollars are dollars) so if the total watts you burn are offset by renewable watts you have put in, then you are ahead. If everyone does that then by definition the entire grid is running on renewable energy.
Right, but they aren't connected to the same grid. Google has datacenters all over the world, connected to all sorts of different grids with different source compositions.
>What if your solar plant in Arizona use the Fischer–Tropsch to make fuel, then you drove it in a tanker across the country and poured it into generators in Virginia?
Sure, but you have to account for the emissions from that whole process as well (emissions of solar + emissions of generating the fuel + emissions of transportation + emissions of burning the fuel).
>The way I see it is that watts are watts (just like dollars are dollars)
Nope, a watt generated on one grid does not spend on another because they're not all interconnected.
Until we solve the energy storage problem, the problem is exacerbated further by continually expanding solar installations in one particular region of the US that will generate more than the consumption for part of the day.
I guess nuclear power is non-renewable but carbon neutral. Cannot think of anything that is not carbon neutral but renewable...wouldn't that require creating CO2 matter from nothing?
1. Net usage after selling off excess energy
2. Net usage assuming there is no demand for excess when you have it (maybe everyone on the grid has solar panels)
Achieving 100% in #1 is great and I applaud Google for it. If everyone did this we could drastically reduce overall usage of fossil fuels, but I believe the fear is that people won't know the distinction and will assume "if everyone does what Google does we will all be 100% renewable" and that isn't true.
That is the definition of 100% renewable energy use is it not? You can't push power into a grid if nobody is using it, they don't work that way. So if the world is using 12.3 TW and 12.3TW is being generated by renewable sources, your done.
What about the consumers of my excess energy? Are they operating at 20% renewable and 80% fossil?
If my 20% in excess is counted towards my total renewable energy consumption, and other consumers also claim that 20% towards their renewable consumption, then we're just double counting. No?
I think the lesson learned from that scandal is to keep it simple. On a per grid basis, simply acknowledge that fossil fuel accounts for X% of power, and rally the community behind an honest metric.
Headlines claiming 100% "mission accomplished" will lull most people into a false perception.
Don't get me wrong, it is a good thing they are doing at least this much... but in my opinion they are getting a huge, undeserved PR boon out of this. Companies like Oracle could not get away with this, even if they tried. In my eyes, and until hard evidence is presented to disproof this, I think they are still milking all the goodwill from the "don't be evil" motto from yesteryear.
(I still think their goal and actions are admirable. There are many companies that have been jump started by a major customer pre-buying a product so the company can ramp up manufacturing to fulfill that and other orders. I much prefer this model over government-driven investment as I have more confidence this demonstrates true demand.)
Also that chart is US-only, while the story about Google here is that Google will go 100% renewable energy globally.
Capturing and burning the methane "downgrades" the potency of the emissions, now only emitting CO2.
Also, if done sustainably, you have to keep sustaining life forms that keep doing the photosyntesis trick for you every year. i.e. All those carbon atoms in this year's crop of Christmas Trees were in the atmosphere a couple of years ago, and we are potentially breathing out at least part of the crops from the next few years... But if someone goes and cuts down a big chunk of the Amazonas in order to make "bio-fuels", yeah, that's dishonest, stupid and self defeating.
I love how the author throws in a disclaimed quotation from a big oil Luddite to help reinforce the message of the article. It's a two-fold sales tactic.
1. Discredit the opposition.
2. Tell the buyer what the opposition will say, after already convincing them of the benefits, so they're prepared to defend what they're buying thereby reinforcing their beliefs.
I'm not implying that the article is wrong, but it's something we should be aware of.
The problem with wind and solar is that they are intermittent. Therefore, Google's consumption at any particular moment in time cannot be supplied solely by these intermittent sources because their energy output never matches current demand (at least until some big honking batteries are built to level this out, for which the tech currently does not exist AFAIK -- no, Tesla's battery tech isn't even close to operating at this scale, though its a small step in that direction).
Therefore, Google must take power from the grid from non-intermittent sources that can ramp up and down based on demand, such as coal, gas, and nuclear. And this fact will remain true, even if Google had purchased wind and solar contracts for total energy many times greater than their total energy use.
So even though they may be purchasing contracts for wind/solar energy that exceed their own total energy use, they still cannot even approach viable operation without access to non-wind and solar sources, so its definitely misleading by Google to suggest they are running entirely on wind/solar.
Furthermore, the excess energy produced that Google purchased, but did not use because of the intermittency issue, actually reduces the efficiency of fossil and nuclear sources. Because the wind/solar energy is loaded onto the grid intermittently, it forces fossil and nuclear to cycle up and down to compensate. Just like city vs. highway driving, this is vastly less efficient.
It's true that there's a storage problem to be solved. That doesn't mean that every single renewables project needs to solve it or become a gimmick.
Google is still using dirty energy, they are just offsetting it by returning extra to the grid when the wind is blowing and the sun is shining.
Nobody forces you to buy a cow to put milk on your cofee, but if you go around offering "organic capuccinos" for sale, your customers could normally expect you to either milk the grass fed cow yourself or, more likely, purchase your milk from a certified provider that does so. They certainly don't expect you to claim that since you make a donation to some Amish charity in the other side of the country, you are entitled to slap a "organic" label to the same brand of industrial milk your competition next door uses and offers at half the price.
I don't see how this should be any different with Google.
Let's say for the sake of the argument that Google's datacenters consume 1% of the electricity in the US.
If Google were operating its datacenters with clean/renewable electricity produced on site, it means the technical problems are, for all practical purposes, already solved; and now everybody can just license their technology an imitate them. There still remains the economic and political problem of having the other 9x% of the country to jump on board, and invest the required resources to convert to renewables, but that's simply a matter of how much money is devoted to that. It could take us 100, or 50, or 20 years, but we'd know for certain that we will eventually get there.
Instead, we have Google claim to be operating their datacenters with clean energy, but instead they operate with whatever mix of clean&dirty energy is available from the local provider, and then go elsewhere and finance the production (let's say, same amount of kilowatts/hr) of renewable energy. That is still a good thing, since those investments would never had been done without the incentive, but there are 2 problems with that.
1. You call electricity "fungible", but it is not. You cannot generate the same amount of solar kWh per dollar in Alaska than in Arizona, the same way you cannot generate the same amount of hidro kWh/$ in Arizona than in Mississipi. By following economic rules, it is rational to assume that Google is picking the low hanging fruit first. Therefore, when the other companies try to emulate, they will find that the second 1% will be more expensive than the first, the third than the second, and so on...
2. The other big whale in the room is that renewable electricity, unlike coal-based, cannot be produced on demand. Until the problem with storage is solved, - and for what I have heard from people that has been doing this for many years it is far from solved, - you will need some form of fosil fuel to close the gaps when production is low. It is very easy to add 1% renewables to the general "fungible" pool of electricity in the grid, but 5% would be less so and I don't think it would be possible at all to do 50% with current technology, even if the production capacity was already in place.
So, what do you think? are my concerns really nonsense, or ridiculous?
They simply pretended that they were "cleaner" than coal and gas, which isn't really true:
is there some kind of auditing authority that can actually guarantee that the money spent is actually going towards electrical generation and isn't just slush-fund cash that financial holding companies play around with while nominally saying "green energy"?
They give some details on this here:
When you purchase these renewable energy credits you're essentially bidding up a bonus for producers of the appropriate energy, above and beyond the price they got for the electricity itself. Google has entered some long-term contracts for the acquisition of these credits from renewable energy providers, which in the short term bids up the price of credits for everyone else and in the longer term leads to more and larger renewable energy producers.
Also note that Google isn't doing this entirely via the credits markets — it's also reducing its usage and where applicable is working with energy providers to create nearby generation.
Can someone explain this? It seems like wind energy prices should fluctuate based on weather patterns (more wind == lower prices).
Over the long haul, yes, but in real-time - no. And electricity prices to fluctuate a lot over days, weeks etc..
But indicating that 'price stability' is a point for green is not really fair, when it's almost always going to be considerably more expensive.
Knowing that something is going to be 'consistent, and consistently more expensive', I don't think is a good argument for green.
There are other, better arguments.
You'd deal with this the same way you deal with it in every other commodity market, futures contracts.
Google is in effect doing the same thing as 'buying a future' and it's unreasonable for him to make the argument that this is 'inherently a better thing about green'.
I could buy a futures contract for Oil today and would it be fair for me to say 'Oil is has more stability in prices'?
But Oil shouldn't fluctuate too-too far between $50 and $120 a barrel. And that's just the raw costs. With all other components, which costs will be more fixed, the actual end price of fuel should not change more than 40% in either direction, max.
I'd argue that the greatly varying costs of wind installations, aggregate availability, huge one: supply and demand for electricity overall, specifically for renewable electricity - regulator changes, improvements in technology - definitely create a situation wherein the 'end cost' of renewable electricity is somewhere in the range of the variability for fuel produced by other means.
Uh, sure there are, they are wagering on weather, which is a very common thing to do in various commodities markets.
I have never heard before a theory that greater market participants necessarily increase volatility. Can you cite a source for this please?
A minor detail, but don't forget the cost of transmission. My understanding is that the US has massive wind energy potential (http://www.nrel.gov/gis/images/30m_US_Wind.jpg and http://www.nrel.gov/gis/images/80m_wind/USwind300dpe4-11.jpg) but it's not necessarily near the most densly populated regions. Building and maintaining these transmission corridors is not an insubstantial cost.
Typically, these types of contracts, providing price certainty, are required to build develop new renewable assets.
Similar pegged pricing can be achieved if you trade futures to hedge your price but typically this is only achievable year to year and difficult to do with electricity as it can come from so many different sources and hedging against them all may be difficult. Perhaps there's a megawatt future out there?
EDIT: Yup - regionally-based electricity price hedge. Cool!
Everything fluctuates, but some things are less prone to wild fluctuation, by quite a large margin.
So a war in the middle east or a trade embargo won't impact the price per kwhr. Whereas a tornado would impact both wind and oil (even if just by slowing down the deliveries).
I would argue that wind (and solar (not so much hydro)) are much more susceptible to weather impact, but with sufficiently large capacitors/batteries that can be minimized.
So, whoever has the cheapest power will be able to provide cheapest services, or charge more for capacity located in more energy-expensive regions (for latency).
"For starters, as of January 2016, we’re sourcing or generating enough renewable energy to cover 93 percent of the electricity we use at our facilities worldwide. In fact, Apple is now 100 percent renewable in 23 countries, including China, Germany, Singapore, and the United States. We’re also 100 percent renewable at every one of our data centers. So whenever you send an iMessage, download a song from iTunes, or ask Siri a question, the energy Apple uses doesn’t contribute to climate change.
In the past five years, we have reduced the carbon footprint of Apple facilities by 64 percent thanks to our clean energy use, avoiding over 1 million metric tons of carbon emissions. We’re working hard to reach 100 percent renewable energy for all of our facilities worldwide, and help our suppliers in China and everywhere around the world make the same transition to clean energy as we have."
Yeah, they rank on the list, but they're pretty far behind.
I don’t think the thing to judge by is who consumes the most renewable energy.
One of the reasons the UAE pitches itself as a data centre site.
Renewables are effectively second-class citizens in the energy world until they can match the reliable storage capacity available from dead dinosaur.
While it's a good point, I don't entirely agree with it.
More apt: renewables are not cheap. They are very expensive in almost all parts of the Western world.
I'd expect much more than that given the size of their operations. i.e. just running the entire world.
Considering that they employ ~70k people, plus the hardware to run their business, that sounds about right.
It's not as simple as saying you support companies doing this voluntarily. I also support that, very much so. But the governments have put up legal hurdles because in the past they "protected capitalism" and now just stand in the way.
State courts are used by telcos to sue cities to prevent Google Fiber from coming into the city, to protect the profits of the "private" interstate corporation from the "public" city.
Drug discovery and many other things are hampered by patents to protect the "intellectual property" of pharma corporations from the "freerider" researchers and public funding that would discover new cures.
And so forth and so on. The legal departments and regulations of the old capitalist system (set in place to protect innovation at a different time) is in the way of open source and open innovation of the 21st century.
And the biggest lie of course is jobs, that demand for human labor will never go down and that enough money will always trickle down to the plebes via wages.
I would expect their usage to be higher if they were trying to store renewable power to maintain a base load, but that doesn't seem like enough to outright dispute the article/claim like this.
This is more than a nitpicking point. The reason why renewable energy isn't a totally viable energy source is because of problems with reliability, predictability, geographic concentration of production, etc. It is highly misleading to give people the impression that an operation like Google can actually operate directly off exclusively renewable energy.
What Google is doing will eventually hit a cap and become zero sum. You still need fossil fuels (or nuclear) to support a reliable and dispersed grid, and can therefore only have a a certain percent of the grid powered by renewables. So if you want your company to "run entirely on renewable energy", you'll end up in a bidding war with other companies that want the same thing, to be the technical purchaser (not consumer!) of the renewable energy going into the grid.
That's true in general, but not if you can choose where you are using the electricity. If your datacentres are in Iceland (geothermal), the Middle East (reliable solar), certain parts of Canada, or Norway (hydroelectric), you could actually run the systems entirely off renewables. I agree though that buying 1 unit of renewable power for each unit that you use is a much less meaningful action. Not to be sniffed at, but not what is being implied.
Let's be honest...it is not good for their shareholders. They are buying power at a significant premium and costing their shareholders money. It is a direct hit to the bottom line. This might be good for the environment - at the very least they are helping renewable energy companies get bank financing through purchase guarantees, which will advance the technologies being used and perhaps one day make them economically viable. But to say it's good for shareholders is disingenuous.
>"In some places, like Chile, Google said, renewables have at times become cheaper than fossil fuels."
I love all of the hedging that comes with any article that tries to paint renewables as economical. This sentence has holes a truck could drive through. It could literally mean that at one point, years ago, Chile had extremely high gas prices for a week and Google was amazed that their bill for that week was lower than it would have been with conventional power sources.
Investing in solar on your roof, in most markets, in a better investment than the S&P500 (EDIT: Utility solar has costs approaching 2 cents/kwh and still declining). Google is doing their shareholders a service by hedging their energy costs, no different than when Delta (the airline) bought a refinery.
Ultimately, Google is doing a service for the world if in fact these companies use Google's support as intended. There are probably more effective, direct routes for a company of Google's size though - they could easily just roll out their own renewable energy research division, and then shareholders would own the resulting technologies. The approach that they are using doesn't seem logical to me, and then trying to spin it as if they are doing shareholders a favor by overpaying for power seems like they are making a bad mistake even worse.
Of course not. That'd be a pain in the ass. That's why they sign PPAs (power purchase agreements) that entitle them to the power from those generators (usually on 20-25 year contract terms). Think reserved instances at AWS. You're paying to commit to a resource usage. What does AWS do with that money? Buy hardware and provide the service to you. You simply make a payment.
> They aren't investing in anything, and their shareholders aren't getting anything in return.
I disagree with you, as do the facts.
Show me what they can put on a balance sheet in the end.
I assure you, these PPAs are not overpriced.
Of course in the short term it might impact the bottom line but so long as the business trajectory is positive and Google's stock continues to be perceived as a good investment then that shouldn't matter.
And your issue with the line: "In some places, like Chile, Google said, renewables have at times become cheaper than fossil fuels."
Meh, it's PR, take it with a grain of salt and remember this isn't a peer reviewed paper.
Generally all you have to do is account for all the costs. Just like dumping pollution into a river can seem economical from the point of view of the dumper, but not from the point of view of the people who live in the area.
The bit you quote is presumably going even further, and saying that even without taking into account the wider costs, it still makes economic sense in that narrow sense.
perhaps it's not good for short term profit and volume, but don't equate that with shareholders please.
Google is already in that business.
Monetizing it in the energy sector will be very, very good for shareholders.
Monetary transfers or asset appreciation aren't the only ways to provide value.
It's like me wanting to put solar panels on my roof: I'm not so much interested in cost savings, but want to ensure a massive grid failure doesn't take my household with it.
The main difference between the OATI Microgrid Technology Center and other commercial office buildings is that OATI will generate its own electricity as the primary source of power. The building is powered by:
- Natural Gas turbines: 600kW Capstone C600 natural gas burning microturbine. Paired with absorption chiller and heat exchanger for CCHP
- Solar: 150kw rooftop solar array, with additional expansion array planned
- Wind: 24kW of vertical axis wind turbines
- Electricity storage: 231kWh, at 125kW Ensync battery rated power and energy
- Generator: 1500kW of diesel backup generator
- Utility connection: Connected to local utility Xcel Energy
Very forward looking of them!
But at least they're going to create a greater demand to hopefully create a greater supply sometime soon, plus they're paying into the renewable energy economic machine.
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.
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.
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. 
Modern panels in decent locations are under 4 years.
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.
The prices are 1 or 2 cents higher (per kilowatt) in many areas.
Definitely going to make the switch to the
> 100% renewable energy – for a 2 cent per kilowatt hour premium over current PG&E rates
1095 € for the local "default contract" 
1060 € for a one-year contract from the local default provider
1080 € for their "eco option"
792 € for the cheapest offer overall 
855 € for the cheapest 100 % renewable
1040 € for a reputable 100 % renewable provider that has the certification supported by most environmental organisations
 (which is tricksy, since it is a package deal with massive overage fees, not mostly a flat fee per kWh like the others)
So there is quite a potential for savings by picking the right provider, and high-quality eco options are among the more expensive ones, but if you do research you probably can find a compromise you like. E.g. the expensive eco providers often earmark 3 % or so of the price for long-term investments in building their own projects, instead of only reselling electricity provided by others, which of course adds on top and some people want to support. Some accept biogas as a power source, others don't.
And since many people never cared and pay the default deals, when they think about switching for eco reasons the shock isn't as big, because they often can save and get something 100 % renewable. Many people are willing to pay 20-50 € extra and go for that if available. The expensive options obviously have fewer customers, but I know people from all categories.
When I read 'Renewable' on the title, I thought they would use hydroelectric energy too. Nice to know I was wrong.
The issue is not submarines, it's computers, lights, cars for the rest of us.
Google has a side-show project I think related to nuclear, but it's an extremely regulated thing so it's hard.
Once the developer has signed PPAs they can finance the up-front construction cost. PPAs enable the construction and operation of new renewable energy projects. Of course the output of large scale renewable facilities is mixed with other sources of electricity in a transmission system; it doesn't directly replace coal. (But this is true of any electricity project in a competitive market: when a company builds a new natural gas plant, they don't detach the transmission wires from a competitor's older coal plant and attach them to the new gas turbine.)
Instead it's pricing effects that indirectly decrease use of dirtier fossils when a grid adds renewable capacity. Solar and wind plants have very low variable costs; they'll supply every MWh they can to the grid, and get paid for every one of them via the previously signed PPAs. But fossil generators have significant variable costs for fuel and for labor, so when renewables are producing high output there may not be enough demand to cover the marginal costs of production. Fossil plants reduce or stop production when that happens, so they burn less fuel and produce less emissions. Fossil plants with high enough marginal costs may shut down for extended periods, like some plants that are only used for e.g. meeting peak demand during the winter heating season. Costs rise over time as aging equipment needs additional maintenance or full replacement. Fossil plants with high enough costs eventually can't turn a profit even serving the peak-demand-season niche and are completely retired and scrapped. Even if an old coal plant is replaced by a new coal plant, it's still an environmental win to get the old ones shut down, because (even under the new president) a coal plant that starts construction in 2017 won't get the weaker grandfathered pollution standards of one constructed in 1957, and the newer plant will produce more electricity per ton of fuel due to technical advances.
 At least coal plants have significant labor costs. Modern CCGT natural gas plants are quite labor-thrifty too.
Effectively renewables have an externality cost of their own: they make "reliable" sources of power much more expensive.
Imagine that wind/solar power supplies 25% of the grid's needs - this means 75% of the time you need a gas power station to fill in the lulls in production. Fine.
Then imagine you triple the amount of renewable energy on the grid to 75%. The gas plant is now only operational 25% of the time, but still requires capital cost of construction and maintenance to be ready to go. Ok, it doesn't need as much fuel but apart from that you have tripled the cost of the plant.
So the problem is that it becomes uneconomic to build the electricity generation that stabilises the grid and makes it possible to have any renewables in the first place.
We're now seeing huge premiums being paid to companies by the various electricity grid operators to cover this problem. This then goes on ratepayers bills to cover "distribution" or "grid reliability". And it then turns out your $0.03/kWh wind is not actually that at all and gets more distorted as more wind gets generated.
We need better energy storage options before renewables really will fulfill their promise.
No politics on HN this week I thought?
I think this is lost on most people in Silicon Valley.
And there are so many pollution externalities: increased risk of disasters like spills, and people who get respiratory diseases/cancers from breathing polluted air and die early.
If they want to see their 10xs breakthroughs (which they sure talk a lot about) they should be throwing money at condensed matter physics research and just support in general for physics. Its easy to get a 10x in software but energy is a whole different game.
That way the smartest people won't have to decide between taking a high paying engineering/programming job or going into physics.
It all comes down to probabilities and the amount of people in the field is really the best approach towards making a breakthrough.
Engineering is much different than software architecture.
One requires hard sciences competency, whereas one you can go into for years with none, and be successful.
Without the support and money from companies who care, many will choose the highest paying option