When the chips are down

When the chips are down

15 October 2021 | Natalie Bannerman

Cover

As the world continues to cope with the global semiconductor shortage, Capacity’s Natalie Bannerman explores the effect of this on the telecommunications industry

The perfect storm created by the Covid-19 pandemic and The China-US trade war has placed restrictions on China’s Semiconductor Manufacturing International Corporation (SMIC) and made it harder for it to sell to companies with American connections. Along with other contributing factors, this has led to what is now one of the biggest global shortages in semiconductor chips.

“As the world shut down because of the Covid-19 pandemic, many factories closed – making the supplies needed for chip manufacturing unavailable for months,” explains Biju Nair, president of global connected living at Assurant.

As a result, orders began to surge and manufacturers struggled to keep up with accelerated demand, creating a backlog, and “an infinite loop, where supply struggled to meet demand”, adds Nair.

“This was compounded by a labour shortage, which not only impacted the mining for minerals needed to make semiconductors, but also the ability to process and transport them.”

“The positive elements of this storm include the acceleration of digitisation by years, new technologies coming to fruition sooner than anticipated and a V-shaped economic recovery,” says Jodi Shelton, CEO of the Global Semiconductor Alliance (GSA).

Affecting everything from the automotive industry to video games and general electrical devices, this shortage has been having a pronounced impact on the telecoms industry. Beyond their use in mobile phones and handheld devices, semiconductor chips are required for core networking, across passive infrastructure, IoT sensors, routers, wireless communications and everything else in between.

“For the telecoms industry, chip shortage impacts are translating into cost increases across the board,” says Yatin Trivedi, AVP for systems and semiconductors at Capgemini Engineering.

“These increases will be passed on first to their enterprise customers and eventually to consumers. For telcos, the shortage will slow down both deployment of the new 400GB (wired) technologies and 5G roll-out, including densification, by 18-24 months. Beyond this timeline, it is hard to predict, since the duration of the chip shortage is still unclear.”

We have already begun to see the impact of this across the sector. September saw Reliance Industries, through its subsidiary Jio Platforms, delay the launch of its JioPhone Next smartphone, a project developed in partnership with Google, citing the semiconductor shortage as a reason.

May saw Italy’s Telecom Italia (TIM) address the shortage, with CEO Luigi Gubitosi commenting during an analyst conference call, “We are marginally starting to see some shortage on chips around the market,” adding that the company had purchased modems in advance that will enable its customers to watch the upcoming season of football in a deal with sports streaming service DAZN “to avoid a supply chain issue”.

While in May 2021, during the JP Morgan Conference, John Stankey, CEO of AT&T, said: “If there is anything that I kind of sit back in and worry about a little bit, and I’ve mentioned it before, supply chains are stressed in a way I have never seen.”

“And it doesn’t matter whether you need pool equipment for your home, broadband routers for your customers or connectors: everything seems to be in short supply and kind of hand to mouth.”

The long-term impact on telecoms, according to Shelton, “depends on the reaction of both companies and policymakers. Unfortunately, the ongoing tensions between the US and China, combined with the overall shortage, has brought this to the forefront for policymakers.’

In its report, What does the semiconductor shortage mean for telecoms?, telecoms research firm STL Partners predicts that the shortage will last for 2022 and 2023 “until foundry capacity, substrates and component demand softens out”.

Further, it uses Apple as an example of one of the biggest companies to have stockpiled for this very eventuality, a practice that the report says is exacerbating the situation. According to findings, the company’s Q1 2021 revenue increased by 54% to $89.6 billion and its iPhone sales increased by 66% to $47.9 billion, accounting for 54% of all revenues in Q1 2021.

This was achievable only by advance ordering and stockpiling its critical components. Samsung is reported to have been doing the same.

As a result, Susquehanna Financial Group’s trend research into the lead time between ordering a chip and taking delivery found that in July 2021 the lead time increased to 20.2 weeks, up from 17 weeks in April 2021. The increasing lead time gap is a sign of buyers making future supply orders in order to avoid shortages but is also evidence that stockpiling and over-ordering are widespread.

Unsurprisingly, the solution to the problem is in no way straightforward, with Nair commenting: “There are several factors that contribute to this issue, so it is not a single-source solution.”

One possible solution is to diversify global manufacturing, which at present is dominated by Asian companies, with China as the biggest exporter of semiconductors, largely to the US, with an estimated $350 billion in chips exported in 2020, an increase of 14.6% from 2019.

“Currently, the majority of chipset manufacturing happens in Korea and Taiwan. Sourcing chip fabrication operations closer to assembly points and creating geo diversity is critical,” says Nair.

The other part of the answer is to increase the fabrication capabilities of such facilities, as “most of the shortage is coming from manufacturing line capacity being maxed out for components like power management chips”, he continues.

“These chips are manufactured in pretty obsolete plants and use older manufacturing equipment. Increasing the capacity and modernising the technology here is also a critical part of the solution.”

US president Joe Biden has already begun to address this, having issued an Executive Order (EO) on America’s Supply Chains in February 2021, which directs federal agency action to secure and strengthen America’s supply chains.

Specifically, the EO ordered two key assessments. The first is a set of reports on critical supply chains (i.e. semiconductor manufacturing, high-capacity batteries, critical minerals and pharmaceuticals) that were due 100 days after the EO was issued; and, second, reports on sector-specific supply chains that are due to be published early in 2022.

“We have fallen very far behind where we need to be with respect to semiconductors,” commented US secretary of commerce Gina Raimondo during the Fortune CEO Initiative collaborative discussion in July 2021.

“Semiconductors are the building blocks of our entire digital economy and, unfortunately, we are vulnerable now because we don’t make enough leading-edge semiconductors in America.”

With companies such as Newport Wafer Fab, in Newport, and Graphcore, in Bristol, the UK has long held a strong position in the chip industry. But as Simon Beresford-Wylie, CEO of Imagination Technologies, commented, “The recent global shortage has highlighted just how crucial the chip industry is, but also just how concentrated it is when it comes to manufacturing, with the majority taking place in the US and Asia.

“To help address the shortage and mitigate future risk, the UK needs to take advantage of its experience and expertise in design and apply it to manufacturing,” adds Beresford-Wylie. “The UK needs to champion supply chain diversification, as it draws up its post-Brexit, post-pandemic future.”

Acknowledging that a long-term solution will be complex, in the short term, Trivedi says, the answer will involve the cooperation of industry and government, varying according to geography.

“The semiconductor industry itself has the ability to migrate existing high-demand chips to alternative (available) processes; fabs (existing and new) can bring new processes online, albeit at great cost; governments can choose to set manufacturing priorities and/or fund high-cost recovery measures.”

On a more positive note, Shelton says that the silver lining of the current situation is that “demand is robust, new technologies are being embraced and the criticality of the role of semiconductors is now recognised.

“There is real risk with overreach from policymakers. Government involvement can distort markets as they are bad at picking winners and losers. They can inadvertently incentivise the wrong type of behaviour. The government can’t fix the problem, but they can perhaps make it more conducive. The shortage crisis will be fixed by the industry.”

Looking ahead, preventive measures will surely be put in place to ensure that supply is able to meet demand. For Nair that lies in the fabrication process.

“This means increasing the capacity of fabrication facilities, ensuring fabrication operations are geographically diverse, and modernising plants that use older fabrication technology for lower-end chipsets,” he says.

For Trivedi, this includes better design portability and making second sourcing part of the design workflow for every new chip design.

But as the world continues to grapple with this crisis, Trivedi reminds us that, despite the current global turmoil, it is unlikely to happen again.

“This has been the ‘perfect storm’, yet the circumstances causing today’s shortage are not likely to happen again in the same combination.”