Govind Choudhary, Head of Service Providers and Market Strategy, Asia Pacific, Digital Realty, April 2024

Recent developments in generative artificial intelligence (AI) have ushered in an era of accelerated technological advancements. From healthcare to manufacturing and financial services, businesses of all stripes are looking for ways to leverage AI-powered solutions to streamline processes, improve decision-making, and transform entire sectors.

As the data center industry moves to support the new generation of high-power systems needed to train AI models and perform inference, the developments in AI have also opened the door for service providers to offer the innovative, high-performance solutions that organizations need to enable their cutting-edge AI initiatives.

Data centers as connected data community hubs for AI

Even as AI reshapes industry demand and the broader data center landscape, an often-unspoken topic is the fundamental role of data. Data is, in effect, the lifeblood of AI – without vast amounts of diverse and high-quality data, AI models cannot effectively learn, adapt, and deliver the desired results.

The volume of data we generate continues to grow at an exponential rate and must be properly managed, stored, and processed. Data attracts data and the creation of new data-centric services, which in turn drives the demand for even more sophisticated data centers.

However, as data sets grow inexorably larger, the resulting Data Gravity is not only driving companies to rethink their IT architectures, but also highlighting the critical need for hybrid IT infrastructure to support the complex data flows of AI.

Hybrid IT architecture, a mix of on-premise, colocation, public cloud, and private cloud, provides the flexibility and scalability required to handle the diverse data needs of AI. This is because:

AI development often involves both public and private data: Public datasets can be used to train base models, while private data is needed for fine-tuning and customization. Hybrid IT enables secure and efficient access to both types of data.
Data processing requirements can vary: Some data processing tasks might be computationally intensive and require the high performance of on-premises infrastructure, while others might be more cost-effective to run in the public cloud. Hybrid IT allows organizations to allocate resources based on specific needs.
Data privacy and security are paramount: Hybrid IT provides organizations with greater control over their data, allowing them to implement appropriate security measures based on data sensitivity and regulatory requirements.

This is reflected in the increasing adoption of hybrid IT, with an IDC APAC cloud survey¹ reporting that 77% of organizations have a multi-cloud or hybrid cloud environment, and this number is expected to reach 84% within two years.

The greater adoption of hybrid IT architectures means that organizations are seeing an increased need for enhanced data center-to-data center or data center-to-cloud connectivity with which to integrate their increasingly complex digital infrastructure deployments. As a result, they are looking for multi-tenant data centers that offer flexible options – secure, private dark fiber solutions that offer customers maximum control over their connections, as well as cost-effective lit fiber solutions customers can lease more economically for less demanding traffic.

Taken together, what's clear is that data centers are no longer standalone buildings for server and networking hardware but crucial data hubs that serve as a digital clearing house for the unceasing ebb and flow of data in our world today. Hybrid IT infrastructure, with its ability to support both public and private data flows, plays a critical role in enabling organizations to leverage the power of AI while addressing data privacy, security, and cost considerations.

The power of an open ecosystem of data centers

Either between departments or with external parties, the ability to seamlessly exchange data can make or break modern businesses. Carrier-neutral data centers are the obvious choice when it comes to interconnection between network providers and colocation customers, and this is another reason our customers come to Digital Realty.

With PlatformDIGITAL^®, we have built an open, collaborative platform to bring our service provider partners together, giving enterprise customers the capabilities and choice to grow quickly through an ecosystem for services and connectivity offerings, all under a single contract and readily discovered and accessed from a single portal.

With a proven foundation and methodology to scale digital business and solve for Data Gravity, customers can access a rich ecosystem of service providers with global reach for reliable, cost-efficient data connectivity. This allows them to build and orchestrate their ideal solutions with ease and efficiency, opening the door to innovation without the constraints of physical location or closed ecosystems.

With enhanced connectivity options, access to internet exchanges (IX), and on-demand services, Digital Realty’s customers can connect with a vast data center network and open a new world of possibilities, achieving this in a significantly shorter amount of time than was possible in the past. On their part, service providers get to grow with their customers as they expand or establish a presence across the region over time, on top of the ability to cross-sell brand new services.

A global ecosystem of providers

Digital Realty has invested a significant amount of capital to build a worldwide platform and support our customers’ footprints across all major metros and edge markets, including high-count fiber access to a connected data community of service providers.

We continually seek partnerships with new service providers in Asia Pacific and globally to expand our offerings and strengthen our platform to support our customers’ growth and innovation:

Frequently Asked Questions (FAQs) About AI and Data

Where is AI data stored?
AI data are often stored in data centers, from traditional centralized data centers, cloud storage platforms, edge data centers, to on-premises servers. Additionally, AI data may also be stored in distributed databases or data lakes to facilitate efficient access and processing.

Will AI require more data centers?
As AI applications become more prevalent and data-intensive, it is likely that the demand for data centers will increase. Data centers play a crucial role in supporting AI applications that require real-time processing and low-latency interactions.

What is meant by data in AI?
In the context of AI, data refers to the information used to train machine learning models, make predictions, and perform various tasks. This data can take many forms, including structured data (e.g., databases), unstructured data (e.g., text, images, videos), and semi-structured data (e.g., XML, JSON). High-quality, relevant, and diverse data is essential for training accurate and robust AI models.

What's the power of data?
The power of data lies in its ability to provide insights, inform decision-making, identify patterns, and drive optimization across various domains. By harnessing the power of data through advanced analytics and AI techniques, organizations can gain a competitive edge, improve operational efficiency, enhance customer experiences, and unlock new opportunities for growth and innovation, which is why data centers are taking centre stage.

1.Digital Realty-sponsored IDC InfoBrief “Unlocking Innovation and Efficiency: Harnessing the Power of Hybrid Cloud and Data”, #AP241438IB, August 2023

Fortaleza Data Centers