Innovation Lightbulb: The Geopolitics of Moore’s Law

Innovation Lightbulb: The Geopolitics of Moore's Law

What exactly are we looking at here? How can policy makers better understand the message and implications of this graph? 

This graph, well-known among chip industry insiders, illustrates a trend concerning the building blocks of electronics: the micro-switches—or gates—that combine to create complex circuits in a computer or device.

In essence, more gates/transistors = more computing power.

When the first transistor was invented at Bell Labs in 1947, it was the size of a desk ornament. Over the decades, scientists and engineers have shrunk them to the nano-scale, with the most cutting-edge chips on the market today boasting 300 million transistors per square millimeter of chip.

The density of transistors on a single chip has historically doubled every two years, while the cost-per-transistor has halved. This empirical observation was made by co-founder of Intel Gordon Moore in 1965, and has since been dubbed “Moore’s Law.” The persistence of this trend has meant an explosion in computing power, the reduction of its cost, and the ubiquity of chips across industries over time. This explosion has enabled innovation in technologies which make us more productive, more connected, more comfortable, healthier, and safer.

However, chip density- and cost-scaling have slowed in the last decade, with cost-scaling even reversing towards more advanced nodes.* The chart above shows that after scaling hit 28nm (a process first commercialized in 2011), cost-reduction stopped due to the extreme complexity and number of process steps required to make an advanced chip.

As a result, designing and manufacturing a chip at the cutting-edge has become increasingly capital intensive and cost-prohibitive for most firms. This is amplified by the winner-take-all nature of the chip industry: if you make the best product, you capture an overwhelming majority of the revenue which can be reinvested; if you fall behind, good luck catching up.

From a geopolitical and policy standpoint, this means that a handful of firms—and the countries in which they reside—dominate entire segments of the semiconductor supply chain, on account of their technological pedigree developed from decades of capital investment and government support. 

China has committed to closing the gap between itself and the rest of the world through a strategy of massive state investment in all segments of its domestic chip industry and state-coordinated, large-scale industrial espionage.

While not a strategy worth replicating, this adds more geopolitical and national security weight to the health of the U.S. semiconductor industry and the innovation ecosystem which support it.

Without a long-term commitment to investing in both our leadership in STEM and our advanced manufacturing capacity, the U.S. risks becoming increasingly dependent on China for the technologies that power the modern economy.

*The term “node” roughly refers to a chip’s complexity and scale. While now used more as a marketing term rather than as an accurate reflection of scale, it denotes the size at which chips are mass-manufactured.

Data visualization by Jaehyun Han