As Moore's Law slows, engineers look to new approaches and opportunities in chip innovation
Smaller, Faster, Cheaper, Over: The Future of Computer Chips — At the inaugural International Solid-State Circuits Conference held on the campus of the University of Pennsylvania in Philadelphia in 1960 … Tweets: @markoff Tweets: John Markoff / @markoff : Will Silicon Valley come to terms with the fact that the world is no longer entirely up and to the right? http://www.nytimes.com/...
Context & Ripple Effects
John Markoff's piece lands at an inflection point the trade press had been documenting for months: in early 2016 the industry's own roadmap formally abandoned its pursuit of Moore's Law as computing shifted toward mobile, breaking a planning convention that had governed chip design for decades. What followed was a scramble across three fronts — process technology, packaging, and materials.
This article frames that search as an engineering agenda rather than a crisis narrative. The subsequent coverage shows each front maturing on its own clock: leading chipmakers weighing EUV lithography for mass production, designers assembling processors from multiple small pieces of silicon called chiplets, and researchers eyeing 2D materials like graphene as eventual silicon replacements.
First-order effects
- Chipmakers lose their default scaling playbook: with the roadmap no longer committing them to Moore's Law cadence, R&D budgets shift from shrinking transistors to EUV equipment, multi-die packaging, and new architectures.
Second-order effects
- Design effort migrates toward specialized silicon — the rise of FPGAs, GPUs, and AI accelerators that Markoff later documented — because gains now come from tailoring chips to workloads rather than waiting for smaller nodes.
Third-order effects
- If the pattern holds, the industry reorganizes around heterogeneous engineering: process toolmakers (EUV), packaging innovators (chiplets), and materials labs (2D semiconductors) become the competitive axes where node-shrinking once was, with no single successor guaranteed.
The trend: The end of transistor-scaling as the industry's organizing principle is pushing semiconductor innovation into a portfolio of substitutes — lithography, chiplets, specialization, and eventually new materials — rather than one dominant path.