There’s no denying that advanced applications like Artificial Intelligence, Machine Learning and High-Performance Computing are having their heyday. Groundbreaking use cases from these intelligent platforms are enabling businesses and teams around the world to enhance resilience with predictive maintenance, personalize customer experience and a whole lot more. Within this single article alone, we see IBM outline a whole new world of opportunities for healthcare, education, manufacturing, energy and beyond — but there’s another side to this coin.
The ‘AI’s resource demands are huge’ discussion is nothing new. In fact, the staggering data, energy and space demands created by generative AI and these other innovative systems have become somewhat of a point of awe. For every data center company around the world trying to meet these needs, these realities have the opposite effect.
In a world on the precipice of an explosion of new IT developments, the reality is that capacity shortages are threatening to strangle progress in these key areas.
Meeting Applications Where They Need to Be
From rack densities to power sources, the road to supplying the resources these applications need is complex — and more pressure is being applied all the time. While the technical demand on data centers expands, the push to become more efficient, green and spatially/environmentally conservative is insisting on new, novel methods from the other end of this equation too. Data centers now find themselves in the middle of a challenging quandary. With higher rack density comes more heat, with more heat comes more power draws for cooling, and with more cooling comes greater resource consumption, making carbon footprints bigger.
Add on to this the fact that planning new data centers, funding their construction and actually building them takes time — and AI, ML and HPC are hungry now. Finding the capital and time to supply infrastructure only compounds the desperate hunt for space and energy, leaving advanced applications in the lurch. Even connecting to a viable, sizable power source capable of keeping up with new demands is becoming harder.
Across the nation, grids and power providers are struggling. As The Washington post notes, “In Georgia, demand for industrial power is surging to record highs, with the projection of new electricity use for the next decade now 17 times what it was only recently.” Georgia isn’t the only one struggling. In Arizona, the state’s largest utility is projecting a total dearth of transmission capacity before the end of the decade without major upgrades. Even on a global scale, generative AI is expected to swell worldwide electricity consumption in the years to come— the European Commission estimates it will increase 60% by 2030.
Ultimately, this leaves IT innovators wondering where they can go to escape the squeeze and find data centers that have a steady pipeline of power — with room to grow.
Data Center Investment In the Driver’s Seat
Today, while AI, ML and HPC applications are driving innovation, the arbiters of true progress are the data centers. Their ability to source and provide ample space and power will determine just how successful these new applications can be — and how widely and fully they can be adopted.
Investing intelligently with a careful eye on the future is key. Ongoing projects and capacity planning are crucial for staying ahead of long lead times and new build projects, allowing data center operators to maintain a constant foundation of growth that can match that of AI and its peers. Here’s the good news: AI innovators and others like them have their pick of the litter when it comes to data center location, freeing them up to find the providers delivering necessary resources.
Yes, most of today’s experiences depend on low-latencies — but AI workloads for training and tuning tend not to be. This means data center space and power can be sourced from many places on the map. Fortunately, for both latency-sensitive applications and non-speed-obsessed workloads, the secret weapon lies in markets that are off the beaten path. Yes, in metros like Omaha, IT innovators can find the best of all worlds: Access to the edge, access to secure colocation and (if you choose wisely) access to the space, power and sustainable practices needed to support the future of technology.
If you look in the right places, you can even find data center teams that are well-equipped with both expertise surrounding AI, ML and other advanced applications in addition to the infrastructure needed to power them. In fact, we might just know a team that checks all the boxes…
Want to learn more about how we provide those strong foundations here at 1623 Farnam? Check out our location advantages here and hear more about our recent power investments here.