Why the global semiconductor shortage could harm real estate
Anyone who has tried to buy a new car recently will have an acute understanding of how the global semiconductor shortage has affected that industry.
New car sales in the UK are currently a quarter lower than they were in June 2021 – which, bear in mind, was during the pandemic. Buyers worldwide are being told they may have to wait for months, or longer, to pick up the keys.
But this isn’t due to a shortage of steel or vital engine components – this is for want of a handful of microchips made with silicon that would normally cost a few pennies. Now car manufacturers are going to extraordinary lengths to get around this problem – using chips meant for washing machines, rewriting programmes to use less silicon or ditching some features altogether. Ford has stripped out some, promising to add them back in a later date.
Anyone who has tried to buy a new car recently will have an acute understanding of how the global semiconductor shortage has affected that industry.
New car sales in the UK are currently a quarter lower than they were in June 2021 – which, bear in mind, was during the pandemic. Buyers worldwide are being told they may have to wait for months, or longer, to pick up the keys.
But this isn’t due to a shortage of steel or vital engine components – this is for want of a handful of microchips made with silicon that would normally cost a few pennies. Now car manufacturers are going to extraordinary lengths to get around this problem – using chips meant for washing machines, rewriting programmes to use less silicon or ditching some features altogether. Ford has stripped out some, promising to add them back in a later date.
What does this have to do with real estate? Well, what is currently happening with cars could soon happen to buildings.
As an industry, real estate is increasingly reliant on data. That data depends on tech and that tech is built on tiny bits of wafer-thin silicon.
“The danger is that the enthusiasm for smart buildings, which is pretty strong at the moment, will be punctured,” says WiredScore president and managing director William Newton.
Instead of “riding the tailwinds” of a post-pandemic drive for a better tenant experience, or the need for more sustainable buildings as we recognise the impact of the climate crisis, Newton believes the long lead times and increased cost caused by semiconductor shortages will cause the adoption of smart building technology to stall.
“Landlords will think, ‘Well, this is going to cost more and take longer,’” he warns. They may decide it can wait and do it retrospectively. “Or they will delay refurbishment works or improvement works completely.”
In other words, planned smart buildings may become dunces. “We will start to see that happening in a pretty meaningful way,” he says.
Hot chip
It could be argued that this is a fairly fringe issue. After all, the whole of the property industry accounts for barely a percentage of a percentage of the world’s entire output of semiconductor chips. There are more pressing issues for real estate, such as the rising costs of steel, finance and talent.
But the need for chips is growing exponentially. Smart buildings are currently the exception. Yes, we hear about them all the time, and it seems that every new building that emerges from a construction site is smarter and more connected than the one before. In fact, according to a recent report by Juniper Research, there are presently around 45m connected or smart buildings in the world. But it accounts for less than 1% of the world’s real estate.
“We are seeing acceleration in adoption due to the nature of ESG and regulations worldwide, and that continues at pace,” says CBRE’s head of digital sales Nick Wright. “And the more you begin to have a standardised set of what layers of technology are expected or wanted, I think you will see more and more buildings come into line with that.”
As a result, that 45m is expected to grow to 115m by 2026. And in order for real estate’s tech revolution to happen, it, like everyone else, needs more chips. According to the Juniper study, global shipments of sensors used in smart buildings will exceed 1bn annually in 2026 – up from 360m in 2022 – a growth of 204%.
That’s a lot more chips.
“The scarcity of semiconductors will only be exacerbated as the real estate sector becomes more reliant on proptech solutions to achieve the myriad of data and sustainability targets that are currently being set,” says Johnnie Wilkinson, chief executive of real estate fund manager Greenman. “This is one of the reasons why we have moved to a more integrated investment management solution, with the capability to provide many of these services in-house.” To do this it has created joint ventures with Greenman Energy, Potager Farms, TechMash and its Luxembourg fund administrator Dinamik. “By providing these services directly, rather than through a third party, we will be in much better control of the process.”
Even so, demand from real estate is unlikely to exceed that of the current top chip consumers any time soon.
The majority of demand comes from computers and related kit, consumer electronics (can’t get hold of the PS5 “for the kids”? That’s because of the shortage of chips), telecoms and industrial electronics; and defence and space. The automotive industry represented less than 5% of demand in 2019 and about 8% last year. With the shift to EVs and smarter onboard systems, that is expected to hit 15% by the end of the decade, says McKinsey & Co.
According to Juniper’s estimates, real estate could become a consumer of chips on a similar scale to the automotive industry.
The scarcity of semiconductors will only be exacerbated as the real estate sector becomes more reliant on proptech solutions to achieve the myriad of data and sustainability targets that are currently being set
Johnnie Wilkinson, Greenman
And that highlights the problem. Real estate, despite being a whale in many other respects, would still be small fry in the chip market. It simply wouldn’t have the heft, buying power or mature supply lines to compete.
“Buildings aren’t going to be the major user of semiconductors, in the same way the automotive industry isn’t the biggest demand,” says WiredScore’s Newton. “But it absolutely needs them. Semiconductors are critical to the automotive industry and without them production is gridlocked.”
Exponential growth
Real estate could face the same issues as car manufacturers. The automotive industry is currently ordering 20% more chips than it needs – at inflated prices – to make sure it doesn’t get squeezed out. While some chips can be left out, others are essential. After all, you can’t sell a new car if the door won’t unlock when you push the little button. As a result, costs have soared. Electronics have gone from being just 18% of a car’s cost in 2000 to being 40% of its cost in 2020. This is projected to rise to 45% by 2030.
And, as with buildings, demand for chips is rising as the cars get smarter.
In fact, demand is rising across the board, by an estimated 8% a year.
In December, Deloitte warned the chip shortage was likely to last the whole of 2023. Now it has revised its estimate, saying inflation, the cost of living and the threat of a global recession will free up a fair few feet of wafer, as fewer people will be willing to spend what little cash they have left on a PlayStation or smartphone.
But that will do very little to change the course of the general trend of demand.
The chip industry, from Taiwan to Ohio, is planning to increase production to unprecedented levels to cope with this. The three largest players in the market alone are expected to spend $200bn (£169bn) in capex by the end of 2023, and that is expected to reach $400bn by 2025, according to McKinsey & Co.
Semiconductor sales are expected to grow by 6-8% each year, with the market expected to be valued at $1tn in 2030, the consultancy adds. So that will solve the problem, right? Perhaps not. Demand is being driven by consumer electronics, 5G, AI, the internet of things and so on. Like a seagull at the seaside, those that have always needed chips now need more chips.
But where demand is expected to go through the roof is where the parts of our lives that are not yet fully connected need to be smart. Like buildings.
The commercial buildings currently vying to be smartest will just be the tip of the iceberg, Juniper states. Nine-tenths of the growth predicted by its report will come from commercial property, as landlords and tenants recognise the virtues and cost-efficiencies of connecting their buildings and mining their data.
“Smart building platform vendors will understandably focus on non-residential use cases, as these provide a stronger return on investment,” says Dawnetta Grant, industry analyst at Juniper Research, and research co-author. “But they should not neglect the importance of residential deployments as environmental concerns intensify.”
In fact, the smart money is on smart technology becoming mandatory in all homes and commercial building before 2050, as long as the world sticks to its climate targets.
To achieve that, billions upon billions of chips will be required, with real estate becoming ever more dependent on them.
But CBRE’s Wright is not convinced this will happen any time soon. “If every building in the country became smart tomorrow, the shortage of chips would be a lovely problem to have,” he says. “But it is going to become a standardisation. At what level that standardisation will be, I don’t know. Will it be just four or five different technologies in a building? Or all-singing, all-dancing digital twins? I think we are a way off that.”
Eggs, not chips
In the medium term, he believes there will be a base layer of around 10 technologies that people will expect in a building. Some of those will be app-based and rely on human data collection and existing kit, like smartphones. But the other half will require sensors. Billions of sensors. The question is not whether there will be the demand, but when.
And there is potential for an even bigger issue.
Back in 1990, about 33% of chips were produced in the US and around 10% in Europe. As of last year, just 12% of chips were made outside Asia. In fact, only 35% were made outside Taiwan. Around 53% are made by one company, Taiwan Semiconductor Manufacturing Company.
The problem, the US government says, is not with the chips, but with the eggs – they are all in one basket.
[caption id="attachment_1044880" align="aligncenter" width="847"] WiredScore’s William Newton[/caption]
“The absence of foundries in the West that can make the kind of chips that TSMC makes is a worrying global oversight,” says WiredScore’s Newton. “Should an aggressive China take control of Taiwan, that could lead, well… to the end of the world as we know it.”
Not that Newton is predicting an invasion. In fact, he thinks it is improbable – unlike the US state department, which thinks it is highly likely between now and 2025. But it wouldn’t have to be an invasion, or even a pandemic, to cause chaos.
“When Taiwan experienced a drought, there wasn’t enough water for the incredibly water-intensive process of making the chips,” Newton points out. “That had a meaningful effect on global supply chains.” So too when the Suez Canal became blocked last year. “The number of component parts of the supply chain that have to function really well for us to be able to have anything resembling the supply we need is just astronomical. You only need a few small things go wrong and it all falls apart.”
US president Joe Biden has vowed to correct that, offering $50bn of state aid to build more foundries in the US. In January, Intel said it would spend $20bn building two cutting-edge foundries outside Columbus, Ohio, in addition to the two foundries it announced in 2020 in Arizona.
The scale of investment means these areas of the US are going to become boomtowns, with the real estate sector reaping the benefit. Within five years, they should be up and running, as will some of the associated development – homes, offices shops and schools – that will rise up alongside.
Similarly, a further $80bn or so is being spent to build foundries in Europe, while the possible sale of Newport’s Wafer Fab is being treated as a matter of national security.
But these facilities will take decades to come online. And while real estate may reap the benefits creating those boomtowns in the short term, unless it secures its own supply of silicone, the industry will be left back in the Dark Ages.
Newton is optimistic, though. Yes there will be cost implications, and maybe WiredScore’s dreams of a fully connected world of smart buildings may take a little longer to crystallise. But it will happen, he says.
“At the moment I’m trying to buy a car,” says Newton. He has been waiting for months, the costs have gone up, the spec has been changed. “It’s a pain, but it doesn’t mean I’m not going to buy a car. Same with smart buildings.”
To send feedback, e-mail piers.wehner@eg.co.uk or tweet @PiersWehner or @EGPropertyNews
Main image © Shutterstock