Corning predicts three major trends in data center development in 2024

January 23, 2024
Latest company news about Corning predicts three major trends in data center development in 2024

In 2024, artificial intelligence (AI) will continue to drive the development of energy consumption and computing power in data centers, and bring new business models to the segmented market of data centers. Michael Crook from Corning recently released his predictions on the development trends and industry of data centers in 2024.


Whether it's daily apps such as mobile banking and social media, or cutting-edge emerging technologies such as artificial intelligence (AI) and immersive gaming, data centers, as invisible engines, have been silently driving the evolution of all these applications. With more and more applications migrating to the cloud and the wave of artificial intelligence, the energy consumption, cooling, and security requirements of data center operations have also changed.


Looking ahead to 2024, we believe that data center operators should pay attention to the following three prominent trends:

1. Artificial intelligence will continue to drive change and innovation


In last year's forecast, we saw that compared to traditional data centers, artificial intelligence and machine learning have a higher demand for power density, which is three times that of traditional data centers. The specific processing mode of the Large Language Model (LLM) requires a large number of fiber optic connections, while also placing higher demands on equipment power supply and cooling.


Artificial intelligence will continue to be one of the driving forces behind the development trend of data centers in 2024. We will continue to see artificial intelligence and machine learning (ML) widely deployed in data centers, especially in large-scale data centers, which requires better optimization management of energy consumption and resources. In addition, as companies continue to build large language models, they also need to establish new inference networks for prediction through analyzing new datasets, which require greater throughput and lower latency.


At present, the vast majority of work in artificial intelligence is built on these large language models, such as OpenAI's GPT-3 and GPT-4, Meta's LLaMA, and Google's PaLM2. Running billions of calculations during the development process of these models is essentially teaching them the knowledge they need. Now, most of the major participants in this field are beginning to shift towards utilizing these models to extend specific applications. This will once again change the demand for energy consumption and computing power in data centers.


What kind of outcome may the transition from development to reasoning bring? This will promote the development of edge computing. With the development of specific artificial intelligence applications, enterprises will seek processing capabilities that are closer to the location of application usage. They will seek smaller data centers to bring heavy computing closer to actual usage areas (such as manufacturing parks, universities, hospitals, etc.).

2. Multi-tenant data center space will usher in a shining moment


Usually, large-scale operators design and build the largest data center parks. However, as the energy consumption and space requirements to support artificial intelligence, machine learning, and other emerging applications continue to increase, large-scale data center operators may need to research alternative methods to build various facilities.


This brings development opportunities for multi tenant data centers (MTDCs). These operators have both development capabilities similar to real estate companies, as well as technical capabilities (they own the site), and they know how to meet power and cooling needs. So, in areas with limited space and energy consumption, multi tenant data centers are a good choice when ultra large scale operators need to operate facilities.


Enterprise level users also hope to leverage these emerging technologies, but building data center facilities is a significant capital investment. For example, multi tenant data centers and other new "cloud" service providers provide "artificial intelligence cloud services" by leasing dedicated server space to an organization (regardless of its size) to run artificial intelligence computing tasks.


As enterprises will seek computing power closer to the application deployment location, multi tenant data centers will also play a role in the rise of edge computing.

3. The advancement of optical modules helps data center operators maximize space utilization


The adoption of various new technologies requires data centers to generate exponentially increasing computing power and transmit more data faster. Given space and power supply limitations, operators are well aware that simply adding more fiber optic interconnections to meet these needs is an unsustainable strategy.


Especially for ultra large data centers, operators have started deploying 800G fiber optic transceivers to support applications, and it is possible to see some 1.6TB prototypes in 2024. High performance computing applications such as artificial intelligence and machine learning are driving the deployment of 800G fiber optics. The latest network switch used for interconnecting artificial intelligence servers within data centers, supporting 800G interconnection. In many cases, the optical module ports on these network switches operate in branch mode, where the 800G line is divided into two 400G lines or more 100G lines. In this way, data center operators can improve the connectivity of switches and interconnect more servers. When we discover the upgrade of fiber optic transceivers, which means that optical wavelengths and fibers can carry more data, we will also find that we will use optical modules with fewer connections and higher speeds during operation, reducing cable congestion in the rack and improving air circulation, benefiting data center customers.


The advancement of fiber optic technology has made it possible for fibers and wavelengths to carry more data. A typical multimode 400G SR8 fiber optic transceiver is equipped with 16 fiber optic connections, suitable for short distance applications, but the 400G SR4 optical module (reducing the number of fibers to 8) is entering the market. These optical modules, as well as other new types of optical modules, play a significant role in helping data centers meet the growing demand for data.


Related to this trend is the progress in connector miniaturization, and the development of solutions such as ultra small connectors will help data center operators make greater use of limited space.




CIOs and CTOs should grasp these emerging trends to ensure that data centers can support emerging business processes and new use cases. Just to keep up with rapidly developing technology, piecing together individual component solutions may be tempting, but a comprehensive engineering solution that meets the current and future data needs of customers will always be a more powerful strategy.


This article refer to:



tag: data center, AI, 400G, media converter, optical module, fiber optic transceivers