Instead of relying on in-demand standard chips, some of the world’s largest technology companies are developing their own semiconductors.
According to company announcements and media reports, Apple, Amazon, Facebook, Tesla, and Baidu are all avoiding established chip firms and bringing certain aspects of chip development in-house.
“Increasingly, these companies want custom-made chips that meet the specific requirements of their applications rather than using the same generic chips as their competitors,” said Syed Alam, Accenture’s global semiconductor lead.
“This gives them more control over the integration of software and hardware while distinguishing them from their competitors,” Alam continues. According to Russ Shaw, a former non-executive director at Dialog Semiconductor in the United Kingdom, custom-designed chips can perform better and be less expensive.
“These specially designed chips can help to reduce energy consumption for devices and products from the specific tech company, whether it’s smartphones or cloud services,” Shaw explained.
The ongoing global chip shortage is another reason why big tech firms are reconsidering where they get their chips, according to Glenn O’Donnell, research director at analyst firm Forrester. “The pandemic threw a huge wrench in these supply chains, hastening efforts to do their own chips.”
“Many already felt constrained in their innovation pace due to chipmaker timelines,” O’Donnell explained.
At the moment, there isn’t a month that goes by without a major technology company announcing a new chip project.
The most notable example occurred in November 2020, when Apple announced that it was abandoning Intel’s x86 architecture in favor of developing its own M1 processor, which is now found in new iMacs and iPads.
Tesla recently announced the development of a “Dojo” chip for training artificial intelligence networks in data centers. In 2019, the automaker began producing cars with its own AI chips, which aid on-board software in making decisions based on what’s happening on the road. Baidu launched an AI chip last month that is intended to help devices process massive amounts of data and boost computing power. The “Kunlun 2” chip, according to Baidu, can be used in areas such as autonomous driving and has entered mass production.
Certain semiconductor projects have been kept under wraps by some of the world’s most powerful companies.
Google is said to be getting closer to releasing its own central processing units, or CPUs, for its Chromebook laptops. According to a report published on September 1 by Nikkei Asia, the search giant intends to use its CPUs in Chromebooks and tablets that run the company’s Chrome operating system beginning in 2023. Amazon, the world’s largest cloud service provider, is developing its own networking chip to power hardware switches that move data around networks. If it works, Amazon’s reliance on Broadcom will be reduced.
In 2019, Facebook’s chief AI scientist told reporters that the company is developing a new class of semiconductor that would operate “very differently” than most existing designs.
At this point, none of the tech behemoths intend to do all of the chip development themselves.
“It’s all about the chip’s design and performance,” Shaw explained. “At this point, it’s not about manufacturing and foundries, which are very expensive.” A modern chip factory, or foundry, such as TSMC’s in Taiwan, costs around $10 billion and takes several years to build.
“Even Google and Apple are hesitant to build these,” said O’Donnell. “They’ll build their chips at TSMC or even Intel.”
According to O’Donnell, there is a scarcity of people in Silicon Valley with the skills required to design high-end processors. “Over the last few decades, Silicon Valley has placed so much emphasis on software that hardware engineering was seen as a bit of an anachronism,” he explained.
“Doing hardware became ‘uncool,'” O’Donnell explained. “Despite its name, Silicon Valley now employs a small number of true silicon engineers.”