By Raul Martynek, CEO DataBank
In 2023 we saw the explosion of AI and its immediate impact on the data center industry. This development alone has completely transformed traditional business models and once-accepted norms. It’s a transformation in the future of power, data center capacity, and even hyperscalers’ new requirements going forward.
What will 2024 bring? It’s hard to tell for sure. I predict the following developments related to AI and the way it will affect the entire data center industry.
Looking back, I think we will remember 2023 as the birth year of AI. It’s hard to believe it was just a little over a year ago when OpenAI unveiled ChatGPT. It was truly a jaw-dropping, game-changing innovation. It immediately captured the imagination and attention of the technology and business community. ChatGPT has spawned a surge of investment and experimentation in AI companies, tools, and capabilities.
At this time last year, very few people were talking about AI – or included it in their list of 2023 technology predictions. Now it seems to consume all the oxygen in the room. So, what will 2024 hold for new AI innovations? At the risk of sounding like a contrarian, I’m tempted to say, “not much.” Let me explain.
I firmly believe in the future impact of AI and that it is similar to the advent of the Internet in the mid-1990s. Fundamentally, it will change every human activity. Also, I believe in 2024 organizations will begin wrapping their heads around incorporating it into their businesses.
There are a number of reasons for this:
I think while we will continue to read a lot about AI in 2024 (and forever more), we will not see any huge breakthroughs. Instead, most organizations will spend their time developing comprehensive plans for incorporating Gen-AI into their specific business models. This was true in the 1990s with the introduction of the Internet, and I think the same will hold true here.
It reminds me of one of my favorite sayings from renowned futurist Roy Amara: “We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run.”
As we are coming to understand, the deployment of AI is inextricably linked with data center capacity. I think 2024 could be the “Year of Data Center Scarcity.” By all accounts, 2023 was a record year for data center leasing. Many companies, like DataBank, effectively leased two years of capacity in one year. The vast majority of the data center leasing for AI workloads still needs to be commissioned; much of the sector will be spending all of 2024 and beyond converting the huge backlog into production.
This has driven vacancy levels to all-time lows in all major markets. We expect continued investment in AI workloads. We expect continued public cloud growth and enterprise digital transformation in 2024. This means data center capacity will be even more constrained in 2024. Further, we expect pricing to continue to increase throughout the year as buyers compete to secure what little space is available.
Although the industry, including DataBank, is working to add new facilities, data center construction is now a 24-48 month endeavor. Therefore, the wave of demand from 2023 is not possible to be filled by next year.
As I mentioned above, I expect public cloud growth to accelerate in 2024. Of all the major public clouds, Microsoft clearly has the best-defined strategy for incorporating Gen-AI into its product offering and identifying a means to monetize that investment. The rollout of Co-pilot across the Azure stack, likely, will accelerate as enterprises understand the productivity benefits of the offering. Based on conversations I have with developers, Co-Pilot for GitHub radically improves productivity for software development. There is already early evidence of this. On its most recent quarterly earnings call, Microsoft attributed a 3% increase in Azure cloud revenue to AI alone.
Oracle is also very well positioned to benefit from Gen-AI given its partnership with Tik-Tok and Nvidia. Tik-Tok’s entire business is based on “internet AI.” Undoubtedly, they will continue investing to make the algorithm even more powerful, and Nvidia will continue to leverage Oracle’s relative semiconductor chip “neutrality” to deploy its CUDA GPU stack within Oracle data centers. Oracle reported a 50% increase in public cloud growth during its most recent earnings call. Also, they indicated it had billions of dollars of incremental cloud/Gen-AI demand.
AWS is beginning to make news about its investment in AI startup Anthropic and collaboration to develop and train more capable large language models (LLMs). Recall, AWS is an IaaS company – their platform is a technology stack for developers to build other products on top of it (think Slack). Their monetization strategy will require developing a compelling alternative to Nvidia CUDA and convince developers they should build their Gen-AI applications on their (proprietary stack) not Nvidia’s. Given AWS’s track record and scale, undoubtedly, they will make inroads. Ultimately, they are dependent on enterprises and technology companies to article to launch their Gen-AI roadmap (see above).
Google recently released its own AI model, Gemini, and claims it outperforms ChatGPT in widespread testing. Like AWS, Google Cloud depends on other developers to consume their services. I suspect Google’s energy direct Gen-AI toward protecting and expanding its lucrative search business first. With all these developments – many that seem too similar – we’ll have to see newly differentiated offerings to allow these vendors to stand apart from the competition.
It’s an exciting time for the entire tech sector.
About the Author: Mr. Martynek joined DataBank in June of 2017 as the CEO and is a 20+ year veteran in the telecom and Internet infrastructure sector.
Discover the DataBank Difference today:
Hybrid infrastructure solutions with boundless edge reach and a human touch.