The story so far:

California's Self-Generation Incentive Program (SGIP) is an $83 million per year renewables incentive that provides a generous subsidy to distributed energy resources such as CHP, wind, advanced energy storage and fuel cells.

The subsidy is intended to spur development of low-emissions distributed generation. But the complicated calculation of just what qualifies as "low emissions" has been the subject of debate at the California Public Utilities Commission (CPUC), since the calculation could exclude fuel cells or small gas engines or even storage from qualifying for the program.

First, SGIP's CO2 math test

If you're installing distributed energy generation in California and want access to the sweet SGIP incentives offered each year by the state, you soon might have to meet the new SGIP GHG Eligibility Factor as defined below.

Got that? (We dive into the details of the formula here.)

The fix is in

In the midst of the debate over the exact math to use to judge just what qualifies as low emissions, the CPUC decided to offer about $40 million of the 2016 SGIP total for awards on February 23 of this year.

As we reported, Stem and one other firm were able to secure the first 56 applications in this solicitation and were able to monopolize the online submission process for the first 2 or 3 minutes of the live opening.

So the CPUC requested that Energy Solutions, the administrator of the “first-come, first-served” applications process, "provide a technical analysis of the February 23, 2016 automated SGIP application launch."  

Investigating server and application logs

Energy Solutions reviewed server and application logs from the day of the program opening to see if applications were submitted in a way that violated the Terms of Usage (TOU) or the SGIP Handbook requirements.

Energy Solutions found that:

  • "The application database accepted submissions and assigned timestamps to 658 reservation requests totaling $181 million in requested incentives in the first 10 minutes following program opening at 8:00 a.m. [There were 898 reservation requests on opening day.] Under heavy load, the system responded slowly to user requests, and many system users complained of significant delays in attempting to submit their applications. The timestamps recorded in the database showed that two applicant companies, Stem and Sunpower, were successful in submitting and receiving timestamps for 56 total applications in the first two minutes, during which time no other applicant companies in the program completed a successful submission."
  • "Stem and Sunpower both appear to have been submitting applications from within the Rackspace server network. The online portal is hosted on Rackspace servers. Stem attempted to submit each of its applications more than once, for an average of 2.37 attempted submissions for each application. This was not unique, as 36 of the 79 applicants attempted to submit at least one application to the system more than once."

Energy Solutions identified three kinds of activity as potentially in violation of policies in the SGIP Program Handbook or the Terms of Use:

  • "Three applicants used a POST-only technique to submit applications without using the graphical user interface. This violates two sections of the Terms of Use." [You will not alter, modify, create derivative works of, decompile or otherwise attempt to extract source code from us unless we give you express written permission.] These applicants are Stem, Sunpower, and Swell Energy."
  • "Many applicants attempted to submit individual applications multiple times. This may be construed as a violation of the Terms of Use, but did not in intent appear to violate the policies in the SGIP Program Handbook relating to submission of duplicate applications." According to the rulebook, "Duplicative applications are considered a program infraction."
  • "Some activity appears to indicate the sharing of username and password information among multiple people, although this cannot be proven conclusively. If this were true, this would be a violation of the Terms of Use."

Stem: Winning hacker

According to the memo:

  • "Stem utilized both the POST and the manual methods during the first ten minutes of program open. With these methods the applicant was able to successfully submit 95 percent of their total 246 applications in the first ten minutes."
  • "For the 233 applications successfully submitted in the first ten minutes, Stem initiated 569 submission requests (389 POST/180 Manual)."

Note: SunPower’s applications under the SGIP program were submitted by Stem, their storage manufacturing partner, on SunPower’s behalf. 

  • "SunPower utilized the POST method and submitted 100 percent of their 21 applications in the first ten minutes of program open."
  • "For the 21 applications successfully submitted in the first ten minutes, SunPower initiated 33 POST submission requests and two manual requests which had an insignificant impact on the outcome for this applicant."

Swell Energy: Honorable mention hacker

  • "Swell Energy attempted to first access the database with the POST method prior to the program opening which failed due to the lock-out (by design). This was the only clear evidence of an attempt to directly access the portal prior to 8 a.m. (Other attempts were to load the homepage or other page of the portal)."
  • "Swell then attempted the POST method 807 times for their 81 applications, which failed to successfully submit their applications because they requested the URL incorrectly."
  • "Swell then utilized the interface to manually submit their 81 applications, all of which were received after 08:10:00 a.m."

Stem has a response

Stem looks to keep the current program results and notes, "Stem has  provided information regarding its application submission process to Energy Division on a confidential basis, and has offered to provide the Commission with any other information it needs to further confirm the clear explanation and conclusion regarding the February 23 program opening."  

Stem claims, "Energy Solutions made clear in their presentation that the various techniques used by parties to submit applications were normal and anticipated uses of the Portal as it was designed and not malicious or outside the site’s terms of use."

Changing the subject, "Stem strongly supports the implementation of a lottery mechanism that includes a merit-based weighting factor as the best solution for more effectively distributing SGIP funding. This merit-based weighting factor should be calculated based on both historical performance of an individual  developer (e.g., total  number of applications submitted versus total amount of projects seen to completion) as well as individual project viability per application."

Ted Ko of Stem told GTM, "We will reiterate that Stem operated within both the SGIP program rules and the Portal's Terms of Use, and that this fact was validated by the country’s foremost firm in internet law, specifically in the construction and litigation of website terms of use. Again, it is important to note that Stem submitted 37 percent of the overall applications on that morning, and our project acceptance rate was well within the historical norm."

Ko also points out that the term "hacker" implies "malicious intent and is certainly not validated by any legal analysis of these proceedings." Andrew Meyer at Swell Energy didn't seem to like the term either.

SolarCity has a suggestion

SolarCity, along with Avalon Battery, Johnson Controls, Sharp Electronics, Sonnen, and Swell Energy (our honorable mention), has suggested that the funds allocated in the February 23, 2016 solicitation "should not be awarded, and instead the results should be voided and these funds should be rolled into future SGIP solicitations."

SolarCity comments on the SGIP method and offers an alternative process:

Specifically, SolarCity believes the current SGIP structure of 60 percent cost coverage has resulted in reported project costs that are inflated to maximize incentive levels.

What is needed is a pricing discovery mechanism that will force SGIP participants to reveal the lowest incentive level necessary to develop projects such that the maximum number of projects can be funded with the limited SGIP budget. Once initial pricing discovery is achieved, the incentive levels should decline over time in a predictable manner to drive cost reductions.

Thus, rather than attempting to set SGIP incentive levels administratively through speculation in the program re-design, the Commission should conduct a “pricing discovery round” so that the Commission can see what SGIP incentive levels the industry as a whole would require for advanced energy storage (AES) projects, and how different projects, developers, and technologies may vary in the incentive levels they require to move projects forward. The fundamental goal of this pricing discovery would be to set an incentive level whereby the highest possible capacity of projects could receive funding, thus bringing demand for limited SGIP funds more in line with supply. 

In order to achieve this pricing discovery, SolarCity recommends that the Commission conduct a single reverse auction whereby developers would submit requested incentive values for each of their AES projects that would then be ranked on a $/kWh basis. The Commission could consider this reverse auction as “Step 0,” with the results being utilized to then determine the appropriate SGIP incentive level for “Step 1” of a California Solar Initiative (CSI) style program. 

Thereafter, incentive levels would decrease in a step-down structure as proposed by the Commission Energy Division staff in November.  To ensure this competitive framework in the AES category does not otherwise negatively impact the ability for new technologies, smaller companies, and smaller projects to participate in SGIP, SolarCity would recommend the following measures be taken to ensure adequate funding remains available for all parties: 

Instituting a 5 percent residential carve-out within the AES category for systems <10 kW.  This carve-out would utilize a similar incentive step-down structure with Step 1 being determined by the Commission as necessary to spur demand

Institute a 1 MW incentive cap for AES projects, to ensure no one project receives an inordinate amount of funds. This cap would not prevent >1 MW projects from participating in SGIP, but those projects would only be eligible for SGIP incentive for the first 1 MW of capacity.  

Instituting a higher reservation fee of 5 percent of requested incentive, requiring that checks be submitted at the time of application submittal, diligent enforcement of proof of project milestone (PPM) dates, and strict penalties for recurring project cancellations. 

A 20 percent developer cap should be instituted within the AES category to ensure no one developer can consume an inordinate portion of funding within any given step

Green Charge, FuelCell Energy comment

Green Charge Networks adds, "First, if there was true malicious activity in the process or on the system that allowed for this prioritization, all applications associated with those vendor(s) should be rejected. This is reasonable, as limited funds should not be allocated to vendors that intentionally circumvent a fair and equitable process and it is prohibited by program rules." 

FuelCell Energy suggests, "In order to avoid potential legal challenges from damaged participants, and to preserve the reputation of SGIP, the commission has no choice but to completely rerun the application process."

Not “criminal behavior”

Energy Solutions concludes, "However, while we cannot say definitively if any laws were violated, we can offer our 'lay' perspective that none of the activity observed is of the type that would be commonly be considered 'criminal behavior' as would malicious 'hacking' or intentional denial of service (DOS) attacks. The types of activities observed -- sharing passwords, multiple submissions, reverse-engineering web pages, and scripting access to websites -- are activities that are commonly employed on the internet for valid reasons, although they often do violate standard terms of use as they may circumvent attempts to regulate access and maintain security as was the case in this instance."

Surely, the current implementation of the SGIP, with its broken applications process and flawed incentive structure, is not in the best interest of the environment or the people of California. 


The SGIP is the original California renewables incentive, predating the California Solar Initiative. The SGIP was intended "to provide incentives for any distributed generation resources that the commission determines will support the state's goals for reductions of emission of greenhouse gases.

Source: Base SGIP Incentive Levels for Eligible Technologies (from 2015 SGIP Handbook)