C3 Energy, the Silicon Valley startup founded by software billionaire Tom Siebel, has quietly been building on an audacious promise: a big data integration and analytics engine, hosted in the cloud, that can aggregate and put to use all of the world’s information, practically speaking, as it pertains to the complexities of big energy systems.
That can include demand-side energy and resource management from the building-specific to the enterprise-wide scales, which was the business that C3 Energy first launched back in 2009. Its list of customers for that line of work now includes Cisco, Hewlett-Packard, General Electric, Constellation Energy, Adobe, Dow and Masdar City, to name a few.
But it can also include utilities, energy companies, and the makers and implementers of smart grid technologies -- and over the past year or so, C3 has shifted its focus squarely onto the utility and energy sector, company President and CTO Ed Abbo said in an interview at the startup’s offices this spring. On top of its existing customer engagement portals and five initial grid applications, C3 Energy plans to release “a comprehensive set of smart grid applications that span the demand side as well as the supply side of the grid,” over the next twelve to eighteen months, he said.
The modules include grid-centric tasks, such as voltage optimization, asset management, fault detection and outage restoration, and integratingsolarpanels and batteries, to customer-centric systems like demand response, revenue protection, load forecasting and customer segmenting and targeting. Its first customer, Pacific Gas & Electric, is starting with a “revenue protection” module, marrying its existing customer management and energy efficiency analytics engine (also built by C3) with grid data to catch and prevent energy theft or unbilled energy use.
C3 is also working on unspecified projects with utilities including PG&E, Constellation, DTE, Commonwealth Edison, Entergy, Northeast Utilities, Southern California Edison and Iberdrola, Abbo said, either with partners like GE Energy or on its own. Indeed, the data engine is only the first step, he said. The next steps will come as developers build new applications that use, and add to, the analytics engine's total store of knowledge and capabilities.
In short, it’s a hugely ambitious goal, and not one C3 is tackling alone. The list of “big data” companies in the smart grid, energy efficiency and supporting IT integration fields ranges from giants like IBM, Teradata and EMC to startups like AutoGrid, Verdeeco, Opower and Bidgely, to name a few -- and several of them, such as SAIC, Cisco, Wipro and Accenture, are also C3 customers or partners.
And, as with so much in the big data world, C3’s data integration holds the promise of unlocking far greater value than its relatively small cost to implement, Abbo said -- though he didn’t provide any price specifics, and the startup hasn’t said how much it’s charging for its software and services so far. To prove out the point, C3 commissioned (and used its big data engine to help conduct) a study of ten of the biggest U.S. utilities, and calculated that C3’s approach to handling its long list of smart grid functions could provide insight into how to achieve a collective $12 billion in annual savings and additional value, he said.
That’s on top of simply implementing each separate smart grid feature as a standalone unit, he noted. In other words, when it comes to big data, the whole is meant to be much greater than the sum of its parts.
The Big Data Foundations of Energy Management
C3 has attracted lots of high-profile board directors and executives, including deep political-energy expertise (former Secretary of State Condoleezza Rice, former Energy Secretary Spencer Abraham, former DOE Assistant Secretary Cathy Zoi) and industry expertise (former U.C. Berkeley Dean of Engineering S. Shankar Sastry, former Constellation Energy CEO Mayo A. Shattuck III).
It’s also invested about $100 million over the past four years in support of what’s now expanded to an energy industry and grid-scale big data project. According to U.S. Securities and Exchange Commission filings, C3 raised $15.8 million in two rounds in 2009, closed a $30 million round and raised another $30 million of a $48 million offering in 2010, and most recently closed $15 million of a planned $30 million round, putting its reported investment at just more than $90 million.
As for what it’s built with the money so far, Abbo and Houman Behzadi, C3 Energy’s senior vice president of products, walked me through a demonstration at the company's Redwood City, Calif. offices, starting with the underlying structure.
On the right are all the sources of data that C3 is collecting, at whatever speeds they can make themselves available -- and it’s a pretty exhaustive list. To drill down into building characteristics and energy analyses, for example, C3 taps data stores like CalStar, Data Quick and Google for richer, address-specific data, as well as benchmarking data stores like Energy Star and CBECS, plus utility rebate and pricing programs around the country. Smart meters (AMI) or meter data management (MDM) systems fill in the energy information -- though C3 can use more old-fashioned monthly billing data to start, and some customers go deeper into building-specific energy use, Behzadi noted.
Then you’ve got all the enterprise and operational systems that are both receiving data from, and feeding data back to, the C3 engine as a whole, he said. That can include ubiquitous enterprise billing systems, customer relationship management and call center data, workforce management platforms, phone and gas cards, and the like, as well as the specialized energy systems like asset management, outage and distribution grid management, GIS and mapping systems, or demand-side load control and distributed resource management, that tie it to the utility operations side.
Behzadi chose as his demo subject PG&E, which uses C3’s demand-side energy management (i.e., customer and building energy analysis) to support its biggest power users like Costco, Sutter Health, the U.S. General Services Administration, and others, as well as medium and small-sized businesses.
“This is more than 30 billion rows of data right now,” he said as he clicked from a map of a big-box store’s properties in California’s Central Valley, then drilled down to interfaces that benchmarked properties from best to worst in terms of energy cost per square foot, or checked the status of every maintenance and repair order and efficiency project on a property and portfolio level. Abbo noted that C3 Energy is running over 10 terabytes at PG&E via its software-as-a-service (SaaS) delivery mechanism.
Charting best-to-worst energy performers among a portfolio of buildings is not a unique idea to C3 Energy, of course, nor are many of the other core energy analysis concepts it’s applying. But what most other building energy data analysis software platforms don’t do, Abbo said, is pull their data in as near to real time as each individual source allows, and then reintegrate it in a closed-loop fashion, more or less on the fly.
“That is the big difference here,” he said. “Where you start to get to where real time matters, the ability to handle events as they’re actually happening becomes significant.” Company documents get into more technical details on just how it does this, noting its support for both Hadoop-based batch data processing and real-time data streaming services, its multi-layered security model from physical computing to network and application stacks, its data quality services to identify missing and erroneous data, and its use of a metadata-based model-view‐controller (MVC) framework in the user interface, to name a few.
What does all of this mean to the enterprise customers trying to use it? Ali Ahmed, senior manager of Cisco’s global energy management and sustainability practice, told me in an interview last month that C3’s ability to manage streaming data and automate ETL (extract, transform, load) functionality was one differentiator that helped Cisco pick it out of a list of fifteen contenders.
Cisco has done a lot of work on building automation and energy management technology, and is pulling all other sorts of information about its properties in hourly or 15-minute intervals, as well as at the speed of alerts or warnings, he said. C3 Energy now integrates data from Cisco energy management partners including Johnson Controls, Ecova, Entech and CB Richard Ellis to help it convert what used to be a months-long enterprise reporting process into minutes of behind-the-scenes calculations, he said.
“We’re always expanding the data streams we’re putting into C3,” he added. In Europe, Cisco is connecting its company cars and fuel cards to the engine to calculate Scope 1 and 2 emissions, as well as to run analyses of how to cut costs from those budgets, for example. Starting later this year, Cisco is planning to start crunching the data on its corporate roster of efficiency and sustainability projects, and using it to analyze potential projects on its roster -- an example of how early projects help feed data back into the engine, making it more accurate and useful over time.
The Big Data Engine at Grid Scale
All of these real-time, massive-scale data integration and analysis capabilities have applications in the utility and energy industry. Unlike the slow-developing market for enterprise efficiency and sustainability software, the markets for next-generation IT integration and data management expertise in the utility sector now stands in the billions of dollars, GTM Research estimates. Then, of course, for a startup promising to squeeze inefficiency, waste and sheer lack of information out of the world's energy streams, there's no better place to start than the energy companies themselves.
C3 Energy’s first forays into the grid itself, however, stem from its work on the demand side of the equation, that is, power customers. C3 got started on the smart grid front with its May 2012 acquisition of Efficiency 2.0, a startup with an online platform that connects utility residential customers, and has rebuilt it on the C3 Energy data analytics platform such as Pacific Gas & Electric, DTE Energy, Entergy, Northeast Utilities, Southern California Edison, NYSEG and others, as well as in partnership with smart meter vendor Itron.
Marrying building energy insight into revenue protection for PG&E was a logical next step into the grid for C3 Energy. Using smart meter data and its suite of property, mapping, benchmarking and financial analysis resources, the utility can scan regions, neighborhoods or individual properties to see where potential theft or tampering with power has occurred, along with the costs and revenue opportunities associated with dealing with each instance, Behzadi said -- all the way down to past and present compliance actions against individual customers. (Editor's note: PG&E has a fair amount of energy theft, notably in support of Northern California’s marijuana growing industry, though it has nothing like the double-digit non-technical losses of utilities in developing nations such as India and Brazil.)
PG&E can also use the set of tools it’s developing with C3 to find inactive meters still consuming power, unreported outages that might appear to be stolen power, and other such variations from the norm, he added. Telling one cause from the other is part of the data analysis package, as is putting together top-line figures like how many millions of dollars could be recovered, at what costs, from different strategies for dealing with all those issues, he said.
As we’ve noted before, the idea of tapping tons of smart grid data to yield useful, cost-benefit defined information for utilities isn’t unique to C3. Simply collecting and generating bills from millions of new smart meters requires some hefty data management chops. Integrating multiple utility systems presents a manifold big-data challenge, one that utility IT and systems integration giants like IBM, ABB/Ventyx, Oracle, SAP, Microsoft, Accenture, Capgemini, Logica and many others are all tackling from one angle or another.
At the same time, smart grid startups from home energy analysis and efficiency startup Opower, to demand response and customer data integration specialist AutoGrid, are using the latest in big-data technologies to tackle specific utility needs. Partnerships on the big data-smart grid integration front include Dell and OSIsoft, Itron and IBM, Silver Spring and EMC, or Ventyx and E.ON, to name a few examples.
But Abbo said that he hasn’t seen any other efforts including both the breadth of smart grid functions that C3 Energy is targeting and the real-time nature of its underlying architecture. In particular, the platform’s speed lends itself well to managing grid events, such as voltage management and dynamic line rating, that happen very quickly, he said -- “We can scale this up to handle tens of millions of messages per second.”
And, because it’s a cloud-hosted service, “The platform is extensible, which allows our customers to build out their own analyses on the data, and configure their applications to add to this set,” he said. That’s allowed partners like GE to use C3’s big-data smarts to build a set of grid analytics capabilities that bring its own expertise to bear, not only in GE’s massive installed hardware base, but in its significant push into software and analytics to support its goal of networking everything from power turbines to dishwashers.
Likewise, Cisco is working on integrating the C3 analytics engine with several utilities in the coming year, Ahmed said. That move could open up its combined building energy smarts to customers interested in using it for the grid, either by building their own apps if they’re big utilities, or tapping one of C3’s modules, or a partner’s module, or a module co-developed with one or several partners, etc.
C3 Energy has even worked with Accenture and McKinsey & Co. on a "value-based modeling" service that enables utilities to identify and validate the unique value proposition of smart grid analytics for their specific operations," according to corporate documents. In other words, it's a big data platform built to tell customers whether or not they should invest in more big data projects or not -- for both customer and utility applications.
At some point, one’s stance on the future potential of all this technology relies on faith in the promise of its creators that it can indeed learn as it goes, turning ever-more-granular data points into ever-closer approximations of hugely complex real-world systems. At the same time, capturing the maximum value out of this massive data integration and analysis task will take time.
C3’s projection of $12 billion in savings for ten big U.S. utilities, extrapolated to the country’s 3,500 or so utilities, equates to $50 billion a year in value, Abbo noted. But it also assumes a scale of integration across traditionally siloed utility business units and IT architectures that’s hard to contemplate today. C3 used its own data platform to come up with its conclusions, and is offering the same platform as the long-term solution to capturing all that value. It remains to be seen how accurately its projections map to reality.