Data-driven: new technologies fueling the rise of data centers - United States
Data-driven: new technologies fueling the rise of data centers
Steady… and slow.
Fundamentally, real estate is a prized class of investment because there’s no escaping its necessity. You’ll always need a place to live, for instance. But as much as the people-driven nature of real estate grants it stability over time, it also sets a natural limit on its growth. You’ll only ever need one place to live, after all, and one place to work (if not less). Additional demand requires additional people, and with U.S. population growth projected at just 0.3% annually for the next 30 years, relying on new real estate occupants to drive demand looks like a losing proposition.
In short, traditional real estate is driven by populations growing at slow rates that tend to slow further over time. But while populations grow slowly, technology continues to advance at an exponential pace, and there is one class of real estate that combines the people-serving necessity of real estate with the growth potential of technology:
The global datasphere – that is, all the data processed and stored in the world annually – doubled between 2019 and 2021, and it’s projected to double again within a couple of years. Understanding growth of this magnitude can be a challenge without some understanding of where that data’s coming from, and while there are far too many contributors to mention, this article explores some of the recent and emerging technologies that are driving unprecedented demand for data center real estate.
Consider this scenario: you’ve just upgraded from last year’s iPhone 13 to this year’s 14 Pro. You immediately notice that the new phone takes stunning photos. You don’t immediately notice that the photo file size is 3x larger than on last year’s model, or 15x larger than on the original iPhone you bought in 2007.
You haven’t changed your own behavior at all, yet your photos now require several times more storage capacity than they used to. And if you thought the increase in photo file size was dramatic, remember that online mobile video sharing hardly existed in 2007. Today, sharing full HD videos is a primary form of communication. The data storage implications, especially for social media giants like Meta and ByteDance, cannot be understated.
Speaking of online video, the world has truly come to rely on it in recent years. The pandemic drove many companies to shift their workforces to remote work, which led to a surge in demand for video conferencing to meet the needs of daily business operations. Zoom users surged exponentially from 10 million daily users in December 2019 to 300 million in May 2020. At the same time, Zoom stock climbed from $68 a share at the beginning of 2020 to a high of $174 per share, while the company’s valuation ballooned to over $48 million.
Though demand for video conferencing is no longer as high as it was during the height of the pandemic, it has become the new normal for many workers and the industry is expected to grow from $6.87 billion in 2022 to $14.58 billion in 2029, according to Fortune Business Insights. Depending on video quality and the service provider, video streaming can consume between 225 and 540 MB of data per hour for one-on-one calls and between 450 and 810 MB of data per hour for group calls.
However, even the highest-resolution video conference lacks most of the depth of a face-to-face interaction, and the data-intensity of teleconferencing will likely continue to increase as the industry adopts augmented-reality and virtual-reality elements to help bridge the gap.
Augmented and virtual reality
Immersive conference calls are just one of a near-infinite list of applications for augmented and virtual reality (AR and VR, respectively), and these technologies will become mainstream forms of entertainment, communication and education. A wide variety of simulations can be used to immerse military personnel, athletes, coaches, as well as medical and dental students in life-like situations, which lower the costs and risks of training. VR can create immersive spaces in which young students engage in safe hands-on learning or in which therapists help patients overcome traumatic events or phobias.
Aside from its many applications, the most impressive thing about virtual reality is the amount of data it consumes. VR can produce up to 1 TB of data per hour, which is equivalent to 17,000 hours of recorded music. Moreover, data transfer must be near-instantaneous in order for a multi-user simulation to feel natural or for augmented-reality models to line up neatly with reality, which necessitates more digital infrastructure closer to the systems’ end users, on what is known as the edge.
The virtual reality space has attracted the interest of startups and established tech giants alike. Meta – Facebook at the time – was an early mover, acquiring Oculus VR in 2014 and developing a suite of hardware and software products from that platform. Sony and others have invested heavily in the intervening years, and Apple is expected to unveil a VR platform of its own in 2023. The most compelling virtual reality products available have been on the shelves for only a short time, and it’s safe to say the industry is still nascent. It will be difficult to comprehend the number of digital experiences that will be created in VR over the coming years, or the amount of user data that will be recorded within those experiences.
Autonomous vehicles are poised to have a tremendous impact on global transportation infrastructure, but few people appreciate the way in which they will transform digital infrastructure. Even basic autonomy (think driver-assistance systems) collects more than 1 TB of data per hour, while high-level autonomy can generate up to 20 TB. The number of high-level autonomous vehicles (levels 4 and 5 on a scale of 5) in the United States is projected to reach 20.8 million in 2030, which could equate to trillions of terabytes of data generated every year.
Much of this data must be offloaded from vehicles for storage and processing, and this need will likely cause digital infrastructure to evolve along with electric vehicle charging stations. After all, cars visiting a station must plug in and sit idle anyhow. In addition to downloading data at charging stations, self-driving systems will also influence the rollout of edge data centers and 5G, which a car would use to communicate in real time with other vehicles and infrastructure while driving.
A self-driving future is closer than most people appreciate. Drivers using Tesla’s Autopilot system experienced accidents once every 4.31 million miles – about 9 times less frequently than an average human – chipping away at the notion that driving is too complex a task to be automated. Public opinion has begun to shift in response; the share of Americans that believe self-driving cars will not be safe dropped from 74% in 2017 to 48% in 2020, which is major progress considering the fact that no autonomous systems are yet certified for full self-driving on public roads.
Data is being created at an ever-quickening pace. Hardly anyone could have accurately predicted just a few years ago how much growth we would see by 2022, and the data center market is probably not done surprising us just yet. The technologies we’ve explored today are just a few of many, and even these will see new use cases emerge as they mature. Moreover, entirely new technologies that are yet unimaginable will come to market. Let’s just hope some of those technologies help us build data centers faster!
Subscribe for data center insights and reports
Commercial real estate services for data centers
Avison Young provides consulting, brokerage, tenant & landlord representation and real estate management for data centers across the United States.