Marc Andreessen famously said, “software is eating the world.” So why hasn’t it devoured the energy industry yet?
I’ve spent much of my career in energy trading, witnessing firsthand how powerful market signals and software-driven automation can be in reshaping energy systems. Over time, I've concluded we're severely underestimating the potential of software—not just to complement but sometimes entirely replace traditional hardware-based energy solutions.
Seeing the bigger picture in energy markets
In my trading days, I watched electricity prices spike on hot days, quickly followed by demand dropping—not through new infrastructure but because smart software automatically adjusted consumption based on price signals. Rather than investing billions into new power plants that might only run a few days each year, digital intelligence optimised existing assets, managing supply and demand in real-time.
The International Energy Agency reinforces this, noting that smart grid technologies can reduce costly infrastructure needs by intelligently balancing supply and demand.
Why isn’t this better understood?
Software remains overlooked primarily because it’s intangible compared to massive wind farms or sleek battery factories. Policies and investments have typically favoured large hardware projects, partly because they're easier to visualise and politically attractive.
However, we're entering a new era where advanced software—particularly AI-driven solutions—can significantly shift this balance.
AI agents: Transforming energy management
AI agents, capable of learning and adapting in real-time, can orchestrate energy systems "over the top" of legacy infrastructure. Instead of waiting years for grid upgrades, these agents use existing data streams to instantly optimise resources.
This isn’t hypothetical: Google’s DeepMind has collaborated with the UK’s National Grid, significantly reducing wasted standby power through AI forecasting. Similarly, Tesla’s virtual power plant (VPP) in Australia showed how AI-managed home batteries could instantly replace a coal plant outage, stabilising the grid faster and more cost-effectively than traditional methods.
Strategic opportunities with software
- Dynamic demand and supply: Software-based demand response programmes automatically shift energy use to off-peak times, significantly reducing the need for additional power plants.
- Virtual power plants (VPPs): The U.S. Department of Energy projects that VPPs could avoid billions of dollars in infrastructure investment by harnessing distributed energy resources intelligently.
- Automated energy trading: AI-driven trading systems continuously optimise energy assets, increasing the economic value of renewables. Google boosted the economic value of its wind farms by roughly 20% using AI forecasting.
Overcoming the roadblocks
Despite clear benefits, adoption has been slow due to policy inertia, risk aversion, and legacy systems locking away valuable data. However, AI’s ability to integrate disparate systems presents a practical solution to these challenges.
The road ahead
While we undoubtedly need more wind turbines and solar installations, the critical role of software and AI-driven solutions must not be overlooked. These technologies represent powerful accelerators in energy efficiency and flexibility, delivering faster, cheaper, and more adaptable decarbonisation paths.
As someone deeply familiar with both energy markets and technology, I'm optimistic that software's full potential in the energy transition is just beginning to emerge. If we acknowledge and invest in software with the same enthusiasm as hardware, we can unlock new opportunities and more rapidly achieve our energy goals.