How Much Data Is Created Every Day? (2026 Statistics)

The world generates roughly 402.74 million terabytes of data every day — about 0.4 zettabytes — according to Statista's latest estimate. That daily flood adds up to a global datasphere on track to hit 394 zettabytes a year by 2028, more than double the 2024 figure. This page rounds up the most current, sourced numbers on how much data we create, where it comes from, how much of it is never used, and how AI is reshaping the curve.

How much data is created every day in 2026?

The most widely cited figure comes from Statista, which estimates that around 402.74 million terabytes of data are created, captured, copied, and consumed each day. That is the equivalent of roughly 0.4 zettabytes per day, or about 147 zettabytes over a full year of daily generation. For scale, one zettabyte is one trillion gigabytes, so a single day's output equals hundreds of billions of gigabytes.

A few ways to picture the daily total, using Statista and Domo's "Data Never Sleeps" estimates:

• 347 billion+ emails are sent every day. (Venngage/Domo via Exploding Topics)
• Google processes on the order of 14–16 billion searches daily. (Exploding Topics)
• Hundreds of millions of social posts, photos, and short-form videos are uploaded daily across the major platforms. (DemandSage)

These figures are best read as latest-available estimates (2024 baseline data, carried into 2025–2026 reporting) rather than precisely metered counts — no single body audits every byte. But the order of magnitude is consistent across Statista, IDC, and Domo.

How big is the global datasphere by year?

Statista and IDC both track the total annual volume of data created worldwide in zettabytes. The trajectory is steep and accelerating: the datasphere roughly tripled between 2018 and 2024, and is projected to more than double again by 2028.

The 182-zettabyte estimate for 2025 means the world stored, copied, and consumed more data in that single year than existed in total just a few years prior. IDC's separate Global DataSphere Forecast, 2025–2029 reaches a similar conclusion, projecting the datasphere to roughly triple across that five-year window, driven by structured data from automation and unstructured data from AI and media.

How fast is data growing?

Annual data creation has been compounding at a high double-digit rate for over a decade. The volume climbed from around 2 zettabytes in 2010 to roughly 182 zettabytes in 2025 — about a 74-fold increase in 15 years, per Statista-based analysis. The most quoted shorthand for this pace: roughly 90% of all the data in the world was generated in the last two years, a ratio that holds true precisely because the curve is exponential.

The replication factor matters too. A large share of the datasphere is copied rather than newly created — the same video streamed to millions of devices, the same file backed up across clouds — so "data consumed" runs well ahead of unique, original data. Statista tracks both, and replicated data has been outpacing unique data as cloud and streaming scale.

Where does all this data come from?

The bulk of created data is unstructured media, with video dominating by a wide margin. Internet traffic breakdowns from Sandvine put video at roughly four-fifths of all bytes moving across networks.

Beyond consumer streaming and social feeds, machine-generated data is the fastest-growing source:

• Internet of Things (IoT): There were an estimated 21.1 billion connected IoT devices at the end of 2025, up from 18.5 billion in 2024, and the count is projected to reach 39 billion by 2030. (IoT Analytics) The data thrown off by these devices is forecast at roughly 79.4 zettabytes in 2025. (Statista)
• Social and user-generated content: billions of posts, images, and clips per day across platforms feed the unstructured pile. (DemandSage)
• AI systems: training corpora, model outputs, logs, and synthetic datasets are now a meaningful and rapidly expanding data category (more below).

How much data goes unused?

Despite the staggering volumes, most data is never analyzed. Seagate's Rethink Data Report found that only about 32% of data available to enterprises is put to work — leaving roughly 68% unleveraged. Other estimates of "dark data" (information that is collected and stored but never used for decisions) range from about 55% to as high as 80%, depending on the study and definition.

• Nearly 1 in 3 organizations say 75% or more of their stored data is dark or obsolete. (DataStackHub)
• In one survey of 1,300+ IT and business leaders, 60% said more than half of their data is dark data. (Splunk, State of Dark Data)

The gap matters financially and environmentally: storing data that never gets used still costs money, energy, and carbon, and it widens the security surface. As IoT and AI push volumes higher, the unused share is widely expected to keep growing unless governance improves.

How is AI changing data creation?

Generative AI is doing two things at once: consuming enormous volumes of training data and manufacturing vast amounts of new data. Synthetic data — information produced by algorithms rather than captured from the real world — is the clearest signal of this shift.

• Gartner projects that by 2028, around 80% of AI training data will be synthetic, up from a small single-digit share just a few years earlier. (Gartner via World Economic Forum)
• Some estimates suggest more than 60% of data used for AI applications was already synthetic as of 2024. (World Economic Forum)
• Generative AI adoption is mainstream: one 2025 mid-market survey found 91% of firms using generative AI, up from 78% a year earlier. (HatchWorks)

This is why IDC and Statista both expect the curve to steepen rather than flatten: every AI interaction generates prompts, responses, embeddings, and logs, while models increasingly generate their own training material. Much like the way high-frequency digital activity — from streaming to real-time online platforms and crypto-gambling sites — produces continuous data trails, AI workloads add a new, self-reinforcing layer to global data growth.

Frequently Asked Questions

Sources

• Statista — Volume of data created, captured, copied, and consumed worldwide
• IDC — Worldwide Global DataSphere Forecast, 2025–2029
• Exploding Topics — Data generated per day (Statista/Domo)
• DemandSage — Big Data Statistics 2026 (Sandvine traffic data)
• IoT Analytics — Number of connected IoT devices
• Seagate Rethink Data Report / dark data analysis
• Splunk — The State of Dark Data
• World Economic Forum — AI and the growth of synthetic data (Gartner)