On August 12, 2019, Texas electricity prices reached $9,000 per megawatt-hour. Innowatts later reported that its day-ahead demand forecast missed by 0.54%; ERCOT’s missed by 1.72%. The gap looks modest until someone must buy the electricity. A decimal point, under these conditions, has acquired the manners of a debt collector.
- Innowatts turns meter readings into forecasts and commercial decisions for energy providers.
- Its reported results are promising; its trials also expose the importance of data and delivery.
- GridX acquired it in September 2025, connecting forecasting with electricity rate analysis.
That is the useful starting point for understanding this Houston software company. Electricity suppliers have to anticipate consumption. Buy too little, and they may need expensive replacement power. Buy too much, and the surplus becomes another commercial problem. The forecast is a working assumption with money attached.
When the decimal point becomes expensive
Innowatts’ October 2019 announcement compared 32 days in July and August. It said its models beat ERCOT’s day-ahead forecast on 27 of them, with average error of 1.5% against more than 2%. These were company-reported results for a particular period, rather than a promise about every future summer.
Eric Danziger, then chief revenue officer, supplied the business explanation: “When demand soars, so does the cost of forecasting errors.” The attraction is easy to grasp. Average accuracy matters, but a buyer should also ask whether the model works during the hours when being wrong is particularly expensive.
A previous customer announcement gives that argument some commercial weight. In 2016, Direct Energy selected Innowatts’ smart-meter analytics and forecasting for its U.S. residential supply business. Direct Energy said daily and intraday forecasts had improved performance and lowered procurement costs. It expanded the partnership to carry those gains into more of its North American operations.
A meter is a customer biography
The underlying idea is to build demand from individual meters upward. A household’s electricity pattern contains clues about weather sensitivity, heating and the timing of consumption. Put enough of those patterns together, and a supplier can examine the customers inside a portfolio rather than rely exclusively on an aggregate curve.
Innowatts combines energy data with weather information and machine learning. Its current company description says its models have learned from more than 55 million smart meters worldwide. That measures the breadth of learning data; it does not mean 55 million people have bought an Innowatts subscription.
This approach becomes more useful as homes acquire rooftop solar, electric vehicles, batteries and heat pumps. Consumption and generation behind the meter complicate what the wider system sees. Innowatts offers identification of those resources, customer segmentation, weather sensitivity analysis and forecasts at different levels of aggregation. The meter’s old job was recording a bill. Its newer assignment is explaining a changing load.
Selling a forecast that reaches the trading desk
The company’s customers include retail energy providers, utilities, grid operators and large commercial or public energy users. Its present suite is organized around Connect, Extract, Forecast and Deliver: bring data in, interpret it, predict demand or generation, and support decisions such as load scheduling, risk management and grid planning.
The business is B2B SaaS, supported by integration and operational service. Its retail-provider brochure lists interfaces including APIs, cloud synchronization and file transfer, plus round-the-clock support. That detail matters. A forecast may be numerically impressive, but someone still has to receive it in the system where a purchase, schedule or alert happens.
Public records place its origins in 2013, with commercial operations beginning in 2014. The founders were Siddhartha Sachdeva, Akhlak Ahmed and Sudaksha Sachdeva. The earlier product names, PowerEASE and eUtility, reflected an ambition to personalize energy services and integrate retail operations. Today’s materials put meter-level intelligence and forecasting prominently in view.
Shell Technology Ventures led a $6 million Series A in 2017. Energy Impact Partners led an $18.2 million initial Series B closing in 2019, joined by energy-sector investors including Evergy, Shell, Iberdrola and EEI. These two announcements total $24.2 million. Financing built the business; customers still have to judge the economics of their own deployments.
The data stopped the experiment
Consider the less polished evidence. Innowatts reported strong results in an EPRI building-level forecasting trial covering June 2019 to June 2020. Nine models from six vendors were tested. Its three models led across the evaluated scenarios after missing forecast uploads were excluded. That qualification deserves to travel with the result: operational delivery belongs in a buying decision.
A forecast earns its keep when it arrives, fits the data, and changes a decision.
What the trial record suggests
A separate SSEN trial encountered a more basic obstacle. Its 2021 closedown report said the utility lacked some detailed data required to evaluate Innowatts fully. Half-hourly forecasts had to be aggregated into monthly figures for one comparison. The report recorded minor deviations and deferred further deployment at that stage. It anticipated another trial when more granular data became available.
Readers can copy the evaluation discipline: compare matching intervals, test costly peaks, count missing outputs and decide in advance who will act. An organization without suitable history or usable customer detail has work to do before a sophisticated model can be fairly tested. The algorithm cannot supply an evaluation dataset by charm alone.
GridX buys the next step
Innowatts occupies the analytics layer between energy measurements and business operations. It faces internal forecasting teams and overlapping vendors such as Bidgely, whose utility software also uses meter intelligence for customer and grid applications. Innowatts’ distinguishing emphasis is bottom-up forecasting tied to supplier economics and planning. That is a positioning argument, rather than proof of universal superiority.
In March 2025, Honeywell announced Innowatts analytics on Forge Performance+ for Utilities. In September, GridX acquired the company to add forecasting and utility planning to its Enterprise Rate Platform. The fit is practical: forecasts describe when electricity may be used; rate analysis examines what that usage will cost. Together, they put the decimal point closer to the decision that gives it a price.