The Future of Datasphere: An Intersection of AI and the Metaverse

NNicholas October 5, 2023 8:42 PM

This article delves into the intersection of AI and the metaverse, and how they'll impact the datasphere. It explores the evolution of AI, the demands of the metaverse, and the response of data centers worldwide.

Blooming AI and the Computational Challenge

The second decade of the 21st century witnessed significant strides in the field of AI. Large language models (LLMs) have successfully elevated AI performance to near-human levels, and in some cases even beyond. However, this success comes at a cost. The resources required to train and operate these LLMs are growing at an alarming rate. Initiatives like DeepMind’s AlphaGo Fan, which defeated a professional Go player, mark the beginning of this large-scale era. Since then, the computational requirements have skyrocketed. For instance, OpenAI’s GPT-4, launched in 2023, needed 55 thousand times more computational resources than AlphaGo Fan. With this rapid pace of growth, future AI systems will demand exponentially larger computers.

Metaverse and its Impact on Data Demands

The metaverse is an integrated, immersive web accessed through augmented and virtual reality (AR and VR) tech. This novel paradigm is expected to add significantly to data requirements. Data transmittance, measured via bitrates, is calculated to be much higher for AR, VR, and holograms, all integral to the metaverse. These technologies top out at 300 Mbps. Additionally, AR and VR demand extremely low latency (less than five milliseconds) to prevent motion sickness. This means that not only will the metaverse increase the amount of data transferred (an estimated 644 GB per household per day), but it will also need to do so swiftly.

Evolving Data Centers: Bigger or Decentralized

With the exponential growth in data, data centers across the globe are evolving to meet the demand. There are two primary strategies in play. The first approach is to build massive, hyperscale data centers. Examples include the Inner Mongolia Information Park and the Hohhot Data Center in China, and Switch’s Citadel Campus in the U.S. The second strategy is to decentralize data centers via edge computing, which improves latency. This approach is particularly useful in applications where speed equals safety, such as self-driving vehicles. Investors are recognizing this trend, with global spending on edge data centers projected to reach $208 billion in 2023, marking a 13.1% increase from 2022.

As we stand on the cusp of the zettabyte era, data production is expected to surge. According to the International Data Corporation, the annual data output will reach an astonishing 221 zettabytes by 2026. This growth, at a compound annual rate of 21.2%, underscores the critical role data centers will play in the coming years. Companies like HIVE Digital, which exports renewably sourced computing power, are stepping up to meet the demands of burgeoning technologies like AI.

More articles

Also read

Here are some interesting articles on other sites from our network.