According to Elon Musk, artificial intelligence (AI) is on the brink of surpassing human intellect as early as next year. Is there truth to his prediction? The potential for a “Gutenberg moment” in technology looms large, promising a seismic shift in innovation. Yet, even if this transformation unfolds more gradually over the coming decade, one thing is certain: it will demand colossal energy resources to fuel AI’s appetite for cloud computing.
Data centers stand at the front and center of this burgeoning demand. In other words, data centers are powering the AI revolution. Owners, investors, and builders are racing against time, grappling with the herculean task of scaling up infrastructure to meet the insatiable appetite for computational power.
Beyond the Hype: The Cloud’s Unsung Catalysts
The cloud’s growth is being propelled by three market forces:
- Decarbonizing our built environment. For years, cloud technology has been instrumental in construction, fostering enhanced collaboration and driving design optimization. Today, it stands as a pivotal tool in our quest for sustainability and resiliency. Generative AI engines like Autodesk Forma, cove.tool and others are optimizing building operations to reduce carbon and extending these engines to embodied carbon are on the horizon.
- Supercharging workforce productivity is another area where AI will fuel a renaissance. Innovations in machine learning, natural language processing, and image pattern recognition are advancing at breakneck speeds. This technological surge is spawning a new generation of startups, each poised to boost productivity across a myriad of occupations.
- The third trend reshaping the cloud landscape is digital twinning—the intricate simulation of our physical world. As we endeavor to model the past, present, and future of our built world, the data demands are staggering. The Internet of Things (IoT) is expanding exponentially (30B devices by 2030, Statistica), capturing vast quantities of sensor data that must be meticulously analyzed.
The Cloud’s Triple Engine: Driving Computational Demand
These three trends form a triad that is driving cloud computing forward, fueling an unprecedented demand for computational horsepower. With cloud computing as an enabler, the data center landscape is undergoing a metamorphosis. For example, new design paradigms like de-centralized distribution, reuse of existing building stock, and heightening sustainability requirements are forcing a new calculus for design and construction that factors in project complexity, speed to market and changing sustainability goals.
Image: Common challenges facing data center design and construction
To gain deeper insights into this dynamic sector, James DiNoia of the STO Building Group recently hosted a compelling podcast titled “The Slingshot Effect: Today’s Mission Critical Market.” In this series, DiNoia engages with data center owners, shedding light on the challenges and opportunities presented by the current market conditions.
The Capacity Conundrum: Balancing Demand and Delivery
Despite the quickly increasing demand for data centers, the industry’s capacity to deliver remains a step behind. Construction teams are compelled to innovate, reevaluating traditional methodologies to keep pace with the rapid deployment required in today’s market.
Supply chain challenges are leading to project schedule drift and cost escalations. Bringing trades onboard early is a solution but this still pressures design teams to deliver before the design is complete. While new construction methods like prefabrication are being explored, they are not yet ubiquitous in the marketplace (McKinsey, 2024).
One trend we do see successful at speeding up data center delivery is the adoption of more proven delivery methods used in pockets across the US and Europe and also in the Engineering Procure Construct (EPC) market. For example, when the core and shell is designed with structural steel, engineering teams integrate the steel connection design and deliver connected 3D steel models as part of their bid deliverables. Doing so can shave 6 to 8 weeks off the bidding and fabrication schedules. One recent study here at Qnect showed that optimizing the schedule on a data project would capture several months of schedule by integrating the connection modeling and engineering early but also using steel connection types that speed up steel erection. Removing just one bolt per connection (while meeting design requirements of course) by optimizing can speed up installation on the jobsite and save time in the shop as well.
Over the past year, we’ve seen several data center owners adopting this method either by vertically integrating with in-house design teams, or partnering with outside engineering-detailing teams. This is a creative approach for data center owners that repurposes proven delivery methods to speed up their capital project execution.
So where should data center owners start? Talking to their design teams is a good first step. They should together explore the latest design strategies to meet the changing programming requirements, procurement realities and project delivery mechanisms. This includes leveraging off the shelf methods for delivering the core and shell (steel and concrete) that takes procurement off the critical path. For structural steel, this means using a delivery that is being coined the early connected model.
We as an industry can meet the looming demand for data center construction to support the growing appetite of the cloud. However, we need to take our heads out of the clouds (no pun intended) and re-imagine how we deliver construction projects.
