What do the founders of Microsoft have to do with the electricity grid?

In early 1973, the young Bill Gates, still in high school at the time, received an unexpected phone call. On the other end of the line was Bud Pembroke, an experienced IT consultant working at the time with Bonneville Power Administration (BPA), the federal body that generated and distributed electricity in the states of Washington, Oregon, and California. Among other things, it managed the massive Grand Coulee Dam on the Columbia River.

BPA had a lofty ambition: to computerise the electricity grid, switching from a manual system to one run by software capable of intervening based on the real-time needs and requirements of the market. It had entrusted this goal to TRW, a giant defence contractor, but the project was past deadline and over budget. What they needed to complete the contract was the support of specialist programmers, specifically experts in PDP-10, a powerful computer at the time. In other words, the exact professional profile of Bill Gates and Paul Allen, the two friends who would go on to found Microsoft just a few years later.

Both were then experts in PDP-10, which they had used extensively both in their school projects and in their start-up, Traf-O-Data. And so it was that, despite their youth (Gates was just 17 years old), the two received a contract with BPA. In his memoir "Source Code", the founder of the IT giant remarks on the episode: «I'm not sure how much came down to our skill versus their desperation».

To work on the project, Allen temporarily dropped out of university, while Gates received permission from his headmaster to consider the experience a form of study. So began an extraordinary adventure, both for their education and for the energy sector in the United States.

The mission: to computerise the system for greater efficiency. Until this point, balancing energy produced, and grid demand had been handled manually. In other words, operators would call each other up to report on changing demands, making the necessary adjustments to turbines and power stations as a result. BPA’s plan was to computerise these processes to allow dynamic, automated balancing, thus creating a fully-fledged smart grid before such a term even existed.

The main problem they faced was the variability of the grid: the energy supply provided by the dams fluctuated depending on the season and on rainfall, while demand changed from hour to hour, with peaks in the evening and a baseload demand at night. To be efficient, the software would need to constantly analyse data on system capacity, consumption and environmental conditions, and make real-time decisions to achieve the right balance.

Initially, Gates and Allen were tasked with writing system error messages: an apparently menial job, but one which allowed them to immerse themselves in the architecture of the software. «When we arrived at the control centre in Vancouver, it was like being on the set of Time Tunnel, only better. A wall-sized screen tracked the state of the power grid and every dam and power facility in the Northwest. Our job was to computerize that process: no more phone calls, but a system that would automatically ensure a balance between supply and demand», writes Gates. «We worked night and day, sometimes not leaving the control room for nearly a hundred hours straight. I subsisted on energy drinks and wrote code until exhaustion took over».

The turning point came when senior programmer John Norton, sent in by TRW to save the project, revised Gates’ code. «He'd torn my work apart, but he was so right. It was a crucial lesson for me, because I understood that software wasn’t a game anymore: a bug could have blacked out an entire region», Gates recalls to this day. And so the team of programmers put together by TRW succeeded in their goal, assembling the first software capable of computerising the American electricity grid.

A lesson we can still learn from. Over fifty years later, the BPA experience seems prophetic. The software that Gates and Allen helped to develop is now the beating heart of modern electricity grids. But above all, the philosophy of digital transformation that inspired it is the same one that guides investments and projects all around the world to this day. In Italy, Terna is moving in the same direction. The update to the 2024-2028 Industrial Plan outlines record investments of 17.7 billion euros, with a focus on innovation, digitalization, resilience, and the integration of renewables.

The Italian energy transition requires the management of an increasingly complex grid. By 2030, wind and solar capacity will grow from 50 to 107 GW, and, just like BPA with its hydroelectric power, Terna will have to balance variable, intermittent generation with real demand, while taking into consideration such new factors as the electrification of transport and the growth of energy communities.

To rise to this challenge, Terna is relying first and foremost on its National Control Centre and advanced solutions: digital twins, machine learning, artificial intelligence, which all have a contribution to make to the issue of climate resilience too. Extreme weather events in Italy increased by 600% between 2015 and 2024: just as Gates and Allen once did, the grid will again have to adapt while guaranteeing continuity despite fluctuations and unforeseen issues.

Innovation serving peoples’ well-being. Gates writes in his memoir: «It was the first time I understood that writing perfect code was not just an intellectual exercise. It was essential to keeping the lights on». «Keeping the lights on», sustainably and at low cost, is the goal, now more than ever, of the digital revolution: the very same revolution sparked by two talented kids in a control room in the early 70s.

But energy is a subject that remains important to Bill Gates even in more recent times. During last year’s Breakthrough Energy Summit in London, Gates emphasised how AI can help identify new ways to cut consumption and drive the transition towards clean energy, while recognising that the pace of green energy development might not yet be sufficient for the zero-emissions goals. In fact, the founder of Microsoft believes that, despite AI’s high energy consumption, this technology will prompt an overall reduction of more than 6% in energy consumption, and that the increase in energy demand from artificial intelligence is not a threat, but an opportunity for the future of sustainable innovation.