DG vs. utility-scale solar
Say you are looking to invest directly into solar projects. Generally speaking, your money can go into either A) huge, centralized, utility-scale projects that feed energy directly onto the transmission grid, or B) smaller, distributed projects that deliver energy onto local distribution grids or directly to a single customer. Now imagine you were forced to choose between investing all-in to either A) or B) – which would you choose?
My answer would be B): distributed generation (“DG” for short).
Obviously, no one is forced to make such an absolute choice. DG and utility-scale solar are sub-sectors of an industry with a great secular tailwind that is growing in strength with more significant government mandates on the horizon and continued improvement in costs and technology. Both segments have and will continue to benefit from these trends. So you might assume that my reasons for preferring DG over utility-scale are found in the weeds such as differences in unit economics, the amount of capital flowing into one versus the other, the various incentives, the supply chains, the operators involved, or engineering challenges. Actually, my reasons are high-level and stem from some basic principles about risk. These ideas have been aptly described by two well-known thinkers / authors, Morgan Housel and Nassim Taleb. The following is a quick briefing on their ideas before we talk about how it applies to solar.
As Morgan Housel notes in his blog, real risk is found in the tail end of the distribution. The low-probability, high-impact events are all that matter. Wars, pandemics, and economic depressions hopefully only occur once every long while, but they do indeed occur; and when they do, the consequences are huge and widespread. Nassim Taleb referred to such events as ‘black swans’ in his 2007 book of the same name.
In Taleb’s 2012 book Antifragile he explains the ‘triad’ concept. To summarize it concisely, a risk factor being realized will generate one of three outcomes for the entity exposed to the risk (whether it be people, communities, businesses, etc.).
The entity is damaged by the risk event. With respect to this risk, the entity is deemed to be fragile to this risk.
The entity is unaffected by the risk event. The entity is considered robust to the given risk.
The entity is improved as a result of facing the risk event. This entity is considered antifragile to the risk.
A classic example of fragility is the large nation-state that is brought to its knees by war, pandemic, or natural disasters. Hasn’t every major centralized power in history been brought down by one or a combination of those stressors?
Antifragility is summed up by “What doesn’t kill you, makes you stronger”. An example along the same lines as above is Switzerland. The country is essentially governed from the bottom up, a bunch of municipalities called cantons which constantly bicker and resolve minor internal issues. The stress events cause no permanent damage and the process actually helps each canton (and, as a result, the nation as a whole) improve incrementally. Moreover, Switzerland has long been a safe haven for people and their wealth during times of global political crisis. Thus the country actually benefited from black swan events like the two world wars.
So let’s bring this back to the question about centralized solar vs DG solar. Compare a 100 MW utility-scale project located in the Mojave Desert, to a 100 MW distributed portfolio consisting of 200 individual 500-kW projects serving power to small and medium-sized businesses throughout a broad swath of Southern California. Institutional investors focused on solar are choosing between opportunities like these two all the time. Let’s see how some low-probability, high impact events are likely to affect each of these assets.
1) Natural disaster - lightning strike, fire, tornado, earthquake.
At the utility-scale project, this kind of event can knock a minor portion, a large section, or even the entire project out of commission for a while, depending on where exactly the damage is. There are hundreds of thousands of individual solar panels, but all located in the area of perhaps one square mile. Generated power is routed through just a handful of inverters, so just one inverter going down is a big problem. And, the entire project’s output has to pass through a single substation and transformer to connect to the grid, so there is a single point of total vulnerability.
For the DG portfolio, with individual projects spread out over a wide region, a single natural disaster is unlikely to affect many of them at once. A wide-ranging earthquake may be a possible exception, but keep in mind that each individual 500 kW project is only 0.5% of the entire DG portfolio.
Both utility and DG are fragile to this kind of risk event since they both suffer, but utility-scale is much more so.
2) War and terrorism.
Being critical infrastructure, large centralized power plants are high-value targets for enemies. For example, the Palo Verde Nuclear Generating Station in Arizona was known to be a target of the Soviet Union during the Cold War, and when the Iraq War started in 2003 the National Guard was deployed there to protect the site, even though there was no specific threat to it.
As for DG, it makes little sense for an enemy to attack hundreds of individual small power stations.
Utility-scale is fragile, and DG is robust to war and terrorism.
3) Grid failure.
The risk of the grid failing might have appeared remotely low to most people until recently, but blackout events in California during the 2020 summer and in Texas during the February 2021 cold snap have raised this concern in many peoples’ minds. The grid is vulnerable to more than just extreme weather – a large coronal mass ejection from the sun that sends charged particles towards Earth could create havoc on the grid. These ejections are what cause the Northern and Southern Lights so we know that they happen regularly. And of course, the grid has been the victim of cyberattack.
How would a grid failure affect DG solar? When the systems are located behind the meter (meaning they are built on site to serve a single customer), they do not rely on the grid to deliver power to the host customer. Then, during the immediate aftermath of the event, there would be a spike in demand for DG equipment (solar panels, batteries, and backup generators) to restore power. Finally, there would be a step-change in awareness and demand because the black-swan event would get seared into people’s memories; a high proportion of the people directly affected, plus many who were not but felt lucky to have avoided the disaster, will look to DG as a necessity. Thus, DG would be antifragile in the event of a grid failure.
The utility-scale project, on the other hand, is dependent on the main grid functioning properly. A grid failure would lead to the project being “islanded” and unable to deliver power. Thus utility-scale solar fragile to grid failure.
These are certainly not the only tail risks to which solar is exposed. The pandemic did not have an apparent effect on one over the other – snarled supply chains and rising costs are problems for solar projects of any size. We could try to think of and discuss more known risks, but Morgan Housel notes that “risk is what’s left after you’ve thought of everything”. Meaning, we’ll never be able to come up with every possible risk event that could occur in the future; consequently we can’t conclude that DG is unequivocally more resilient to all major tail risks than utility-scale. Nevertheless, the risks discussed here have happened before and/or can certainly happen in the future, and with respect to these we can say that DG has the clear advantage.
Sources:
https://www.collaborativefund.com/blog/the-three-sides-of-risk/
Antifragile, Nassim Taleb
https://en.wikipedia.org/wiki/Palo_Verde_Nuclear_Generating_Station