The Power of Electricity Innovation

[Seattle, 16 November 2017] Today, Brian Klebach and CAPRE organized the Third data center summit in the Pacific Northwest. Here is an outline of my points during a panel discussion on ”Innovations in Data Center Power Generation” with Erik Stockglausner and Sebastian Lykke Moller.


Electricity generation and distribution present interesting opportunities for data center innovation; and I would use innovation in the sense commercial development of technology, a new solution or a new business model. Motives for development in this area are certainly not lacking. First, there are strong (defensive) motives to tackle hurdles and problems. Second, new power technology and novel use of assets can generate substantial value both to data centers and the energy markets.

So, this is where Plato meets Edison.

“Necessity is the mother of invention”

Innovation in data center power generation can be motivated by a defensive strategy to invent around failures and shortcomings that are jeopardizing the data center’s operation and expansion. Solutions in this category involve on-site generation with gas engines or turbines and off-grid power production with solar panels.

The common factor for defensive solutions is that there is necessity to invent around failures or shortcomings that would otherwise have a negative impact on the data center’s business. These are, typically, either grid failures or policy failures. Grid problems involve insufficient capacity for power distribution, transmission losses, excessive grid prices, risk of power outage due to grid failure or monopoly behavior by the utility. Policy problems involve insufficient market incentives for renewable power generation or distorting regulations favoring on-site power production. Interestingly the commercial motives for defensive innovation are stronger the more severe the grid and policy problems are.

It could be argued that the most straightforward solution is to deal with the source of the problem, i.e. a bad grid or a bad policy. Insufficient or unreliable grids can be strengthened with additional investment and renewable power can be fostered with good policy.

Hoping for a better grid or a better policy is often unrealistic particularly for a single company. In these cases, the most straightforward and efficient way to deal with a problem is to make a site selection that meets the need for a reliable and strong grid as well as abundant supply of renewable power. It could be hypothesized that with globalization of the data center industry, i.e. improved connectivity and the expansion of international cloud companies which made computing more footloose, could – over time – drive a larger share of capacity to locations with less grid and policy problems.

“Our greatest weakness lies in giving up”

Interestingly there are complementary offensive motives to make energy innovations in the data center industry. As power capacity and energy consumption represent a significant cost, incentives to develop technology and solutions that are more efficient such as free cooling, direct current distribution or distributed redundant UPS topology, are quite strong. This is straightforward.

More thought-provoking and less obvious are motives that relate to benefits that stem from external synergies with energy markets. This is an area where there are potential revenue streams from utilities, energy companies and grid operators that could make contributions to the data center’s fixed assets.

First, energy storage is an obvious candidate for value creation both for the data center and the energy market. With intermittent energy, particularly wind and solar, expanding rapidly in many markets the need for energy storage that can match supply and demand over time is increasing. Data center UPS systems, e.g. built for two-way supply and with a lithium ion battery bank, can serve as a valuable system for uninterruptable power supply for the data center while also providing frequency containment services to the power grid.

Second, flexible and rapid onsite power generation is another candidate for value creation. Power systems are sensitive to peak capacity to maintain stability and to avoid outage. Reserve power generation as a back-up can serve as a very useful asset both to the data center and the energy market.

Third, value creation with server heat presents a very interesting area of development. Traditionally, cooling has been a problem and a cost to the data center. Recovering and recycling waste heat can turn this cost into revenue if heating is sold to a district energy company. As recycled energy has the potential to replace electrical boilers or power consumed by heat pumps in the heating market, it will make contributions that free up power generation capacity on a system level. Furthermore, recycled heat could go into valuable energy production, including heat-to-power solutions or production of biomethane.

A final word

One of the great advantages of the evolution of the hyperscale data centers is the energy efficiency gains these data centers bring. Moreover, these facilities – thanks to size – have the potential to make material contributions by generating new synergies at utility scale.

Mattias Ganslandt, CEO and President of Multigrid Data Centers