Introducing a big data prediction engine for the power grid

A startup called AutoGrid is looking to provide big data analytics — the same type of analytics that companies like Amazon(s amzn) and Netflix(s nflx) use for their web recommendation engines — for the power grid. Tools like this will be needed, given that down the road there could be more data created by the power grid than by the Internet.

AutoGrid is officially launching on Monday, though the Palo Alto, Calif.-based company was founded in 2011. While a variety of big data energy startups have emerged in recent months (like Bidgely, and Stem) AutoGrid has a particularly impressive team, it’s got investors that know the smart grid sector well, and it’s already got some utility customers, including the City of Palo Alto Utilities and Sacramento Municipal Utilities District.

The company’s team includes founder and CEO Amit Narayan, who was the director of Smart Grid Simulation Research at Stanford University; CTO Chris Knudsen, who was the Director of the Innovation Lab at PG&E;  and VP of Business Development Andrew Tang who also hails from PG&E. AutoGrid has raised $9 million from investors including Foundation Capital — which has backed other smart grid plays like Silver Spring Networks — as well as Voyager Capital and Stanford University, and the company is also working on projects with the Department of Energy’s ARPA-E program.

So what is AutoGrid selling? It’s created a software platform it calls the the Energy Data Platform (EDP), which takes petabytes of grid data — both structured and unstructured — and crunches it to predict and analyze what it happening on the grid in real time. AutoGrid can then provide services to its utility customers like predictive applications, optimization of the grid, or analyzing trends in energy usage. The platform is cloud based and can be delivered via private or public clouds to utilities and broadband providers (like a cable company) and requires no extra hardware.

The data that will be ingested into AutoGrid’s analytics comes from the utilities’ increasingly connected devices like smart meters on every home, and connected substations and transformers, as well as publicly available data like weather and demographics. AutoGrid’s Narayan tells me that because they don’t sell or make any hardware, they can more easily partner with all of the companies that are making these types of connected home and grid energy devices, like a Nest  smart thermotstat or a RuggedCom router for the power grid.

While the power of such a platform could be huge, AutoGrid needs to start off by selling specific applications. It’s first product is analytics for demand response, which is when utilities curb energy consumption of its customers at different peak times of day (like a hot summer day in the afternoon) to better manage the grid. AutoGrid says its demand response application can help utilities review their programs and cut the cost of their demand response systems by 90 percent and also increase the response rates of the customers in the program by 30 percent.

Both the City of Palo Alto Utilities and  Sacramento Municipal Utility District (SMUD)are using this demand response product. Utilities, which tend to be conservative when it comes to using new technologies, could be interested in AutoGrid’s tools, partly because it doesn’t require a huge commitment of time or money to test them out. The utility just needs to give them access to their data. Narayan tells me utilities can use its tools to be more innovative and to try out new services. Remember AutoGrid has two former PG&E execs in its upper management, who are familiar with the difficulties of rolling out smart grid tools from the utility’s perspective.

For the power grid to be much more efficient — and be able to include electric vehicles and clean power like solar and wind — it needs the type of real time intelligence that AutoGrid’s platform can provide. Some companies are working on similar software tools — for example, Tendril recently has shifted its focus away from hardware and more towards creating a software layer for the power grid. Stem is a startup that recently revamped and uses data and analytics — as well as energy storage — to predict and analyze the energy use of companies’ buildings. Bidgely is a new startup that is looking to crunch smart meter data to deliver information about home appliance use in real time.