Advanced Computing in the Age of AI|Tuesday, October 27, 2020
  • Subscribe to EnterpriseAI Weekly Updates:  Subscribe by email

Globalfoundries Sees Pandemic ‘Pulling’ Chip Demand 

Globalfoundries, the custom semiconductor manufacturer, announced advancements on several fronts this week, including the latest version of its FDX platform aimed at low-power, edge devices as well as a chip design verification kit that incorporates machine learning tools.

The foundry also unveiled chip intellectual property protections aimed at safeguarding customer data. Also announced was glucose monitoring technology based on its 22FDX platform.

In previewing the upgrades and designs, CEO Tom Caulfield noted the pandemic has reversed the traditional “push” of chip technology to enterprise customers. Instead, Caulfield asserted, the pandemic in pulling technology to new applications.

“We want ‘pull’ technology to come faster,” he said during a briefing with analysts and reporters. “We’re seeing accelerated demand. It’s not a bubble, it’s an acceleration.”

GlobalFoundries dropped its 7-nm process development in 2018, redeploying its FinFET technologies mostly to the 12-nm node. Caulfield downplayed the drive to finer chip geometries, noting there is plenty of volume at higher nodes. “Twelve nanometers and above is the sand we play in,” he added.

Those efforts have yielded a “plus” version of the foundry’s 22FDX platform aimed at Internet of Things and 5G mobility applications. The new version unveiled during a virtual event is based on 22-nm FD-SOI process technology, or fully depleted silicon-on-insulator.

The upgrade addresses ultra-low power RF applications such as connected devices. GF said it has so far shipped 350 million 22FDX chips to global customers.

The upgraded version includes digital and RF enhancements designed to boost frontend module designs, an emerging “pull” market for 5G applications.

The 22FDX+ chip is manufactured at Globalfoundries’ 300-mm production line in Dresden, Germany. The company said it would be available in the first quarter of 2021.

Meanwhile, the chip maker and IC design tool vendor Mentor released a “design for manufacturability” (DFM) kit that embedds machine learning capabilities. The kit is based on the Siemens unit’s Calibre design platform, and targets AI training and inference applications.

DFM technology is intended to help chip designers spot defects that could show up in production devices. The tool is ready for production at Globalfoundries’ Fab 8 in Malta, NY. It will be rolled out on its 22FDX and 12LP+ platforms during the fourth quarter of 2020.

As the U.S. military strives to secure and diversify its semiconductor supply chains, greater emphasis is being placed on baking securing into chip designs. With that in mind, GlobalFoundries also this week unveiled a “Shield” program aimed at safeguarding chip IP and the resulting products. The goal is to extend security steps used in military and aerospace components to commercial customers.

Along with cybersecurity, the framework would integrate information, product and operational security, the company said.

Healthcare is emerging as another technology puller. To that end, GlobalFoundries and medical device startup Movano Inc. rolled out a continuous glucose monitoring technology. The “needle-free” monitor is based on the chip maker’s FD22 platform and Movano’s proprietary RF sensor technology.

The glucose monitor employs machine learning, ultra-wideband, multiple antenna RF technology and operates in the cloud. The RF sensor is manufactured on Globalfoundries’ Dresden production line.

Caulfield also this week reaffirmed the chip foundry’s plan to go public by the second half of 2022.

About the author: George Leopold

George Leopold has written about science and technology for more than 30 years, focusing on electronics and aerospace technology. He previously served as executive editor of Electronic Engineering Times. Leopold is the author of "Calculated Risk: The Supersonic Life and Times of Gus Grissom" (Purdue University Press, 2016).

Add a Comment

Do NOT follow this link or you will be banned from the site!
Share This