In 20 years the only money to be made in energy will be made by grid operators.

In 20 years the only money to be made in energy will be made by grid operators.

[This article hasn’t been edited, but with certain recent news it seemed worth posting even it its rather choppy, 1st draft state. Especially as I’m not likely to get around to finishing it until after I move!]

By 2035, 2045 at the latest, only moving electricity around will be profitable; utility scale producers will be racing each other into bankruptcy. Even many solar and wind producers.

Given that this is exactly the opposite of the current state of energy around the world, I expect some resistance to the idea. Just a touch. Let us explore…

It’s not just solar… but yeah. Solar.

RCA 630-TS Television Here’s the thing. As little as ten years ago, residential solar was basically at the same stage as televisions in the 50’s, only not as cool. Massive, hideous boxes that took four people to move and any sane person would hide behind curtains. Ugly as they were though, those lead boxes did amazing things, so everybody wanted one. But, maybe not so ugly, please? And perhaps not $10,000?

Fifty years later, around the turn of the millennium, the average American home hosts something like three televisions. Screen sizes are starting to creep upward: the 19″ television of the early 90s gives way to the 30″, which starts, ever so slowly, to give way to Plasma and LCD screens above 40 inches.

Now blink. In the past ten years, nearly every behemoth television sold before 2005 has been replaced. Not moved into a bedroom. Replaced. Three televisions per household, replaced with a vastly superior specimen for a quarter the cost of the originals. Many of those televisions were replaced not once, but twice in the last ten years, because we’ve had two distinct generations of television technology in the last decade (three if you count the newest generation of 3D televisions, but consumers really haven’t bitten). From CRT we went to affordable LCD and Plasma, which did have better image quality, but mostly just bigger screens and less floor footprint for a lower cost. Then came the High definition revolution, which brought movie-theater quality into the living room.

That’s not even close to all. In 2010, Apple released the iPad. The word ‘television’ started being replaced by ‘media device’: and ‘media devices’ started showing up absolutely everywhere. Airports went crazy, going far beyond replacing the walls of CRTs. Retail shops now hang screens in windows, displacing mannequins and posters. Mass transit stops display schedules, stadiums and auditoriums have replaced posters in hallways – literally replaced, putting static images on screens and only updating the image between shows, or every few minutes. Restaurants are even replacing chalk boards with LCDs. Chalk boards!

So. Residential solar. Ten years ago residential solar started breaking out of its rather limited circle of interest and into the larger world. Costs were still high, but in many places were still easily justified. Installations were ugly as sin, but mounting systems started getting some attention – like the 50s television, wrapped in expensive wooden boxes that we can only assume were considered attractive at the time. They buying process started becoming less of an ordeal, and the risk (perceived or real) of ending up with a defective product and bankrupt installer began rapidly decreasing.

Five years ago residential solar was more like television in the year 2000. Specifically, the 40″ and up plasma and LCD televisions. Sure, it’s expensive, but the benefits are pretty obvious, at least for those in sunnier climates. Adoption rates really started climbing. Research results from unrelated fields augments direct investments in solar manufacturing and helps drive manufacturing costs down. Over supply of the raw materials needed for panels starts pushing panel prices down as well. Analogous to the early slimming of the TV, solar installations started trimming the fat: huge posts in the side yards and big, industrial-warehouse-looking roof top mounting systems giving way to slim installations that lie flat against the roof, or a few inches above. Solar, like the big screen television, starts to get interesting to a lot more people.

Right now, approaching the end of 2014, solar is about to become… boring and mundane. Just like big screen televisions did right around 2010. And just like big screen televisions in 2010, we’re going to start seeing solar absolutely everywhere very soon.

Although the first round of fat-trimming in solar is still not pervasive, the second round has begun. This time it’s not just putting the same panel on a prettier mount, or hiding it against the roof more skillfully. Advances in micro power converters and other electronics required for solar are enabling changes in form factor and placement. We’re starting to see the first real integrated solar installations: solar panels conforming to curves, samples of solar “panels” that don’t go on the roof, they are the roof, solar panels that are “rolled out” onto flat roofs, like tar-paper.

Residential solar – which I’m actually using as shorthand for ‘non-utility scale solar’, as you have probably figured out by now – hasn’t had its ‘iPad moment’ yet. Nor do I even have a firm guess as to what it will be: battery storage? Solar shingles? Oil prices going back to setting new record highs every few months? There’s so much research going on at the moment that I’m quite confident we’ll see something extra amazing hit the market soon, and that amazing thing will be what we remember most when we think back on how solar started becoming pervasive five or ten years from now. The iPad, after all, was taking advantage of the existing state of technology; even without it, we’d still have 42″ televisions being used to replace chalk boards.


Cars. Electric ones. Them too.

Electric cars are the other half of this story. Without them, or more accurately, their batteries, I’d have to wave my hands a lot and talk about grid-scale power storage being an unknown, how scaling compressed air storage will be hard, because caves, and how expensive batteries are when you want to buy them by the metric arseload. Etc. But I don’t have to do that. Thank you, Mr. Musk.

The Tesla Model S has a range of around 425km (265 miles). It manages this with an 85kWh battery pack. But how much is 85kWh, exactly? A hair dryer on high for 3.5 days (85 hours) isn’t that enlightening, though it is accurate given a typical hair dryer.

How about some perspective.

In 2012, the average annual electricity consumption for a U.S. residential utility customer was 10,837 kWh, an average of 903 kilowatthours (kWh) per month. Louisiana had the highest annual consumption at 15,046 kWh and Maine the lowest at 6,367 kWh.

Ok, kind of dry. Sorry. But anyway: dividing by 903kWh by 30 days in a month tells us that, in 2012 at least, the average US home used just over 30kWh per day (we’ll ignore the fact that certain energy hogs are skewing this number massively upward. WTH Louisiana?). Which means that with just a tiny bit of conservation, an average home could be powered for three days on the energy required to drive an average sized sedan 425 or so kilometers. Note that this is going to be true for pretty much any average sized sedan, as the Model S is an amazingly efficient car, with very low wind resistance (coefficient of drag) compared to, well, pretty much every other car on the road. Smaller cars with a similar range might need slightly less battery capacity, but the reality is that most cars are going to need more, often significantly more.

Think about it: the average electric car that anyone will want to own can power a house for three full days.

More importantly, an electric car with an 85kWh battery and 425km of range could power a house for just two of those days and still have a 30% reserve – about 140km (90 miles). The average work commute is less than 50km.

With that, let’s talk about how we’ll go from here to killing utility scale power profits dead, in five year chunks.


Five Years: 2020


Lawsuits against utilities in places like Hawaii and Arizona will be found in favor of the citizens who want to put power back into the grid. Even backwards, contrarian utilities will be forced to put real effort into upgrading their ability to cope with distributed power generation. Mostly these public utilities will choose ‘net-metering’, allowing customers to roll their bill back to zero, but never paying them.

Residential grid storage starts becoming noticeable in forward-looking regions such as Germany and Norway – but only to engineers who look very closely at graphs made with very fine lines.

Pilot projects for “smart” and “micro” grids will scale to full neighborhoods and possibly even small towns in the United States. Germany, of course, is five years ahead of everyone. All sane utilities will have begun significant projects related to upgrading delivery systems to handle customers selling power back to the grid. Sadly, however, having been created by incumbent utilities these pilot projects will maintain substantially the billing structure we see today: the operating utility contracts to buy power in advance and passes the cost on to the customer, plus some profit. Power generated by members of the grid will (still) be considered inferior to that generated by registered IPPs and utility scale producers, and this view will continue be used to justify a resistance to charging and paying individual consumer/producers via instantaneous pricing strategies that truly reflect the current state of supply and demand.

Real demonstrations of electric cars storing power when it’s cheap and selling it when it isn’t will have made the news, with a handful of interested but otherwise ‘regular people’ (not associated with a company developing or marketing a related product) doing so as a matter of routine.

At least one SolarCity-like company will be found dipping a toe into the transmission space, leveraging a market area in which they have a particularly dense installation base. A neighborhood adjacent to a business park or fairly dense retail is most likely. Unlike traditional local utilities, this company will encourage all members of this experimental grid to store energy when it’s cheap and sell it back when it’s expensive. Also unlike a traditional utility, the rate paid and charged to members will be based on market price at all times, with one minute or better price resolution (most likely ten seconds or less). Success will depend greatly on the number of customers who have some level of grid storage installed. The company will operate the grid with two distinct business models simultaneously: first and primarily as a transparent broker of energy, charging a small, fixed markup on top of the cost of energy consumed from the grid and taking an equal cut from energy put onto the grid. The grid itself will operate as a market, with individual homes automatically bidding on and placing offers for energy, based on general parameters set by home owners and businesses. Secondly, the company will bundle the entire micro-grid into a single unit and present it to the outside world in the form of an Independent Power Producer. The company will then perform real-time energy trading on the larger regional grid, much like one of the members within the micro-grid. This business model will be an unqualified success for the earliest adopters, and it will eventually utterly destroy the profitability of centralized power production.


O&M costs feel more like S&M

The first coal plant that isn’t considered “old” will have been shut down somewhere, thanks to cheap natural gas, solar and wind.

Executives of utility scale power producers will start to drink heavily as June approaches, and won’t stop until September. In 2020, by late June any power producers with fuel costs will find it impossible to operate at a profit during daylight hours on most days. Given that there are around 15 hours of daylight in late June, and those hours cover all the times in which people are actually awake and using power, this is problematic for power producers who don’t own a lot of solar production capacity.

The practice of taking solar and wind offline in order to keep coal base-load plants from having to idle will already be in court. With natural gas able to fill the base-load gap and desperate for a way to actually be profitable even during sunny June days, coal will be on trial for its life. Nobody will be expecting coal to win.

If a coal plant hasn’t seen flame by 2020, it won’t. No new coal burning power plant will turn on after 2020. Anywhere. Any company that owns a coal power plant today, or interest therein, will have already either sold it or tried to by 2020. The rush to sell off interest in coal power plants and related assets will crash the value of any company whose value is significantly rooted in coal based power production, regardless of the age of their plants.

Solar chalkboards

The solar industry will crush the Energy Departments Sunshot Initiative of 6 cents per kWh, hitting 3 to 4 cents per kWh, down from 11.5 cents today. The vast majority of which will be permitting costs.

A 30% annual growth rate (down from the 32% – 75% annual growth for each of the past five years)  will find China alone installing 200GW of solar in 2020, with the US trailing by half at 100GW. This does not include wind power installs, which will be another large number. For a reference, consider that in the year 1950 the United States had a total generation capacity of about 175GW.

Nearly every single new home sold will have sufficient solar built in to power it completely during much of the summer. Only the cheapest of the cheap new homes will not. This one is a pretty safe call, actually: if a home without solar panels on it takes even a week longer to sell there will be pressure on the builder from the sales team, as well as the associated increase in holding costs. Now, the housing market might stay crazy strong for a while, which could push this prediction out a bit. Maybe. Still, even with a strong market I’d take even odds that 95% of all new homes sold in the United States in 2020 will have enough solar installed that they don’t buy power from the grid most of June and July. Simply because installation will be even cheaper than it is today, and adding solar is one of the rare breed of home upgrades that all parties in the home sale agree add value – seller, buyer and most importantly, mortgage broker.

But mostly it’s a safe call because companies like SolarCity will pay home builders to do it.

Residential system sizes will increase somewhat as panel and soft costs drop, but price reductions will mostly serve to simply balance the expiration of tax credits.

Speaking of SolarCity, that fine little company will be installing more than 10GW per year between their business and residential customers. That’s based on 60% year over year growth. If they maintain the 104% year over year growth they achieved between 2012 and today it would be closer to 35GW. Similar companies will combine to more than match that, as SolarCity will almost certainly still have significantly less than 50% of total market share (it’s at about 35% today). Total non-utility solar generation will grow by at least 30GW in the US alone in 2020, driven in no small part by wildly fluctuating oil prices.

All new commercial buildings with flat roofs will either have a green roof or solar included as part of the design. For one to five story construction, including warehouses, warehouse stores, grocery stores, business parks, etc, many new buildings will have solar siding or other integration solar products, such as windows, solar carparks, awnings, sunshades, etc. High-rise construction will leverage solar awnings and solar windows, with the rare instance of solar panels integrated into the building shell in some fashion.

Solar shingles – replacements for clay tiles or asphalt – will have arrived For Realtm, but they won’t be price competitive with clay tiles on a one-to-one basis (that is to say, without considering power generation). Solar siding products will hit the market at scale, mostly for use in commercial buildings. The majority of the solar siding products will be facade only – not actually providing weather protection (other than sun protection, obviously). Some products will even be attractive; it doesn’t take much to beat the exposed concrete outer wall of the typical home improvement warehouse store, or the standard beige stucco found on most retail warehouses, business parks and mini-malls.

Retailers will once again be found chasing after Walmart and Ikea. Instead of logistics, this time it will be in utilizing the currently idle resource that is the rooftop space of massive retail complexes. As prices fall, the trend will move into parking lots, with more and more retailers and (particularly) leasing companies becoming Independent Power Producers on the side. Install sizes will start jumping around this time, transitioning from producing only some of the buildings energy requirements to fully covering them on long sunny days, and then beyond. For businesses, the initial driver will be cost predictability rather than actual savings.

Pulling into a large parking lot and not see at least one or two rows of covered spaces will feel odd. Typically if there is only one row of spaces covered it will be those spaces with charging outlets for electric cars, which will be rapidly gaining market share.

The explosion in commercial roof installation will be driven in part by roll-to-roll organic cells finally hitting the market in a real way, which will slash installation costs dramatically. Organic panel lifetime will still be measured in years rather than decades, and efficiency will still be half that of the competition. However, the lower system cost and resilience in the face of “single module shading” will more than make up for these shortcomings. More expensive, purpose built infrastructure such as carports will receive correspondingly higher efficiency panels, while roofs and walls will be covered in lower efficiency, more affordable, easy to replace, shade tolerant organic based technologies.


Home, home powered by range. But not yet.

In five years Tesla will have released both the Model X and the Model 3. Nissan will have released a completely restyled Leaf with 400km range, and at least two other major car companies will also finally be offering fully electric vehicles with at least 80kWh battery packs. Middle income households will have at least three desirable electric vehicles, more likely four or five. Sadly, only one or two of those (other than Tesla) will have designed their vehicles specifically for electric, which will mean generally “mushy” handling due to the extra weight of batteries and likely poor weight distributions. Also, no frunk.

Attentive pundits will declare the opening of a new war between auto manufacturers: a range war, to replace the horsepower wars which expensive gas and mandated fuel efficiency standards finally put to rest. The pundits will be wrong: the war started when Tesla rolled out the first Model S.

The range war, however, will be real. As with horsepower, range will be used as the primary marketing differentiator for even non-sporty vehicles. Advertising will  start with a focus on range to counter the decades of FUD sown by the very auto manufacturers now entering the electric vehicle market, but the momentum formed in making range such a central part to early advertising will be essentially unstoppable; for high end cars, performance won’t matter if it can’t go as far as the family sedan. But that’s mostly later. At five years, we’ll just be seeing the tip of the iceberg: to within a rounding error, every single ad for an electric car will mention range.

Five years is long enough that we might see some real improvements to battery technology. But probably not. At least, not much more than a continuation of the current trend, which is incremental improvements to energy density, charge speed, charge cycles and, importantly, safety. The former three stats will likely see improvements to the tune of around 5% per year, which is a respectable 28% increase after five years. The latter is likely to improve as well, but is much harder to put a number on.

All new single family homes will be built “electric car ready”. They won’t include the actual fast charger, but the heavy duty wiring will already be run to the garage or parking space  Similarly, all condo, apartment and business towers designed after this point will, at minimum, take into consideration the need to be able to install fast chargers at a future date – which is to say, without requiring drilling or expensive renovation.

Home charging stations that allow the car to power the home will be only a few percent (10 – 20%) more expensive than charge-only systems. However, regulations have (and always will) vary wildly between regions: there will still be a significant cost involved with obtaining a permit to push power onto the grid, including the special device that makes sure that a home does not push power onto the grid when the grid is not live, such as when a worker is attempting to repair tree damage in the area. Nevertheless, electric car makers will be encouraged strongly to make it easy for buyers to use their vehicles as home battery storage systems.

The pleasure of never needing to pull into a gas station will reach the public consciousness in a big way, along with the absurdly low maintenance needs of electric vehicles.

Tesla alone will sell around 400K units, with competitors somehow managing to scrape up another 300K units. Maybe. If they don’t keep building hybrids. Taking a stab at average battery capacity, I’ll say 60kWh, which will give an annual increase in worldwide energy storage capacity of 42GWh. And I’m absolutely stoked to be able to say that made that number up before looking at this info from Tesla about the Gigafactory and projected production. Although I suppose I should apologize to Tesla for lowballing their production estimates for the year 2020. At any rate, either scenario will prove once again that Elon Musk was right. In any event, Tesla will be a couple years into construction of Gigafactories two and three, with ground breaking on four and five either in progress or coming soon. Probably Gigafactory #2 will be starting limited production by 2020. The only way this won’t be the case is if another company finishes their own Gigafactory before 2018, which is unlikely.

Other car manufacturers join forces with various battery manufacturers to fund another handful of massive battery factories. Because duh.



Continued drought and flooding will do what they do to a society already on the brink. The unrest and outright war in the countries we all heard about two years ago will return, and, with it, the volatility in oil prices. Oil prices will hit $150/barrel, and will never look back. No matter how hard OPEC tries to bankrupt the wee upstarts in fracking country.

As for the upstarts in fracking country, while the return to $150/barrel oil means profits again by 2020 the well played charade is up: the reason oil is at $150 is because everyone can see fracking wells only last a few years. The party isn’t over, but last call has gone out. Natural gas prices don’t spike, but they do start to twitch a lot.


Ten Years: 2025

Sun worship

Remember solar shingles? Within ten years they will cost approximately the same as high end asphalt shingles, last just as long, perform just as well, all while weighing a tenth as much. As asphalt shingles are replaced (or a new layer added) every 20 to 30 years, this will cause a sharp uptick in the number of homes that generate many times their own power needs.

Similarly, solar siding will replace vinyl siding in price, performance and capability. Except, obviously, both of these will produce energy in addition to their primary function.

The number of new homes that completely dispense with traditional siding and roofing materials in favor of solar versions will still be small, but it will become the norm very quickly. An average such home will generate many times as much energy as it needs during the entirety of the summer. In most locations in the US such a home would produce at least twice its own daily energy requirements on sunny days, even in the heart of winter.

In high-rises (new ones, that is), “empty spaces” on the outer shell (those areas all too often filled with robins egg blue panels, like on this building) will be augmented with solar, either by simply throwing a panel in the spot, or by using a semi-transparent solar layer in front of the decorative panel. Windows that harvest energy, via coatings or integrated technology, will become the norm for high-rise commercial and up-scale residential alike. If we’re really lucky, we might have a nice combination of this solar technology with this window shading technology.

All 4000 or so Walmart stores in the United States will generate more energy than they use during the summer months. As will most buildings of similar shape, placement and function.

REI will sell a solar canopy. Not a silly little thing with a few solar panels sewn in. A stiff, canvas-like material that is the solar panel. Begley cloth is introduced to the world, joining geosynchronous orbits, waterbeds and probably a thousand other inventions first described in science fiction books. Efficiency will be low, less than 10%, but cost will be no more than double that of a similar canopy that just provides shade.

Most municipalities will have finally abandoned the practice of requiring design review for residential solar installations for ‘typical’ use cases. A most will continue to demand a straight up tax (permit fee), which is still vastly cheaper than the costs imposed by design reviews and their associated delays. Installers will still be required to be licensed, pay any permit fees and obtain permission to tie into the grid. When this policy is implemented system installation costs will be cut in half, to one cent per watt and below.

City infrastructure starts getting covered in solar. Bridges, overpasses – even highway noise abatement walls (the reflecting versions, not the absorption type) – suddenly transform, going from faded-splotchy concrete colors to an iridescent blue-black.

The first solar+battery streetlights are installed in locations that have ready access to grid power. It will be cheaper in many locations to install solar+battery streetlights than to run conduit and wires. Particularly for parking lots. Interestingly, this will help drive the dimming of public area lighting during the late hours, offering the distant but tantalizing hope that one day we might be able to see the stars from the city again.

Solar nameplate capacity will reach at least 250GW in the US, supplying 50% of all energy during brightest hours of the summer. China will have installed over 500GW, which, due to China’s massive rate of growth, will be around 50-60% as well. This assumes a year-over-year growth of around 20% from 2020 to 2025 – which is almost certainly low.

Germany will have reached a nameplate capacity greater than 150% of consumption, and will be busily pissing off neighbors by undercutting even nuclear from May through October.

At least one OPEC member will have quietly started spending a trillion dollars on solar installations and long-distance transmission capacity with the intention of supplying power to Europe.

Somewhere in the world a gigawatt-scale solar project will have been built almost completely by robots. With labor costs all but eliminated, land being cheap, and solar materials costs being nearly negligible, this project won’t bother with water cooling. The use of automation will allow for better integration into the surrounding landscape and will also avoid much of the  ecological damage associated with the actual installation phase.

Organic solar cells will hit 20% efficiency and 10-15 year working life at production scales, while still using cheap, readily available materials and roll-to-roll manufacturing. Imagine if every newspaper printed during the year 2000 was a solar panel. That’s the price point and production scale we will be chasing sometime around the year 2025.

Solar installations will return to 40% year-over-year growth.

Any sufficiently mundane magic is indistinguishable from science

Battery technology marches on. This story doesn’t require a deus ex machina (or at least not in battery technology), so I’m not going to guess at epic breakthroughs in graphene batteries or some magic marrying of diamonds and potassium to transparent aluminum. Just a nice steady, 5% annual increase in capacity and such, giving the batteries of ten years from now 55% more oomph than they have today.

The five years of electric vehicle sales from 2020 to 2025 with 25% year over year growth will have given the world nearly 8 million electric vehicles, with an associated 475GWh of battery storage. Fully charged, these cars could have powered the entire United States for over 12 hours in the year 1950, or for an hour in 2014. Production is up to 2 million electric vehicles per year, requiring 130GWh of battery packs – assuming 60kWh on average, which is low. Demand far exceeds capacity.

Tractor-trailers will go hybrid in a big way, helped by sweet, sweet gas turbine technology.

Sales of gas powered sports cars are basically zero by this point, as even testosterone soaked high school students are quite capable of comparing zero to 60 times and realizing that the average electric minivan can whomp any gas vehicle with less than around 500 horsepower: it’s a funny side effect of having enough power to muscle a box through the air at 90 miles per hour, while at the same time not having a transmission.

The effect of the range wars is the application of all improvements to battery technology to increased range, rather than reduced cost. The average vehicle gains a 55% increase to battery capacity, from 60kWh to about 90. As all performance model lines will be the first to be electrified, I’m going to increase my guess for average battery pack size to 100kWh for all years past 2025. US vehicles will likely average 120kWh, Europe 100kWh and the rest of the world 50kWh and less. China would be higher, but the massive sales of lower range or smaller vehicles will likely balance out the high performance, long range sales.

Tesla will have five gigafactories running at full capacity and ten more in various stages of planning and construction. Today’s big car manufacturers will have finally pulled their collective heads out of their asses, realizing that if they don’t stop screwing up and screwing off Tesla will just buy them for their manufacturing capacity and fire all the executives: five non-Tesla gigafactories are complete or nearly so, and thirty more are in various stages of planning and construction all over the world. At this point we are using gigafactories as a unit of production measurement, not referring to the number of actual physical buildings that will be built. Gigafactories will get bigger. Much bigger. Interestingly thought, the largest ‘gigafactory’ will never be anywhere near the size of the largest oil refinery.

Needless to say, lithium is the new oil. There are, however, very good reasons to believe that mining lithium at arbitrary scales won’t be a limiting factor to battery production.

Gas prices will be looking pretty insane by this point, well over five dollars per gallon. The europeans reading that are laughing, of course, because that would be insanely cheap in most of Europe, even today.

Nobody in Europe wants buy gas cars, preferring to go on a waiting list for an EV rather than buy any of the latest gasoline or hybrid models. The US isn’t quite as unified in shunning gasoline power, but the shift in opinion is clear. Shoppers will cite “friend and neighbor satisfaction”, cargo space, readily available free charging stations at work/home/shopping, performance, unstable gas prices, the environment, reliability, never having to stop at gas stations and, occasionally, the ability to use their vehicle as a battery for their house.

Alas, two million EV’s per year is only two percent of the world automobile market in 2025, assuming only incredibly moderate growth in total demand. (85 million vehicles were sold worldwide in 2013, 63 million if you only consider passenger vehicles). Nevertheless, manufacturers start to eat losses as delayed purchasing turns once-every-three-year car buyers into once-every-five-year buyers, and once-every-five-year buyers into “screw it, I’ll wait” non-buyers.

The first significant wave of EV battery replacements starts, with Model S and X owners paying around $6000 for a new pack when their original battery is down to about 80% of its original capacity. Their Tesla will now have a range of around 600km (375 miles), up nearly 50% from 425km. The low(ish) cost comes from Tesla purchasing the old battery pack for use in home energy storage systems, as well as cost reductions due to production scale and, you know, science. These cars are, of course, nearly ten years old, and are on their second or third owners. A few heavily driven Model 3s (or those used heavily for home energy storage) will get new battery packs as well. These buyers, being more price sensitive, are likely to wait a little longer and buy a little less, heading in for a new battery at around 70% of original capacity and choosing a new pack that ‘just’ restores the original range. This option will cost around two thousand dollars, assuming the old pack is traded in. 

The beatings shall continue until moral improves

Coal will be bankrupt. There won’t be a single coal plant online able to operate at profit. Solar and wind will make up significantly more than 50% of the nameplate capacity in the US, but in terms of real generation capability it will be more like 25%. As of 2014, coal was responsible for approximately 40% of all electricity production in the United States and solar for less than 1%. As solar and wind capacity increase (exponentially, I must add) from such low values to over 10% in 2020 and to 25% in 2025, it will be coal that gets displaced.

Not just displaced, but kicked brutally to the curb. Solar will ensure that vast swaths of summer are completely off limits during daytime hours to any generation plant that has a fuel cost. Wind will take steal away a lot of those cold winter nights. Combined, however, it’s simply the variability that knocks coal out of the competition: coal can’t spin up or shut down fast enough and efficiently enough to dance to the tune set by solar and wind. Natural gas can.

Investors will, of course, be catching on to this. Voluntary investors that is. Public utilities are the actual owners of a significant number of existing coal plants. Which is to say, “the customers” own the plants. By 2025, exactly zero US citizens will own any part of a coal plant on purpose. The plants will have a negative value: they still need to be shut down and cleaned up, but they can’t even earn enough money to pay for fuel. In order to even run them, utilities will frequently need to force other, cleaner, cheaper sources of energy offline. Because for a significant period – hours! – after a cold start, coal plants can burn half again or more as much fuel as they do after warming up. Basically, it just costs a hell of a lot extra every time you have to shut down a coal plant. And they don’t throttle worth a damn either. In both cases, pollution is extra bad and costs are extra high.

Investors are also bright enough to realize coal plants, which are very big dinosaurs sitting around a tree smoking cigarettes, burn coal. If the only customer for your product is only alive thanks to life support, and the doctors and nurses all hate him, you’re not really in a good place yourself. Nobody will own coal mines on purpose either. These, too, have huge liabilities on their books, and the accounts receivable line is looking pretty sketch even in 2014.

Sometime between 2015 and 2025 coal mining companies will all go bankrupt. It will be sudden. There may be a consolidation first, with corresponding mine shutdowns and layoffs. But with no investors stupid enough to buy into a dead industry, coal mining will become a state supported industry – just like most coal power plants are today. (If the state – being ‘the people’ – is on the hook for all the liabilities, but private investors get all the profits, does that somehow make it not state sponsored?). Coal prices will double and possibly triple during the chaos, and are unlikely to decrease significantly before coal dies completely.

Notice, however, that I am not saying that all coal plants will be shuttered. Alas, we cannot, not even by 2025. Even if we wanted to, the math is still not there: 40% minus about 25% doesn’t work out so hot. Even if we scale up natural gas production we still won’t be able to close down ALL the coal plants. Don’t forget that consumption will also rise between now, 2014, and 2025.

Fuel coal will cost twice as much as it does today, but we’ll still have to run a bunch of the plants at night during much of the year, and during many, many days in the winter.

Still, in 2025 more coal plants will have been shuttered than not. All coal plants will enter planned shut downs from around June through August or September. All coal plants will have a set retirement date – the bell tolls for thee, and for thee the bell tolls…

The spike in natural gas power generation will have more than caught up fracking production, the growth of which will be unquestionably negative by this point. Natural gas prices start to creep upward.

Ashes, ashes, the grid falls down

In 2025, fades, brownouts and blackouts will be regular occurrences in municipalities that weren’t aggressive (and lucky) with grid upgrades. Mostly limited to hiccups that last just long enough to shut the TV off and reset the clock on the microwave, but occasionally more severe issues. Those affected can’t help but notice that neighbors who have either an electric car OR solar panels on their roof (or sort of shiny blue shingles, which will be the solar panel of 2025) seem to be immune to said shortcomings in the grid. Imagine the excitement of public utility managers when this drives solar adoption to an even greater clip.

Micro-grid ‘experiments’ by solar leasing companies will be easily found in many liberal, middle income suburban areas by this time. They won’t be considered experimental any longer. The profitability of the business model will have more than caught the attention of current institutional investors with significant holdings in public utilities. Some smaller public utilities in progressive areas are able to embrace the model, but most public utilities are crushed by a bureaucracy so deep that nothing can save them. Only those utilities that have significant support from both investors and customers will even make motions towards moving to an instantaneous pricing, symmetric billing model; to say that there will be some resistance from entrenched players to the concept of enabling customers to earn money is akin to declaring the ground is down. Still, some utilities will manage to implement “hybrid” models, where customers can reduce their bills to zero and earn credit towards power later in the year (but no payouts), which will keep these utilities stumbling along for longer than those utilities that cannot manage even this. Indeed, we can already find this model today, in 2014, minus instantaneous pricing.

Those public utilities that don’t mange to present their customers to the larger regional grid as an IPP will all be struggling desperately by this time. These struggles will be laid at the feet of the very customers who are being underserved by the utility. Customers of the failing IPPs who have solar power installed won’t feel the pain in their power bills as badly as those who don’t, but only those customers with solar and some form of storage – either standalone battery storage or a battery electric vehicle – will be relatively immune. Sadly, thanks to the failure of these public utilities, instead of helping to stabilize prices these solar+storage customers become a large part of the problem for these grids. Without symmetric pricing, these customers can buy cheap, but have no incentive to sell once they’ve reduce their bills to zero. They only release power into the grid at the worst times – when it’s already free – worse than free, actually: the utility will have to “roll back” the customers meter while *paying* to put this energy onto the regional grid. Yes. By 2025, most summer days will include significant periods where instead of paying for electricity, power producers will have to pay to put it on the grid. This is already happening in 2014. In fact, it started happening long before wind and solar was a thing, way back when Enron worked their magic on Texas there were multiple periods where the cost of a MWh “exceeded” negative $1000.00. A more typical price for a MWh was around $20.

These failing utilities will be a huge problem to society because it will be those who can least afford it who will be most impacted by the wild swings in energy prices. These will be the customers who can’t afford to replace their hot water heater with a new model that is smart enough to not turn on when power is ten times more expensive than average. Who don’t have the resources to re-insulate their home, replace their windows, or switch to a price-aware heating system. Frequently they won’t be allowed to modify the home at all, because they will be leasing. Their home won’t be wired for grid storage, and even if their landlord was on board, many won’t be able to afford it. Low income apartment dwellers will be completely at the whim of the property management: heating and appliance upgrades, grid storage, EV chargers – even if a complex has electric charging stations it will likely be up to the management company to pass any benefit (e.g. net metering) back to the tenant, assuming the tenant even has a battery electric vehicle. Etc.

Regional grids that play host to multiple “new model” (instantaneous, symmetric pricing) micro-grid IPPs will be vastly more stable than regional grids that do not. Members of these micro-grids will be almost completely isolated from the wild price swings associated with sudden changes in solar or wind output (e.g. a storm front moving in), as market pricing will encourage members to sell stored power, leveling prices both within the  micro-grid as well as in the regional grid. In regions that do not host any of these new model micro-grids these dips and spikes in production will play hell with the whole regional grid, leading to “power fades” (if we still had incandescent bulbs, they would be seen to dim momentarily), brownouts and occasionally, in very poorly managed regions, blackouts. In horribly managed grids (and there are plenty) there will even be occasional damage to infrastructure; over-voltages that destroy industrial equipment, transmission line equipment committing suicide to prevent downstream damage (as it’s designed to as a last resort), etc.

Local interest stories featuring interviews and mentions of households that earn hundreds or even thousands of dollars per month as members of market-based micro-grids will fill the gaps on quiet news days occasionally. Typically these cases will be upper middle class families with teenage drivers, three or four electric cars, short commutes, and grid-tied battery storage that came with the house when they bought it. With a 30% reserve, the four cars could contribute around 200kWh of storage. During the middle of the day on during the summer months they will frequently be paid to charge these cars. Depending on the stability of the regional grid, this could range into some crazy numbers occasionally, like a dollar per kWh; those prices, however, will be driven by transient events and will not be common (or at least, won’t be very predictable). A realistic daily “average low cost” will be around (-10) cents per kWh, again, in areas with less stable regional grids. Paid $20 to charge the vehicle they will sell the power later that evening when the sun sets and the wind dies down, getting around 20 cents per kWh and thus earning another $30. Fifty dollars per day is $1500 per month: more than minimum wage.

Indeed, the interfaces into integrated home battery storage and battery electric vehicles will be quite sophisticated by this time, if only the after-market ones. An EV owner will be able to set an absolute minimum charge level, so they can get to work back the next day, as well as multiple buy and sell price levels. For example, “Fill to 70% if price is less than (-5) cents. Fill to 100% if price is less than (-25) cents. Sell to 50% if price is above 15 cents. Sell to 30% if price is above 25 cents.” More sophisticated versions will be available that use learning algorithms. Users of these systems will simply set a minimum range to reserve and a maximum price per kWh to pay.


Fifteen Years: 2030

That’s just crazy talk

In the year 2000 more people used dial-up than broadband. By far – greater than 90%. Google was two years old. WiFi barely existed. Full color LCD screens were expensive, grainy and basically crap. The game platform de jour was the PlayStation 2, which came out in late 2000. The Xbox hadn’t been released. “Time shifting” (that didn’t require programming a VCR) was a new thing; Tivo and ReplayTV were released in 1999. Napster was a year old and hadn’t been sued into subservience yet. Wikipedia didn’t exist. Neither did Pirate Bay.

As we exit 2014, cell phones can connect to the Internet a hundred times faster than the dialup connections of 2000 in many places. I own lightbulbs that I can turn on and off via WiFi. There exist watches with screens that have higher resolution than the best television did 15 years ago. My cell phone has a better screen, in every possible way but physical size, than the best PC monitor available in the year 2000, at any price. Buying an encyclopedia set is a sign of mental illness.

I’m not marching these bits of trivia out so much for the reader as for myself. The numbers start to get so incredibly large fifteen years out that even I need to remind myself just exactly how much can happen in such a seemingly short time. 

Don’t stand in one place too long…

Fifteen years from now everything will be covered in solar harvesting materials. Not literally everything, as in “everything, everywhere”, but rather every type of thing. There will be at least one example of “a thing” being covered in solar materials for nearly every “thing” that can possibly be painted, clear coated, wrapped in cloth or hidden behind a facade. Not every sidewalk in the world will harvest sunlight, but solar sidewalks will be no big deal. Not every street light, awning or semi-trailer will have a coat of solar paint, but many will. Stonework the world over won’t all have received a solar harvesting clear-coat – but many new buildings will install just such stone. Solar harvesting clear-coats will be created for cars, trucks, ‘natural’ wood fences, etc: solar clear-coats work by harvesting the energy from UV and infrared light, protecting whatever it’s covering.

We hit a milestone around the year 2028:  the United States will have covered the equivalent of one Death Valley National Park with solar – total land area: 13,650 square km, or a little over four Rhode Islands. This is assuming an average efficiency for all installed solar of only 5%. China will have covered twice that.

Five more years of capacity expansion at 40% year over year growth; the bulk of the increase in growth rate is thanks to “cheap as newsprint” organic solar.  Cumulative installed solar capacity is up to nearly 1,300GW in the US, 230% of EIA projected total energy demand for 2030. Two Death Valleys. China will have installed 2,500GW total, and over 750GW this year alone – four Death Valleys and one Death Valley, respectively.

EIA demand projections are, of course, complete shit. Forecasting 0.9% year-over-year growth between 2012 and 2040 is just… wrong. Blindly, unimaginatively wrong. China’s estimate of ~5.6% year over is somewhat more believable, but not much. In any case, cheap solar + automation is going to simply stomp the hell out of both numbers, in great part by creating completely new categories of energy use.

The mass migration of large-scale industrial energy consumers to the sunny states will be complete. Concrete production factories and food processing facilities with integrated solar concentration plants, aluminum refining, chemical refineries, steel and glass recycling, etc. Data centers will all be located in areas rich in renewables and completely sheathed in solar gathering materials.   

With the aid of advances in robotics and solar materials it will be possible to deploy huge solar installations in the desert with nearly zero risk to wildlife, at a cost barely exceeding the cost of materials and permits. What’s more, it is quite likely that the shade imparted by giant solar canopies will in fact be life savers, providing much needed shelter from year after year of climate-change induced record temperatures. By dedicating a small percentage of power produced to ‘watering the desert’, solar could very well save vast areas from desertification.

Going, and going, and going…

Year over year electric vehicle production will increase steadily from 25% in 2025 to 75% in 2030, thanks to a profound shift in public opinion that will have made driving a gas car “so yesterday”. (Keep in mind that Tesla is the only company actually increasing total vehicle production – most manufacturers are ‘just’ producing fewer gasoline powered models and more battery electric models).

Worldwide EV production will hit nearly 20 million vehicles per year. The US and Europe accounting for about half of that, China a quarter. Even with the significant reduction in vehicle ownership throughout the US and most of Europe, this is not even close to meeting demand. In 2013, 7.6 million passenger vehicles were sold in the US, 16 million in all of Europe, nearly 3 million in Brazil, 4.5 million in Japan, and a whopping 18 million in China.

By 2030 there will be nearly 51 million electric vehicles on the road and in garages. The combined energy storage capacity of these vehicles, at nearly 5,000GWh, could power the United States of 1950 for a week, or the United States of 2014 for 12 hours.

Annual vehicle production alone will require the equivalent of almost 40 gigafactories worth of battery production. This does not include replacement batteries, which will require another gigafactory or two.

Most middle income EV owners use their EV for some level of energy storage, if only to buffer big power spikes. By now, plugging your EV into a public charging station will be as likely to draw down the battery as charge it, depending on the time of day and your needs.  

The market-priced micro-grid model will be quite common by 2030. In fact, the model will be well enough established that some companies will have taken the risk of buying out and otherwise taking over some of the more poorly run, bankrupt (and despised) failing public utilities.

The median residential customer will have reached net-negative energy consumption by this time, on an annual basis. Public utilities that haven’t abandon net-zero billing are all utterly broken by this point: half or more of their customers cost them money instead of being a source of profit. 

Companies that manage micro-grids will no longer bother with power purchase agreements (PPAs, which are contracts to buy or sell power in the future), at least for wealthier areas where many clients own some form of grid storage (either an EV or integrated into the home). The point of a PPA as a buyer is to avoid risk. With a well established micro-grid, one which has plenty of battery backup, there’s little risk. Prices for power leaving micro-grids will be real-time, just like prices inside the grid. 

Once prices from micro-grids go real-time, solar and home battery storage (and ‘extra’ BEV capacity) will utterly flatten the daily price curve for  power, to below the cost of materials. Why below the cost of materials? Because large numbers of producers will be homeowners, who will consider income from selling power back into the microgrid as cost recovery: a secondary benefit from assets purchased with little consideration for their value as investments. Building new utility-scale solar installations will become unprofitable, as will replacing or upgrading older, panel-based (as opposed to roll-to-roll) installations.

Micro-grid operators, meanwhile, will be making a fraction of a cent on every kWh that crosses their grids.

Sources, etc

Smart Glass:

Up to 85GW of silicone based solar in 2016:

China Solar Targets:

Solar Growth Numbers:


85million cars sold in 2013 world wide:



New home construction plummeted from 1.2 million homes per year in 2007 to about 600 thousand homes in 2013, but I’m going to bet that we’ll see an average of about a million new homes per year between now and 2035. Maybe more if we keep having stupidly destructive storms. We really shouldn’t see any significant population effects start to take hold until sometime around 2035, though migration into cities could decrease that tally somewhat.