IBM has unveiled a radical new chip architecture that stacks transistors vertically like a block of flats, promising a leap in performance and energy efficiency. This British-style engineering triumph could redefine the limits of Moore’s Law, but its impact on global energy consumption and climate goals is a double-edged sword.
The semiconductor industry has long relied on shrinking transistors horizontally to cram more power into chips. But as we approach the physical limits of silicon, IBM’s 3D stacking approach offers a vertical escape route. By building layers of transistors one atop another, the chip can house more components without increasing its footprint, much like a high-density housing estate. This reduces the distance data must travel, slashing power consumption and heat generation.
For a climate correspondent, this is both a beacon and a warning. Our digital infrastructure already accounts for nearly 10% of global electricity demand, and AI workloads are driving that figure toward 20%. Any efficiency gain is a step toward decarbonisation. But the Jevons paradox looms: more efficient chips often lead to more use, not less. The energy saved could be swallowed by the next generation of data-hungry AI models.
IBM’s chip is not a silver bullet. The manufacturing process is complex and energy-intensive in its own right. Building a 12-storey block of flats requires more concrete than a bungalow. Similarly, 3D-stacked chips require intricate cooling systems to manage heat in the vertical stack. IBM has used tiny channels of fluid to carry heat away, but the overall carbon footprint of production remains high.
Yet the potential for climate-positive applications is significant. Imagine edge devices that can process climate models locally, power grid sensors that optimise renewable energy flow, or AI that monitors ice sheet melt in real time with minimal energy draw. This chip could be the engine for the next wave of green technology.
The question is whether we will use it wisely. History suggests we will not. Every breakthrough in computing efficiency has been met with a corresponding increase in demand. The cryptocurrency boom swallowed entire power plants. Large language models now guzzle electricity like a fleet of SUVs. IBM’s block-of-flats chip could either be a tool for sustainable progress or a catalyst for even greater energy profligacy.
Climate scientists have long warned that technological efficiency must be paired with policy and behavioural change. A more efficient engine still pollutes if you drive it farther. A more efficient chip still heats the planet if we use it to mine more Bitcoin or train larger AI models.
IBM’s announcement is a milestone. It showcases the meticulous engineering that has long defined British and European semiconductor research. But it is not a substitute for systemic change. We must marvel at the science while demanding that this power be directed toward solving our most pressing crisis: the biosphere collapse.
In the coming years, we will see a flood of similar announcements from firms racing to postpone the end of Moore’s Law. Each will promise efficiency, performance, and a salve for climate guilt. But the only metric that truly matters is whether global absolute energy consumption from computing declines. Without binding international agreements to cap data centre energy use, these chips will merely accelerate our digital addiction.
For now, I will report the facts with calm urgency. The chip is real. The engineering is beautiful. The potential is immense. But the path to a stable climate does not run through a single silicon block. It runs through our collective will to use these tools for survival, not convenience.
