The switch to renewables is about energy security and climate change.

While climate change is obviously the big-picture issue in the global drive to decarbonize, energy security is an important aspect that is shaping governments’ policies worldwide. From either perspective – strategic or environmental – Europe and the US are far behind China in the race to adopt renewables. China reported installing 198GW of solar capacity between January and May. Effectively, that’s the equivalent of adding 100 panels every second!

With total capacity now more than twice that of the rest of the world combined, China is clearly the prime mover here and looks likely to reach its 2030 climate goals by next year. Now compare that to Europe, which is expected to add 110GW this year while the US is at just 36GW.

It’s true that progress in the West is slowed by factors like high real estate costs, lengthy planning procedures and other legal hurdles. On the other hand, 69% of all new energy capacity installed is solar, so change is happening – if slowly. Solar has become hugely successful globally, with total capacity reaching 2.2 terawatts in 2024. That’s already more than four times the 2035 target set by the International Energy Agency, although global demand and energy from fossil fuels are also still rising.

Despite that caveat, we can see that solar will satisfy a significant proportion of the world’s future energy needs. This was accomplished through a combination of improved photovoltaic energy-conversion efficiency and increased panel manufacturing efficiency. The next challenges concern managing capacity and ensuring stability as we continue to make our grids greener. The solutions are software-based.

Balancing the energy flows to maintain an adequate supply and ensure appropriate demand response is a fascinating area. This has been a challenge, of course, since the beginning of scalable power distribution, which provided the foundation for electricity to evolve from a utility to a strategic enabler of modern life. The traditional distribution model has enshrined a centralized architecture containing a small number of large power generators operating continuously. Green grids are the polar opposite, comprising many smaller generators driven by intermittent sources.

Renewable generation is noisy and nonlinear, so managing green grids requires greater perception and more responsive control than has been necessary in the past. AI is central to this, enabling systems to ingest vast quantities of data from diverse sources, detect patterns invisible to rule-based systems and connect causes and effects across domains. These include weather forecasts, solar irradiance, wind speed, grid voltage and frequency, and usage data from EV charging stations, smart meters and industrial activity. In addition, AI agents make it possible to orchestrate adaptive responses in real time, optimizing inverter settings, reconfiguring microgrid topologies, charging and discharging storage across thousands of nodes in milliseconds. And let’s not forget AI’s ability to learn from past events, which can help automatically improve responses and guide decisions on future planning and upgrades.

One problem is that we lack the granular, multimodal data – such as microclimate data for solar forecasting and behavioral data for demand shaping – needed to train AI for grid management. Historically, there has been no compelling reason for conventional operators to collect such information, and, in any case, the data would have overwhelmed conventional analytics. Now, however, the need is clear, and we have systems capable of processing such diverse data streams meaningfully.

Although the prospects for solar farms are bright, there will always be limiting factors such as cloud cover, seasonal changes, and the cycles of day and night. These problems go away in space, of course. Building an “always on” solar generator to beam power to earth is a scientist’s dream that could become a reality if Japan’s OHISAMA (Japanese for Sun) project is successful. Aiming to demonstrate a 1kW generator in space, the experimental satellite weighs about 400 lbs. and will capture and convert solar energy for transmission to earth via a microwave link. If successful, it’s reckoned the project could be scaled up to 1GW or more per satellite.

OHISAMA is not the only out-of-this-world proposal in the solar industry. A Californian company, Reflect Orbital, is building a demonstrator to test its concept to increase solar generating capacity by providing extra daylight during peak power demand. The plan suggests using reflectors aboard a constellation ultimately comprising more than 4000 satellites to extend daylight hours after sunset and before sunrise. There are concerns about how this could affect all forms of life on earth, as well as the potential effects on human psychology.

The scale of any space-based building project large enough to make a difference to power generation on earth is also considerable. The US’s Golden Dome defense project, which aims to build enormous constellations to detect and destroy hypersonic missiles, identifies power generation as a key challenge. Among the many challenging aspects of this project – like the high-speed communications and complex handoffs needed to track missiles traveling at up to 20 times the speed of sound – published analysis suggests suitable solar generators could be produced at a rate of one panel a day, amounting to one wing per week.

While these projects offer some fascinating technological development opportunities, here's a sobering thought: in the five minutes you may have spent reading this column, China’s solar farms could have grown by the equivalent of 4000 satellite wings, or one for every space vehicle in the planned Reflect Orbital constellation. That’s bringing power down to earth.

Alun Morgan is technology ambassador at Ventec International Group (venteclaminates.com); alun.morgan@ventec-europe.com. His column runs monthly.

• Next