I still remember watching Steve Jobs’ Stanford commencement address about connecting the dots. He spoke of dropping into a calligraphy class simply out of curiosity, a decision that years later became the foundation for the Mac’s beautiful typography. He couldn’t have known it then. Today, I have a similar feeling as I watch the energy and technology industries converge, connecting the dots between my master’s degree in energy economics and the unprecedented demand for compute that AI is driving.
This interplay is a paradox of progress: AI devours power to fuel innovation, yet it also holds the key to solving our energy crisis. AI is the caterpillar devouring today’s energy grid to transform into the butterfly that will power tomorrow’s world. As we push AI to new frontiers, we’re hitting new limits: hardware shortages, overstretched power grids, and an insatiable demand that efficiency alone can’t moderate. Microsoft’s Satya Nadella invoked Jevons Paradox to warn us: cheaper AI doesn’t shrink consumption; it expands it. The solution, therefore, isn’t just about making AI more efficient; we must fundamentally change our energy supply.
(Image courtesy SandboxAQ)
The scale of AI’s energy demand is overwhelming. Every AI-generated image comes with an energy price tag equivalent to charging a smartphone. Modestly assuming the world churns out one million images per minute, that’s 15.8 gigawatt-hours per day, enough to power a small city. Add in text, video, and scientific simulations, and the curve soars. Fatih Birol, the International Energy Agency’s executive director, said “Global electricity demand from data centers is set to more than double over the next five years, consuming as much electricity by 2030 as the whole of Japan does today.”
This bottleneck is why tech titans like Microsoft and Amazon are now investing in their own nuclear power solutions. It’s a clear signal: technology’s future hinges on energy, and we’re in a race to secure it. Look no further than the public markets for proof. In a stunning pivot, companies that once mined cryptocurrencies are now repurposing their energy-intensive infrastructure for AI compute. In the last six months, Iris Energy (IREN) has soared over 600%, while competitors have surged, with Cipher Mining (CIFR) climbing 435%, Applied Digital (APLD) 271%, and Galaxy Digital (GLXY) 205%. The market is screaming that the most valuable commodity is no longer a digital coin, but raw computational power backed by reliable energy.
Today, this immense energy demand is geared primarily towards generative AI, which creates content from language and information on the internet. But our physical world, created of atoms and molecules, demands more than language-based models. To create new physical products, we need quantitative AI. This is AI that is based on the laws of physics, chemistry, and biology to model molecular interactions, discover new drugs, advanced materials, and simulate the very hydrocarbon reactions that define our energy systems. This brings us to a fundamental truth: energy is applied science. Whether it’s chemical combustion, photovoltaic or nuclear fusion, every power source is governed by the laws of physics, the native domain of quantitative AI.
While much of the news around this technology focuses on pharma and biology with the incredible promise of curing human diseases, the same digital transformation is unlocking the future of our global energy supply. This is most evident in the world of chemistry and materials science. Think of catalysts, the unsung heroes that drive the chemical reactions for nearly every fuel we use. Over 90% of all commercially produced chemical products and fuels involve catalysts at some stage of their manufacturing process. It is estimated that catalytic processes contribute directly or indirectly to over 35% of the world’s GDP.
(Shutterstock)
The 20th century’s energy dominance was built on discovering these catalysts through slow, expensive, and often serendipitous lab work. Today, open-source chemical data like AQCAT25 and quantitative AI allow us to bypass much of that, designing and testing new catalysts in silico. This shift from physical discovery to digital design is the new strategic imperative. Energy is again the critical resource, but the leaders will be those who master the computational tools to create the very molecules that power our world.
We don’t need to wait to connect the dots anymore; AI does it for us. We are now connecting entire industries—computation, materials science and energy—in a feedback loop of innovation. This convergence is not just a technological shift; it is the defining economic and geopolitical challenge of our time. Meeting it will require us to aim our most powerful computational tools not just at creating new content, but at solving the fundamental problem of how we power our future.
About the Author
Fernando Dominguez Pinuaga is VP of Global Outreach at SandboxAQ. Pinuaga joins SandboxAQ from X, The Moonshot Factory, where he was a strategic business leader. SandboxAQ spun out of Alphabet in 2022 and is focused on creating quantum-inspired physics and AI-based tools and applications that can run on today’s classical computing platforms. SandboxAQ also focuses on enterprise SaaS solutions at the intersection of machine learning and physics. Fernando is a driven innovator with a relentless passion for projects that make the world a better place.
The post Connecting the Dots: How AI Can Solve the Energy Crisis It Created appeared first on AIwire.
