Our latest in a series of blogs inspired by Applied Materials’ panel discussion at IEDM explores where and how data should be processed and stored to drive computing efficiency while curbing energy consumption.
Applied Materials moderated a thought-provoking panel discussion during IEDM which showed that while there is no single path to achieving continued improvements in chip performance, power and area-cost, the industry will be well-served to search for solutions together.
What’s clear from the panel discussion I recently moderated with Facebook, IBM, Intel, Stanford and TSMC is that the semiconductor design and manufacturing model is evolving and will look extremely different in the years ahead.
Where can the industry test codesigned chip innovations for the AI Era—to accelerate the future of computing, from materials to systems? The answer is: our new Materials Engineering Technology Accelerator (META Center), which—I’m delighted to announce—is now open for collaboration.
The serial interaction between siloed parts of the industry that produced the computing eras of the past—mainframe/minicomputer, PC/Internet, mobile/social media—will not be sufficient to fuel the next era of computing. Instead, we need codesign from materials to systems to enable the AI Era.
The AI Era is fueling an explosion of data, while Moore’s Law is no longer the predictable engine of progress that it once was. With opportunities and challenges of this magnitude, we need to open the strategic planning lens and decide where we want to be in 2029.
Metrology allows process engineers to see the results of process steps. Today, driving improvements in chip performance, power and area/cost (PPAC) requires novel architectures and exotic material stacks. This is introducing new sources of atomic-scale variation that can negatively impact chip performance and yield. It is becoming imperative to “watch” the deposition process as it occurs to control variability and deliver repeatable performance.
