What is Moore's Law, and is it really ending?

Moore's Law is the observation that the number of transistors on an integrated circuit doubles about every two years with minimal rise in cost. It is not a physical law. Its original form, the pace of transistor doubling, has clearly slowed, and several sources say shrinking transistors is nearing physical limits. Whether the law is therefore over depends on whether it is read as a literal doubling schedule or as a broader goal of steady computing progress. [1] [2]

Intel co-founder Gordon Moore, in a 1965 paper, predicted that transistor counts would double every year for the following ten years. In 1975 he revised the figure to a doubling every two years. According to Intel, Moore did not much like the label "law", which was not in his original paper. It is a projection that depends on continued innovation. Wikipedia describes it as an empirical relationship of the experience-curve type, which quantifies efficiency gains from learned experience in production. [1] [2]

Intel says the observation became a target that the company and its competitors worked to meet, which is why it guided the semiconductor industry for close to 60 years. [1]

An OECD essay by H. Kressel says gate widths now approach near-atomic dimensions, and below a certain width the switching properties of the conventional transistor deteriorate. Patterning such small features needs special ultraviolet light sources, multiple exposures and extraordinary equipment control, which greatly raises cost. As a result, unit transistor costs have started to rise rather than fall, so the most advanced chips are justified only for demanding, high-volume uses such as smartphones and large cloud systems. Live Science adds that at the smallest scales quantum effects such as tunneling cause electrons to leak. [3] [4]

One view holds that the law has effectively ended in its literal sense. Live Science says that in the simplest sense it is no longer relevant, because the two-year doubling is far behind us, though it lives on as an expectation of the pace of progress. The OECD essay likewise says industry sources expect transistor density to double over the next ten years, not two. [4] [3]

A second view is that the law continues through other means. Intel says advanced packaging and new materials keep its pace going, and notes that many have predicted its death before. A third view reframes the law. Georgia Tech's Arijit Raychowdhury is quoted as saying it is fundamentally economic, about the business imperative to deliver better performance, lower power, smaller form factors or reduced costs, not about the physics of smaller transistors. [1] [5]

Wikipedia notes that industry experts have not reached a consensus on when the law will cease to apply. The sources here agree that transistor shrinking has slowed, and differ mainly on what counts as the law. [2]

The OECD essay points to three-dimensional chip structures that use chip area better, with industry sources expecting them to double density over the next decade, though new fabrication plants cost more than USD 10 billion each. It also cites work on removing system bottlenecks. Live Science notes the shift toward chips with many cores that split tasks. [3] [4]