The Explosive Growth of Global Data and Its Infrastructure Demands

GitHub Copilot · Data Infrastructure and Generative AI · By Super Admin

Global data creation has grown explosively, reaching about 149 zettabytes in 2024, driven by smartphones, social media, video streaming, AI content, and IoT devices, requiring massive infrastructure and energy investment.

The sheer volume of data being created worldwide underpins everything AI tools like GitHub Copilot rely on, and understanding its scale helps explain both the opportunities and constraints facing these systems. In 2024, global data creation hit about 149 zettabytes (ZB), or roughly 0.4 zettabytes per day. For comparison, global data creation was about 64.2 zettabytes in 2020 and only 6.5 zettabytes in 2012. A zettabyte itself is one trillion gigabytes, or one sextillion bytes—streaming Netflix continuously for 36 million years would produce a zettabyte of HD video.

Drivers of Data Growth

Several factors are fueling this rapid increase. There are more than 7 billion smartphones in the world, representing about 90 percent of the population, and each uses about 20 GB of data monthly. Social media platforms generate massive amounts of text, audio, and video content, with more than 5 billion people active on these platforms; even smaller platforms like Snapchat see users share more than 500,000 photos every minute. Video streaming, especially 4K video, consumes large amounts of data and represents 82.5 percent of global Internet traffic. AI-generated content is also a growing driver, estimated at 57 percent of content and potentially reaching 90 percent by 2026. Internet-of-Things (IoT) devices—including home wearables, industrial sensors, and smart appliances—are estimated to produce 80 ZB of data by 2025.

Infrastructure and Energy Investment

Handling this data growth requires major investment in infrastructure such as datacenters. The biggest players—Microsoft, Meta, Google, and Amazon—indicated in 2025 that they will spend $215 billion on IT infrastructure, up 45 percent year-over-year. This growth also demands significant energy investment: in 2023, datacenters accounted for 4.4 percent of total U.S. electricity according to the Department of Energy, with estimates rising to 6.7 to 12 percent by 2028. To manage this, large tech companies strategically locate datacenters near energy sources like solar or wind power stations, and there has been renewed interest in nuclear energy—for example, Microsoft's November 2024 partnership with Constellation Energy to restart the Three Mile Island reactor, with a $1.6 billion investment.

Data as a Strategic AI Asset

Data is increasingly described as the "new oil," functioning as a highly strategic asset core to AI development, since data is used for training models. For the most advanced models, training datasets can be enormous, as shown by the evolution of Meta's open-source Llama model: Llama 1 (February 2023) used 1.4 trillion tokens, Llama 2 (July 2023) used 2 trillion tokens, and Llama 3 (April 2024) used 15 trillion tokens. However, quantity alone is not sufficient—data also needs to be high quality, which is critical for tools like GitHub Copilot as they are used for more critical applications requiring high levels of accuracy and relevancy.

Key takeaways

  • Global data creation reached about 149 zettabytes in 2024, up from 6.5 zettabytes in 2012.
  • Smartphones, social media, video streaming, AI-generated content, and IoT devices are key drivers of data growth.
  • Major tech companies are investing $215 billion in IT infrastructure in 2025, including nuclear energy partnerships to power datacenters.
  • Data quality, not just quantity, is critical for training AI models and for tools like GitHub Copilot used in critical applications.
0.0 (0 ratings)

Discussion (0)

No discussion on this article yet — be the first to comment.

Please sign in to join the discussion.