Concerned Citizen: “So, What is the Cost of Building SG 2.0?” The Engineer Ultra: “Ahem……about $407 Billion!”

("....and that’s not all", he said, sotto voce, "we also need about $1.6 trillion for non-Smart-Grid power system infrastructure".)

Final Avatar 80x80-Logo-SG-1-and-2-and-IX-LOGO-e1363114874895-150x150

 Breaking Down the Cost of the Smart Grid

Dom Geraghty


This dialog continues our series on The State of the Smart Grid, providing context for SG 2.0 business cases to follow. This dialog is fairly long, at about five pages (ouch!) but broken up into digestible pieces with lots of bullets -- we are hoping that you will find it gripping enough to make it to the end… is a really important topic, and we provide lots of numbers. All references are available in our Archives here, as well as directly linked to the source.

Definitions are important when you talk about costs (or benefits) of SG 2.0 and its applications*.

10. DSC_1334-150x150AMI is not the Smart Grid (SG) -- it is enabling infrastructure (iX) that supports automated meter reading and a few SG applications (SG 1.0). It lacks the band-width and speed necessary to provide grid automation and control, i.e., to make the grid “smart” and deliver the full benefits of SG 2.0 applications.

*We are using our new standard definition of SG 2.0 (see here) which differentiates SG 2.0 applications from (a) the infrastructure (iX) that enables them, and (b) legacy AMI 1.0 applications.

What Has Been Spent to Date

The power industry has made a very large investment in AMI, i.e., SG 1.0. There are about 40 million (see EIA, Pike) smart meters deployed in the U.S., costing an estimated $100 - $125 per installed unit, for a total cumulative cost to date of $4 billion to $5 billion.

Included in this number is $813 million in ARRA government funds that were provided for AMI deployments. Note that since the completion of the ARRA program, shipments of smart meters in N.A. have slipped 42% to 7.2 million in 2013 from their peak in 2011 of 12.4 million. A substantial amount of additional ARRA funding was also provided for the early deployment and demonstration of selected SG 1.0 applications.

If we want to transition to SG 2.0, it is clear that we’ll have to make an even larger investment than that which has been made to date for AMI. This investment in SG 2.0 will likely be spread out over a 20 – 30 year period as we transition to full SG 2.0 capability.

What Will the Transition to SG 2.0 Cost?

Many cost estimates have been published by knowledgeable entities. Similarly, the benefits of a fully functioning SG 2.0 have been analyzed, estimated, and published.

Overall, the estimates are difficult to synthesize because differing definitions of SG 2.0 have been used, and, in some cases, basic electricity infrastructure (iX) costs have been bundled with SG 2.0 costs. We will focus on ball-park cost numbers rather than try to be deceptively precise.

Bottom line: there is general agreement that a fully functioning SG 2.0 will cost hundreds of billions of dollars (see below).

A Selection of Cost Estimates

DSC_1253-150x150First, we will present a few of the published estimates. Then we will present results from the most comprehensive analysis to date of the costs and benefits of SG 2.0, published by EPRI in 2011.

Navigant/Pike Research estimates that a total of $494 billion will be invested in the smart grid in the period 2012 – 2020, with the majority of this spent on  transmission upgrades, automation of substations and distribution, smart grid IT, and smart meters. Of that total, they estimate that $78 billion will be spent on distribution automation and $47 billion on AMI.

McKinsey has estimated that the net benefits from SG 2.0 applications will total $130 billion annually by 2019, comprising customer applications ($59 billion), AMI ($9 billion), and grid applications ($63 billion). Cisco cites this estimate in its interesting report on the future of the SG (SG 2.0), entitled Gridonomics, which we discussed previously.

Cisco also notes that EEI has forecasted that over $1 trillion will be expended through 2030 on electric iX by IOUs, noting that while the part of that associated with smart grid technologies is not clear, Pike Research suggests that $200 billion might be spent globally on SG 2.0 -- a number Cisco says seems too low.

The Brattle Group, in its report for The Edison Foundation, projected a required investment through 2030 of $1.47 trillion in generation, transmission, distribution, and AMI/energy-efficiency/demand response, which includes “continuing the installation of the Smart Grid”. The AMI/EE/DR portion of the investment was $85 billion. They caveated that since there was no standard definition of the “Smart Grid”, they did not attempt to carve out the costs of SG 2.0 from the total estimated cost.

The American Society of Civil Engineers estimates that the United States spends between $44 billion and $101 billion a year on energy infrastructure. It predicts that if current trends continue the country will face a $107-billion electricity investment “gap” by 2020 and a gap of $731 billion by 2040. To avoid the “dire” consequences of such a “gap”, the report recommends an additional investment of $11 billion per year on electricity infrastructure – we are assuming that this number does not include SG 2.0 infrastructure.

EPRI Report on the Costs and Benefits of SG 2.0

Lastly, in 2011 EPRI published a detailed report on the costs and benefits of the SG, providing considerable granularity along elements of the SG value chain. The report concludes that $338 billion to $476 billion will be needed for SG 2.0 over a 20-year period, and that this investment will create benefits of $1,294 billion to $2,028 billion, reflecting a benefit to cost ratio of 2.8 to 6.0.

The EPRI report provides the following breakdown of SG 2.0 costs over these 20 years:

SG 2.0 Cost Category

$ Billions

Transmission and substations

82 - 90


232 – 339


24 – 46


338 – 476

(averaged cost = $407 billion)


EPRI also estimated the annual increase in customer bills (in nominal $) to pay for the total SG 2.0 costs above as follows:


Annual Increase in Monthly Bill

Increase Over 20 Years


8.4% - 11.8%



9.1% - 12.8%



0.01% - 1.6%



What is concerning about the increase in customer bills is that the numbers above relate to a maximum 20-year capital expenditure of $476 billion for SG 2.0, but we estimate that the total required capital for power system infrastructure and SG 2.0 applications will likely be over $2 trillion (see below).

This raises a major question: if the residential customer’s bill increases by ~10% per year for an (averaged) $407 billion investment in SG 2.0, what will the increase be to cover a $2+ trillion investment? Can current and future customers afford these higher bills?

What the EPRI Report Can’t Tell Us

DSC_1065-150x150The EPRI report is very helpful for ball-parking the potential costs of SG 2.0, but it has shortcomings when it comes to determining whether and how to invest in enabling iX and SG 2.0 application investments.

It estimates the total benefits “bundle” and the total costs “bundle” independently, rather than estimating the benefits and costs of individual SG 2.0 applications, and then totaling those that achieve or exceed some threshold business case metrics. Unfortunately, the report does not provide sufficient granularity to evaluate the business cases for individual SG 2.0 applications.

At a minimum, infrastructure investments with very long pay-back times should be evaluated separately from SG 2.0 applications.

And SG 2.0 applications should first be evaluated individually to assess whether each business case make sense on a stand-alone basis, and then together to prevent double-counting of overlapping benefits.

Now, Let’s Forget About the Exact Numbers…..What Are We Seeing Here?

The power industry needs a lot of capital over the next couple of decades, and, somewhat surprisingly, SG 2.0 applications aren’t the primary use of these funds. Here is a breakdown of the requirements:

  • Incremental investments beyond the historical business-as-usual approach:
    • Replacing aging plant (generation, transmission and distribution assets), estimated by Brattle at $1.47 trillion** through 2030, and including “continuing SG deployment”
    • Policy-driven costs:
      • Compliance with cyber-security mandates, estimated by EPRI at $3.7 billion for SG applications, and $32.3 billion for IT protection
      • Higher-cost power production as a result of RPS mandates
      • Increased reserve requirements due to a higher percentage of intermittent/variable power generation – including both centralized and distributed energy resources
      • Higher costs of DG versus centralized power production
      • Additional environmental regulations, e.g., CO2, Once-Through Cooling
    • Catch-up maintenance  as a result of years of deferred maintenance
  • Upgrading/overlaying of AMR,  AMI 1.0, and other meter-reading legacy infrastructure to support the transition to SG 2.0-enabling iX
    • Faster communications systems with more band-width – it seems logical to use the modern, 5 x 9s, highly secure, high band-width iX of the telecommunications sector to save costs
    • Transition to universal use of the Internet Protocol (IP)
    • Enabling 143 million customers for SG 2.0 applications at, say, $100 each, would cost $14 billion (note: not all customers necessarily need to be enabled immediately)
  • Enabling of SG 2.0 applications, estimated by EPRI at between $314 billion to $430 billion** (excluding customer premise upgrades), including:
    • Equipment such as sensors, EMS, DMS, DA, ADR, synchrophasors, FACTS, etc.
    • “Big Data” warehousing, management, and applications
    • Implementation of interoperability standards
    • ISO system upgrades
  • Customer premise upgrades for SG 2.0 applications, estimated by EPRI at between $24 billion and $46 billion

DSC_1310-150x150In summary, total capital requirements for the power sector over a 20-year period has been variously estimated at around $2+ trillion**.

**Note that given the diversity of definitions, there may be some double-counting of costs in the numbers above.

Some Good News: Factors That Will Likely Offset the Expected Increase in Customer Bills

  • SG 2.0 savings (assuming that they are passed on to the customer)
    • Reduced transmission and distribution losses
    • Other operating cost savings
    • Increased asset utilization rates
    • Deferral of needed generation, transmission and distribution capacity (the Brattle study included savings of $192 billion in generation capacity over 20 years)
    • Use of existing DG assets to offset new capacity requirements in generation, transmission, and distribution
      • Interconnection for dispatch-ability
      • Conversion to natural gas
  • Policy Initiatives
    • Incentive regulation that increases the efficiency of power companies’ operations, as discussed in a previous dialog
    • Faster or at least timely implementation of RTP, TOU, etc., to crystallize the benefits of SG 2.0 applications
  • Cheaper natural gas
  • Potential new monetization and cost-saving opportunities
    • Sale of customer consumption data
    • Rental of SG 2.0 infrastructure to providers of SG 2.0 applications
    • Utilities extending their business model into the edges of the SG value chain to manage DG and other SG 2.0 on-premise services
      • Improves technical efficiency because they are electricity experts with highly competent engineering staffs
      • Improves financial efficiency because they already have in place the required management overhead structure, financing capability, and distribution channels to their customers


For electricity customers, the future is expensive.

We have a reasonably good idea of the cost of an adequate, SG 2.0-capable power system – it is about $2 trillion over the next 20 years.

However, we are less certain about its benefits, given the lack of business cases for individual elements of SG 2.0 – if history is our guide, we should be concerned because the promised benefits of SG 1.0 have not been delivered to date.

Time is our enemy. Utilities and SG businesses need to create SG savings immediately by leveraging our existing AMI 1.0 installations (SG 1.0) – if we can build attractive business cases for SG 1.0 applications, investors will come to the table.

These business cases will generate an increasing “surplus” that we can roll-over as we transition to SG 2.0, off-setting in part the large capital investment schedule that we are facing.

To support these initiatives, regulators also need to move quickly on new policies that support, and reduce barriers to, the deployment of the SG, and especially on price-responsive tariffs that enable key benefits of SG applications to be realized.


We strongly encourage you to challenge any of the numbers presented here! (in the comment box below)

One thought on “Concerned Citizen: “So, What is the Cost of Building SG 2.0?” The Engineer Ultra: “Ahem……about $407 Billion!”

  1. Andrew Osler

    Dom, I found this a useful summary of the various analyses that have been published. i am always a bit amused by all these reports on aging infrastructure, regulatory mandates, smart grid investment needs, etc. as they never seem to address the rate issue you briefly mention. Customers will not tolerate the dollar investments identified above unless they perceive there to be value for the investment. I think there needs to be substantially more focus given to articulating the benefit side of the equation — why should we as consumers agree to this level of investment? What is the bang for the buck?


Leave a Reply

Your email address will not be published. Required fields are marked *