Data Centers Will Never Be the Same Again After AI
- By Lachlan Colquhoun
- October 16, 2023
As Artificial Intelligence gains momentum, questions are being asked in the data center industry.
Is your data center ready for AI? And if it isn't, where do you go looking for a provider who is?
AI has pervasive implications for so many aspects of data center technology, and just as AI is transforming the cloud, it will also change the data center, which is happening now.
This is a threat to existing providers and an opportunity for new players.
Big changes afoot
In the security world, AI is finding its way into corporate firewalls for threat hunting.
Ethernet switch ports are beginning to demand AI-based servers.
Data centers must have servers capable of leveraging AI chips that can perform the vast volumes of computations that are required for new applications.
With the volumes of computations comes massive amounts of data, which also needs to be stored.
“Conventional data centers will be challenged to keep up the pace as more capacity is used for AI.”
The demand for more servers running on higher-performance chips will also require massive amounts of power, posing fresh challenges in sustainability and green IT.
Servers running AI require up to seven times the power conventional servers need.
The investment required will be on a massive scale.
According to Data Bridge Research, spending in the global AI infrastructure market, including data centers, is expected to grow at a compound rate of 44% between now and 2029, reaching a massive USD422 billion.
Ambitious new players
Given the scale of this investment, the USD2.3 billion debt facility raised by CoreWeave in August is a drop in the ocean of what will be spent.
CoreWeave may not be a household name in the technology industry, but it has big aspirations.
Things can change rapidly in the fast-moving world of AI, as Nvidia's market capitalization of over USD1 trillion has shown.
CoreWeave was founded six years ago in New Jersey and has seven AI data centers. The new funding will help the company double that by the end of the year.
They are not the only new players; in the U.S. alone, emerging companies such as DataBank have similar ambitions.
It would appear there is a new generation of data center providers positioning themselves with greenfield facilities as they look to take the opportunity.
AI in the center
While data centers need to evolve to service their customers, AI will also transform their operations.
The first change is around sustainability.
According to a study by the University of California Riverside, when Microsoft partnered with OpenAI to train GPT-3, the exercise consumed 185,000 gallons of water.
This is around the same amount of water used to produce 320 Tesla cars or cool a nuclear reactor.
AI will be needed to help mitigate power and water use in this context. AI can drive optimal operating parameters for temperature and humidity and help save up to 40% on energy use.
Security is also a high priority, and AI and machine learning tools have a role in the data center's physical security and cybersecurity.
At the site, smart cameras and intrusion detection systems can work with mobile robots to maintain security. At the same time, AI has a significant role in threat detection and alerts.
As it learns, AI can become increasingly effective as a protection layer, make data storage and processing more effective, and automate incident responses.
Finally, AI has a role in the performance of the data center as a physical asset and helps reduce downtime.
This is delivered through AI monitoring, capturing and analyzing data to predict and prevent equipment failure and identify necessary maintenance.
Loop of change
All these changes are part of the quickening pace of AI adoption, which is set to snowball exponentially.
In this change loop, conventional data centers will be challenged to keep up the pace as more capacity is used for AI.
Conventional data centers running close to capacity will struggle to deliver to customers moving from pilot AI projects to adoption at scale.
This challenges the data center business model as the industry sits on the cusp of a massive spike in demand for capacity, which in many cases requires significant investment today to deliver on the promise of tomorrow.
Lachlan Colquhoun is the Australia and New Zealand correspondent for CDOTrends and the NextGenConnectivity editor. He remains fascinated with how businesses reinvent themselves through digital technology to solve existing issues and change their business models. You can reach him at [email protected].
Image credit: iStockphoto/metamorworks