Are microgrids that are not interconnected to utility grids (islanded) appropriate to power data centers? Islanded microgrids may be very beneficial to single owner data centers or other mission critical facilities. This column discusses considerations, including legal considerations that relate to an independent microgrid as distinct from a microgrid that operates synchronously with the utility grid.

Although there are variations in microgrid structures, the type considered in this column is an energy generation and distribution system established and operated by the owner of a facility, which has definable physical boundaries. While technical details are beyond the scope of this discussion, briefly stated, the energy distributed by the microgrid can include electricity, gas, steam, and gas and thermal energy. A typical microgrid includes energy generation units, energy consuming facilities, power storage units, monitoring, and integrated control equipment to coordinate elements of the system. Often a combined heat and power (CHP) unit is the heart of the microgrid which, together with boilers, produces electric and thermal energy.

The growth of islanded microgrids is being fueled by numerous considerations such as the desire to achieve resilience to avoid electrical utility grid outages caused by natural or other disruptive events. Other driving factors include permitting the owner to: (i) possess operational control; (ii) enjoy flexibility of operation and the ability to achieve savings; and (iii) achieve environmental benefits (and enjoy any related financial incentives). For example, an owner may be able to manage costs by entering into long-term natural gas purchase and hedging agreements or by incorporating renewable energy generation, such as photovoltaic solar, into the microgrid.



A typical large data center has a reasonably level load profile and operates on a continual basis. It requires a very high degree of reliability and resilience. Its relative inflexibility of demand means that the data center is generally not in a position to take advantage of lower off-peak utility pricing by purchasing power under a Time-of-Use (TOU) service classification: i.e., the data center itself typically cannot readily curtail or “shave” load during periods of high electricity prices in order to purchase power during lower price periods.

While there are technological strategies that enable customers to take off-peak TOU power, such as generating power for peak shaving purposes or using energy storage (such as flywheel or thermal storage) in connection with a TOU service classification, such strategies require a significant investment in light of amounts of power needed on-peak. Although a data center will have backup generation capacity, without on-site generation such as cogeneration, it is “stuck” with the utility pricing, operational capabilities, and service disruptions. Any utility distribution grid outage becomes a data center outage that must be addressed by strategies such as load shifting or backup generation. Without a CHP unit or other form of generation, a data center has limited control, limited flexibility, and limited opportunities to achieve environmental benefits.

Islanded microgrids can be expected to offer data centers cost savings and less risk over time by providing reliability, resilience, sustainability, flexibility in energy distribution, and environmental benefits. A primary benefit is not being subject to events that affect the utility grid such as over or under voltage, defective power quality, or lack of balance.



  • A utility’s additions of plant in-service or expenses (such as the expense of procuring power under a power purchase agreement) are out of a data center’s customer’s control. Intervening in utility rate cases in an attempt to keep utility rates down are unlikely to produce much, if any, relief. Decisions on accepting proposed utility rate recovery and tariff design are in the hands of state public service or utility commissions, which are also concerned with making sure that utilities are solvent and have tariffs that do not result in lower bond ratings (which increase utility interest expenses and drive up rate recovery). Moreover, utilities that generate their own power may have fuel adjustment clauses such that their rates increase to cover increases in fuel costs. Utilities and microgrid generators are each subject to fluctuations in fuel prices so purchasing from a utility does not confer advantages.
  • An islanded microgrid controls its own fuel procurement. Cogeneration is much more efficient than utility generation and the customer avoids paying rates that cover utility capital recovery and debt and the operating costs of lines necessary to transport power from the generating plant to the customer. A data center microgrid procuring its own fuel for microgrid generation is in position to manage and hedge its costs including participating in spot markets. While it will still be subject to utility pricing if it procures natural gas from a utility gas distribution system, in some locations there are alternatives such as sourcing gas directly from landfills.
  • It is important to pay close attention to the provisions of fuel procurement contracts, particularly those that address the imposition of additional charges. Again and again we see contracts in which a client has signed a “standard form” contract with a fuel seller. These can be expected to have provisions that are highly favorable to the seller and adverse to the buyer which can come back to “bite” the buyer. A buyer of large amounts of fuel possesses bargaining power and should routinely have such contracts reviewed prior to signature since experienced counsel will be able to spot and can assist in substituting balanced provisions which protect the buyer. It is noted, however, that purchasers of pipeline gas will be subject to FERC tariffs.
  • Financing considerations. It may be possible to gain incentives and enjoy tax benefits for cogeneration and any equipment providing renewable generation in order to mitigate the substantial capital investment necessary to construct a microgrid. Over the long term, however, the avoidance of paying utility rates and paying for the microgrid’s own equipment rather than paying rates based upon the utility rate base (plant in service and expenses) is a significant benefit in itself. At the end of the day, rather than paying endlessly for utility plant in service, the customer will own its generating equipment. This advantage may be analogized to purchasing a building instead of renting it.



A large traditional microgrid, which contemplates synchronous generation with the utility distribution grid to enable sale of utility electric power, backup, and/or selling excess electric energy to the utility, can require time-consuming interconnection approvals. It may also be necessary for the owner to pay for new or costly upgrading of utility distribution lines (or transmission lines above certain voltage levels). At data center voltage levels, it is necessary to enter into an interconnection agreement with the local distribution utility. Some microgrids might even be rejected for interconnection. Interconnection requirements include the imposition of utility tariffs and the applicability of state PSC or PUC regulations. In addition, independent system operators (ISOs) may impose requirements. Conversely, a microgrid, which is entirely islanded on its own or rented on the premises with no interconnection does not require utility approval.



An interconnected microgrid is subject to all the problems that affect the grid as a whole. Given the wide geographic spread of joined systems, a utility grid problem several states away can “take down” an entire system of linked grids — as happened in the 2005 blackout from Ohio to the East coast in the U.S. and Canada. A properly designed, constructed, and operated islanded microgrid will offer stability and excellent power quality. If the owner elects to have the system operated and maintained by an outside contractor, it will be necessary to have in place an O&M agreement properly protecting the owner. All too often, owners sign the “standard” agreement presented by the contractor without proper review except for the attachments that cover specifications. The “boilerplate” can prove to be very important and the entire contract should be reviewed by experienced legal review which would prove very cost-effective should a dispute arise.



Interconnected microgrids capable of islanding must have sufficient generation to do so and any additions or changes to the microgrid can require additional testing and approvals. On the other hand, islanded microgrids can be upgraded to incorporate additional capacity or different forms of renewable generation without the necessity for further utility compliance testing or approvals.



Depending upon the state in which the facility is located, a potential advantage of a single-owner interconnected microgrid is net metering, in which a customer essentially sells electricity back to the distribution utility at its retail price. The “sale” is accomplished by “running the meter backwards” during periods when excess power is flowing from the microgrid to the distribution utility. Typically, there are a number of restrictions upon net metering such as the amount of net metered electricity provided to the utility may not exceed the customer’s load. For that reason, in light of a large data center’s likely level load profile, participating in a utility net metering program may result in few opportunities to sell much, if any, power to the utility. Another point to bear in mind is that a customer’s ability to net meter is typically limited by state regulation to amounts such as 1 or 2 MW.



Will generation within an interconnected microgrid be a QF subject to regulation? An owner or operator of a generating facility with a maximum net power production capacity of greater than 1 MW may obtain QF status with the Federal Energy Regulatory Commission (FERC) by filing with FERC. In the event that it satisfies the FERC requirements of a QF, the cogeneration unit will generally be exempt from most federal and state regulation. The following categories of QFs are exempt from the Public Utility Holding Company Act of 2005 and state laws and regulations concerning rates and financial and organizational requirements utilities:


  • Cogeneration facilities of any size
  • Small power production facilities 30 MW or smaller
  • Geothermal and biomass small power production facilities of any size
  • Renewable generation facilities below 80 MW and certain other small power production facilities


Conversely, a generating facility to be used entirely for the purposes of a microgrid that is not interconnected and is used to provide power for a single customer on private property would generally not be subject to regulation.



  • Potential utility franchise issues. In the event that a data center installs facilities and distributes electricity over public streets or the electricity generating equipment and wires connecting the loads are under different (unaffiliated) ownership on different non-contiguous geographic sites, the set-up is likely to infringe upon the local utility’s franchise rights. In order to avoid a franchise issue, it is usually necessary to ensure that the entire microgrid is on a single site.
  • “Over the fence” sales. Depending upon the state, it may be possible for a microgrid to sell power to a neighbor (with contiguous property) without interfering with the franchise of the local utility — as long as the lines do not cross public property and the power is not transmitted or wheeled using the facilities of the local utility.
  • Power purchase agreements. In a “hotel” situation, where several data center customers purchase power from a third-party microgrid\generation owner on a single site, it is necessary to have power purchase agreements and, if applicable, thermal energy purchase agreements between the owner of the generating and distribution microgrid and the data center customers.
  • Submetering requirements. To the extent that the third-party microgrid\generation owner sells power that is partially provided by the electric distribution utility to customers on a single site, the state submetering regulations should be reviewed for compliance. These issues are avoided when an islanded non-interconnected microgrid is selling power to its own tenants on the same site.
  • Standby fees. When an interconnected microgrid owner takes standby service classification of the local utility to meet peak demand (or to provide back-up in the case of a microgrid generating outage), the microgrid owner should review and understand the application of standby fees. These can be very high during summer peaking periods when the cooling requirements of a data center are the greatest. An islanded microgrid avoids the risk of high standby fees.


Although there can be significant advantages to an island that is not connected to a microgrid at all, these must be weighed against potential legal and practical impediments in addition to performing a cost\benefit analysis.

In conclusion, if a single-owner microgrid generates and manages all the power it requires and there is no interconnection with a utility grid, the microgrid has the ability to achieve significant benefits and avoid several potential problems. Once the entire “landscape” of considerations is taken into account, islanded non-interconnected microgrids are increasingly seen as the answer to controlling energy costs.

This column is intended to provide general information and not specific legal advice. Any reader should consult an experienced attorney for advice regarding the issues discussed above.