The concept of storing energy in sand has an almost poetic quality. It is ubiquitous, old, and unglamorous. The amount of material beneath your feet is so abundant that it almost seems offensive when you walk on any beach or in any desert.
Google, a business that has never exactly struggled with ambition, is now discreetly testing the use of sand as a thermal battery, storing excess renewable energy as heat and releasing it back into grids when the sun sets or the wind stops.
| Category | Details |
|---|---|
| Company | Google LLC (Alphabet Inc. subsidiary) |
| Founded | 1998 — Menlo Park, California |
| CEO | Sundar Pichai |
| Headquarters | Mountain View, California, USA |
| Core Focus | Search, Cloud, AI, Renewable Energy Research |
| Energy Initiative | Sand Battery Thermal Storage Experiments (2024–2025) |
| AI Ethics Framework | Published 2018; revised 2024–2025 |
| Project Maven Controversy | US Dept. of Defense AI drone footage analysis contract |
| Known AI Product | Gemini 2.5 (launched March 25, 2025 without full model card) |
| Workforce | ~180,000+ employees globally |
| Annual Revenue | Approx. $307 billion (2024) |
| Ethical Criticism Areas | Transparency, bias in AI, data privacy, labor concerns |
The idea itself is not wholly novel. Since 2022, the Finnish startup Polar Night Energy has operated a sand battery system in Tampere, and engineers who spend their days considering grid stability have shown genuine interest in the technology. Sand is inexpensive. Lithium degrades more quickly than sand. The sourcing disputes surrounding cobalt mining in the Democratic Republic of the Congo do not affect sand. It sounds almost too tidy on paper.
However, Google is not a Finnish company. It is one of the most closely watched companies on the planet, operating under intense scrutiny each time its name is associated with terms like “experiment” or “ethical questions.” Furthermore, considering the company’s recent performance, the scrutiny is not wholly unjust.
It was like a door slamming in a quiet room when Google unveiled Gemini 2.5 in late March 2025 without a complete model card, which is the technical documentation that tells researchers and regulators what a system can and cannot safely do. Depending on your point of view, the six-page model card that finally surfaced weeks later is either a start or a gesture.
It’s difficult to ignore the structural tension that underlies both Google’s energy experiments and its AI disclosures: the company moves quickly, publishes selectively, and then asks the public to believe that its internal reviews were thorough enough. The Gemini rollout was likened by Professor Sandra Wachter to shipping aircraft without safety manuals. The analogy endured because it was true in its ridiculousness.
This larger pattern includes sand batteries. Google is testing a thermal storage technology that heats sand to temperatures above 500 degrees Celsius using excess electricity, preferably from solar or wind. After being stored for days or even weeks, that heat can either be used directly for industrial heat processes or transformed back into electricity.
The consequences are significant in areas that rely on fossil fuels for winter heating. However, Google’s plans for scaling, partnerships, data collection methods regarding grid usage, and specific implementation details are still mainly unknown.
It may not seem like it, but that opacity is important. As they say, “personal information is the new gold,” and energy systems produce enormous amounts of behavioral data. Advertisers, insurers, real estate developers, and governments all benefit from knowing how a grid draws power, when demand spikes, and which neighborhoods consume more and less.
What precisely does Google intend to do with the data it gathers from energy infrastructure it helps power or monitor? This is a legitimate question buried in the sand battery story, but it hasn’t been raised loudly enough.
Google pledges to incorporate privacy design principles and be accountable to people in its AI principles, which were published in 2018 and have since been updated with differing degrees of enthusiasm. It’s a promise made with the kind of assurance that sounds comforting in a press release but seems a little flimsy when scrutinized.
Early in 2025, watchdog organizations discovered that Google had subtly removed language related to diversity and equity from some of its Responsible AI pages. This kind of minor change often reveals more about a company’s direction than any public declaration.
Employees at Google have not always watched the company’s ethical development in a passive manner. Workers publicly protested Project Maven, the Department of Defense contract that used Google’s AI to analyze drone footage, a number of years ago.
Executives were reportedly so shaken by the backlash that CEO Sundar Pichai admitted the field was “dynamic and evolving.” That is undoubtedly one way to put it. Observers often disagree sharply about whether that evolution is headed in the right direction.
If used widely, sand batteries might be exactly what their supporters say they are: a low-tech, low-toxicity link between intermittent renewable energy sources and dependable power delivery. The physics make sense. The resources are plentiful. Making a sand battery has a far smaller environmental impact than making lithium-ion batteries. At a time when the global supply chains for rare minerals are being strained by the energy transition, that is truly valuable.
In another version of this story, however, Google’s interest in thermal energy storage is more about vertical integration—managing infrastructure, gathering data, and extending its presence into actual systems that billions of people rely on on a daily basis—than it is about decarbonization. The maps we use, the cloud that stores our documents, and the search engine that determines what we read are all already created by the company. It is not neutral to include energy infrastructure on that list.
Actually, sand has nothing to do with the deeper ethical dilemma. It concerns who is allowed to conduct experiments on systems that are visible to the public, how transparent those experiments must be, and what accountability entails when the experimenter is one of the most influential technology companies in human history.
Autonomous technology carries risks that aren’t always apparent until something goes wrong, whether it’s an AI system determining how and where stored energy flows through a grid or a self-driving car making snap decisions.
As this develops from the outside, there is a sense that the discussion the world needs to have about Google’s energy goals is the same one it has been having about its AI practices: how much trust is reasonable to give to a private company running infrastructure that impacts everyone, and what safeguards are in place in case that trust proves to be misplaced? An intriguing solution to a real issue is sand batteries. The issue that needs a lot more attention is who controls them and under what conditions.
