The past year has seen a tremendous uptick in the number of artificial intelligence (AI) applications and solutions entering the marketplace. It’s also seen a massive increase in the number of businesses and organizations embracing these solutions to help improve their processes, streamline operations, increase efficiency, and automate tasks.
In fact, according to a study by Deloitte that polled more than 2,600 global business leaders, a vast majority have heavily invested in AI and will continue to do so moving forward. The study found that 79 percent achieved full-scale deployment for three or more types of AI applications, 94 percent thought AI was critical to success over the next five years, and 76 percent planned to increase their investments in AI.
It’s clear that the demand for AI solutions and services is only increasing. And data center owners need to quickly scale up data center capacity to keep pace with this rapidly growing demand. To do so, they need data center construction and White Space Integration to happen much more quickly than ever before.
However, this need to rapidly scale data center capacity comes at a time when data centers are more complex than ever before. This increased complexity requires more precise engineering, more complicated design, extraordinary attention to detail, and effective coordination among a growing number of trades.
…data center owners need to team with a trusted White Space Integration partner early in the design process. These companies understand the requirements of a modern data center and the ways in which the data center has evolved from the ‘production’ data center facilities of the past.
The confluence of these two diametrically opposed trends in the industry makes it critical that data center owners begin the White Space Integration process much earlier in the construction lifecycle. It also has made the use of a trusted White Space Integration partner more important to the timely delivery of completed data centers than ever before.
Partner Early in the Design Process
Today’s data centers – especially those that are being constructed to house power-hungry AI workloads and applications – have evolved to be far more complex than traditional “production” data centers. As rack power densities careen towards 100KW per rack, the way data center owners power and cool the data halls in a data center building needs to change. It’s this change that is contributing most significantly to the increased complexity within the data center.
As power densities spike, each rack produces more heat – making traditional air cooling incapable of meeting modern data center heat removal requirements. Instead, data center owners have been forced to embrace more efficient modes of data hall cooling, including the use of liquids at the rack. The need to run a liquid distribution system within the white space – and even directly to individual processors – has added a new layer of complexity to the White Space Integration process.
Something similar is happening with power. As the amount of energy that each rack requires continues to skyrocket, the amount of electrical equipment and cabling increases, as well. More power requires higher voltage distribution, higher amperage circuits, and more hardware, which adds density to the infrastructure and requires increased robustness and structural support in the data hall.
We’ve only discussed two aspects of the data hall – power and cooling – and we’ve already identified massive changes that impact the mechanical and electrical engineering of the larger data center building.
If a data center is designed like a traditional “production” data center, there is a good chance that these changes have not been accounted for. It’s more than likely that the water distribution, power distribution, cabinet weight, and other characteristics of an AI data center simply have not been accommodated in the design given to the General Contractor – and that can set a data center owner up for significant rework or renovation in the not too distant future.
It’s clear that the demand for AI solutions and services is only increasing. And data center owners need to quickly scale up data center capacity to keep pace with this rapidly growing demand.
This is why data center owners need to team with a trusted White Space Integration partner early in the design process. These companies understand the requirements of a modern data center and the ways in which the data center has evolved from the “production” data center facilities of the past.
Liquid cooling requires networks of complex mechanical piping to be intertwined with today’s power-hungry power distribution systems and fiber-rich cabling plants. The white space is truly a new and different environment now. Bringing in an expert who understands this new reality early on in the process – in the design phase of the construction lifecycle – is essential to ensuring these systems are properly accommodated and accounted for in the data center design.
However, the benefits of working with a trusted White Space Integration partner don’t end with the design process. Keeping this team invested and involved in all aspects of the construction process can also play a massive role in eliminating rework, expediting project completion, and keeping costs down. And starting White Space Integration while construction is ongoing can also pay dividends.
Getting an Early Start
In many of today’s data center construction projects, the White Space Integration begins at the very last minute. This has been an industry best practice for a decade because it allows the owner to make white space design changes up to the very end of the construction lifecycle, despite adding a little risk to the overall data hall delivery process. But it’s not the best practice anymore – at a time when data center owners need projects completed quickly to rapidly scale with demand.
We believe the optimum approach is to align the General Contractor responsible for the construction of the data center building with the White Space Integrator early in the process. Coordinating the General Contractor and the White Space Integration team, and starting work on the data hall as soon as possible, can expedite construction.
Coordinating the General Contractor and the White Space Integration team, and starting work on the data hall as soon as possible, can expedite construction.
There are multiple places within the data center where the White Space Integration team’s responsibilities and scope of work overlap with those of the General Contractor or trade professionals. Enabling these disparate organizations to collaborate on portions of the facility where their scopes overlap or connect can introduce significant efficiencies into the project.
For example, all parties involved in the data center construction project will have responsibilities in the power and cooling galleries located just outside of the white space. When the White Space Integration team can coordinate with the base building contractors, the systems installed first are done with the subsequent downstream requirements in mind. Physical space can also be more efficiently allocated and coordinated to enable the proper installation and future maintenance of both the base building systems and white space infrastructure.
By enabling coordination between the White Space Integration team and the General Contractor, data center owners can ensure that all parties know how their work impacts the next deployment phase. This can help to eliminate time-consuming and costly conflicts that arise without coordination. It also helps with the identification of engineering issues that need to be addressed early.
Today’s data center owners – whether they’re hyperscalers, colocation providers, government agencies, or private enterprises – are under incredible pressure to quickly bring new facilities to life to meet their demand for advanced AI solutions. But if they’re going to rapidly scale to meet the demand for AI services and solutions, they’re going to have to partner with a White Space Integration team that they trust, get them involved in the construction lifecycle early, and begin work on their data halls as soon as possible. Otherwise, rework, a lack of collaboration, and oversights in design could result in lengthy construction projects.
To learn more about the impact of AI on the data center, click HERE.
