Solargis says solar’s next challenge is not just hardware, but data

As PV projects move into more complex terrain, hybrid configurations and grid-supporting roles, Solargis sees higher-resolution meteorological data, physical modelling and quality-controlled AI as increasingly central to project development and operation.

Solar power plants are no longer just physical infrastructure. They are becoming digital assets that need to be designed, financed, monitored and operated with increasingly sophisticated data and software tools.

That was the central message from Marcel Suri, CEO of Solargis, in an interview during SNEC 2026 in Shanghai. Solargis, he said, is a digital solutions company providing software and meteorological data for solar project developers and asset operators. Its role in the industry is built around a simple but increasingly important reality: the performance of a PV project is shaped not only by modules, inverters and trackers, but also by solar irradiation, temperature, wind, snow, pollution, cloud movement and site-specific weather conditions.

“Solar power plants are digital assets,” Suri said in substance. They need to be developed and controlled through software tools and meteorological data because the external environment directly affects energy production and system performance.

For much of the early solar market, project modelling was largely built around accurate solar irradiation data and typical-year assumptions. Developers and lenders wanted to know the expected resource level and the long-term average yield. That is no longer enough. According to Suri, customers today are much more demanding about data quality, resolution and coverage. They still need accurate solar radiation data, but they also increasingly require reliable information on wind, temperature, snow and other meteorological parameters.

The reason is that solar projects are under more stringent technical and financial scrutiny. Investors want confidence that a plant will meet its expected performance. Owners need to understand how a system behaves in operation, not only under average conditions but also during unusual or extreme weather events. A project that looks robust in simplified modelling may show materially different results when complex terrain, tracker behavior, bifacial modules, fast-moving clouds or hybrid PV-plus-storage configurations are properly simulated.

This is changing demand for both data and software. Solargis sees growing interest in higher-resolution time-series data, moving from hourly datasets to 15-minute data, and in some project-specific cases even one-minute resolution. Such data can help model inverter clipping losses more accurately, simulate system behavior under fast-changing cloud conditions, and support operational decisions in projects that include batteries or grid-supporting functions.

Suri said the market has also moved beyond the idea that a typical meteorological year is sufficient for project assessment. Developers now need to understand extreme weather conditions — including strong wind, heavy snow, pollution and other events that can affect safety, availability and long-term output. If such extremes are not properly understood, he said, a project may be exposed to technical risk or even failure.

This need becomes more important as solar plants are built in more challenging locations. Many new projects are moving into mountains, deserts, snowy regions and areas with complex terrain. At the same time, project configurations are becoming more advanced. Hybrid plants, trackers, bifacial modules and storage systems all increase the number of variables that must be modelled. For such projects, simplified calculations can diverge significantly from detailed physical simulations.

These trends are particularly relevant for Chinese solar companies expanding overseas. Suri said the challenges they face are broadly similar to those faced by developers in other markets, but the scale of China’s solar industry makes the issue especially significant. As Chinese companies develop or supply projects in more diverse international environments, they need to align historical meteorological data with future performance forecasts, understand local site conditions, and prepare designs that can work with local grids.

The grid is becoming a central part of the discussion. Solar projects are increasingly expected not only to generate electricity, but also to interact intelligently with the power system. In PV-plus-storage projects, developers need to understand how batteries should charge and discharge, how curtailment should be managed, and how the project can support grid needs. This requires high-frequency data and software capable of modelling the physical behavior of the system from the string and inverter level through to battery dispatch and power injection into the grid.

Suri noted that more manufacturers are now talking about smart systems, grid-forming capabilities and grid-supporting functions. But those concepts cannot be realized through hardware alone. They require good data, advanced software and project designs optimized for grid interaction.

Forecasting is another area where digital tools will become more important. Solar forecasting is not new, including in China. But Suri expects it to become more agile and more closely linked to operations. Forecasting will need to respond not only to weather, but also to grid demand, storage dispatch and real-time system behavior. In that sense, the next phase of solar digitalization will require much closer alignment between software, data and hardware.

Artificial intelligence will play a role in this transition, but Suri’s view is deliberately measured. AI can help fill gaps in data analysis and support more advanced interpretation of measurements. It can be useful when dealing with complex datasets, including meteorological measurements and electrical signals from PV systems. But he argued that AI should not replace physical models where physical processes can be mathematically described.

Where solar radiation, meteorological data and electrical output can be converted through established physical equations, physical models should remain the foundation. AI and machine learning become more valuable when they are combined with those models, rather than being used as a substitute for them.

This distinction matters because AI is only as reliable as the data it uses. Suri stressed the importance of quality control. Measurements from weather systems or SCADA platforms are often affected by failures, errors or inconsistencies. Before such data can be analyzed by sophisticated algorithms, it must be cleaned, checked and properly structured. Otherwise, AI may amplify poor data rather than produce useful insight.

For Solargis, the future therefore lies not in treating AI as a universal answer, but in building a stronger ecosystem around data quality, physical modelling, forecasting and selective AI application. The company’s view is that digital solutions must become more deeply embedded in solar project development and operation as PV assets become larger, more complex and more closely connected to the grid.

This marks a broader shift in the solar industry. Hardware innovation remains essential, but it is no longer sufficient. As projects expand into more difficult environments and take on more system-level responsibilities, the ability to model, predict and control performance becomes a major source of value.

Suri’s message at SNEC 2026 was clear: the next stage of solar development will depend not only on better modules or batteries, but also on better data. In a market defined by hybrid systems, extreme weather, grid constraints and rising investor scrutiny, digital intelligence is becoming part of the infrastructure itself.