How Much Data Is Created Every Day? (2026 Figures, Corrected)

The Data Deluge: How Much Data Is Created Every Day in 2026?

The widely repeated 221 zettabytes for 2026 is an older forecast-era figure, not a current measurement. It was a forward projection made four years before the year it describes, not a measurement of what actually happened. Treating it as a current 2026 statistic is the single biggest error in most of the pages currently ranking for this topic.

The more recent, verifiable picture is different. IDC's live forecast product covers 2025–2029 and was published in January 2025; a further IDC-derived update, summarised by Exponential Tracker in June 2026, puts the 2025 Global DataSphere at 218.4 zettabytes and projects 718.5 zettabytes by 2030. No official 2026 point estimate is quoted here; the latest verified figures available are 2025 at 218.4 ZB and 2030 at 718.5 ZB.

Dividing IDC's 2025 annual Global DataSphere forecast of 218.4 zettabytes by 365 yields a rough daily equivalent of about 598 exabytes. As noted above, however, this is purely derived arithmetic applied to an annual total rather than an official IDC daily statistic, and every citation of a "daily" data number should carry that caveat.

Zettabytes and Exabytes: What the Global DataSphere Actually Measures

IDC defines the Global DataSphere as the amount of data created, captured, replicated and consumed each year across the world — a single annual total that folds together everything from a photo saved on a phone to a replicated backup snapshot to a video streamed to a smart TV. Crucially, it is a gross volume metric, not a measure of unique or original information; the same file copied across backups or streamed multiple times contributes repeatedly to the total.

Alongside IDC's January 2025 regional-analysis edition covering the same annual scope, that forecast lineage matters: the 2020 actual figure of 64.2 zettabytes, the 2025 figure (181 ZB under a March 2021 forecast path, revised to 218.4 ZB under the June 2026 forecast), and the 2030 projection of 718.5 zettabytes are all points on the same evolving curve — not competing or contradictory numbers, but successive forecast vintages of the same underlying series.

The Engines of Growth Behind the 2026 Datasphere

IDC's four-stage definition — created, captured, replicated, consumed — is broad enough to include AI training and inference traffic, IoT telemetry, and streaming replication within the same headline total. However, no verified source provides a reliable, dated breakdown of how much each category contributes; rather than repeating uncited percentage estimates, this study avoids circulating unsubstantiated claims — the same discipline applied to the legacy 221 ZB figure.

What is verifiable is the shape of the growth curve itself. IDC figures put the datasphere at 64.2 ZB in 2020 and 218.4 ZB in the June 2026 forecast for 2025, which is about 3.4 times higher — and IDC's June 2026 forecast projects it reaching 718.5 zettabytes by 2030, a further 3.3-times increase. A May 2022 IDC-linked market summary had already flagged that the datasphere would more than double between 2022 and 2026; the growth now visible in later forecasts is consistent with that early signal, even though the precise 2026 endpoint has since been revised.

The Unstructured Data Challenge: Volume Versus Value

A common question from IT leaders is what proportion of all this data is actually analysed versus left untouched as so-called dark data. As with the category splits noted above, none of the sources verified for this piece publish a dated ratio for analysed versus dark data, so we decline to repeat circulating, uncited percentages.

What the IDC framework does confirm is that "replicated" and "consumed" data sit inside the same annual total as "created" and "captured" data — meaning a meaningful share of any headline zettabyte figure, whether 218.4 ZB or 718.5 ZB, represents duplicates or ephemeral consumption rather than unique corporate records that businesses query, report against, or govern directly. For a buyer, the practical implication is that raw volume growth says nothing about how much of that data is structured enough to be actionable; closing that gap is a classification and tiering problem before it is a capacity problem. Use our storage solution finder to match the data actually held to the tier it deserves, rather than storing everything at the same cost and durability level.

Impact on the UK: Real Costs Behind the Global Curve

Global zettabyte totals are abstract; data-centre power bills are not. Rising datasphere volumes directly expand physical storage and compute footprints, driving up facility power draw and cooling demands. For UK colocation buyers, commercial energy pricing is the headline operational constraint: wholesale UK electricity averaged 9.8p/kWh in Q1 2026 and 10.0–11.5p/kWh across May–June 2026, while average commercial electricity including CCL sat at 24.14p/kWh. In turn, a 2026 benchmark cited standard racks at £180–£260 per month and high-density AI/GPU racks at £230–£340 per month.

On 30 June 2026 the UK government published an official special-feature analysis of data-centre electricity consumption in Great Britain covering 2020 to 2024 — the clearest current government-sourced baseline for how much of the country's power the sector already draws, and a useful anchor for any UK-specific footprint discussion.

By way of a household comparison — though commercial facilities negotiate dedicated power tariffs and do not pay domestic capped rates — Ofgem’s domestic electricity price cap sat at 26.11p/kWh through the second half of 2026 (translating to roughly £870 annually for a typical 2,700 kWh home). That domestic figure provides consumer context, but enterprise colocation budgets are anchored by commercial and wholesale power.

From Volume to Value: Managing, Securing and Extracting Value

Rising data volumes only become a business asset once they are matched to the right storage economics and the right governance discipline. Four practical moves follow directly from the numbers above:

• •Tier by access pattern, not by habit — understand storage tiering strategies before defaulting everything to premium flash.
• •Model growth explicitly rather than guessing capacity headroom — calculate your backup and DR needs against realistic five-year volume trajectories, not the legacy 221 ZB forecast.
• •Extend hardware life where durability allows it — consider refurbished storage options for lower-tier or archival capacity rather than buying new for every terabyte added.
• •Treat classification as the bottleneck, not capacity — the datasphere's growth is outrunning most organisations' ability to say what proportion of their own holdings is dark data, so governance investment should scale alongside storage spend, not lag behind it.

The Hidden Costs: Economic and Environmental Footprint

The clearest measurable footprint proxy available here is the UK government's own analysis: the 30 June 2026 special feature on data-centre electricity consumption in Great Britain, covering 2020 to 2024. It is the official reference point for discussing how the sector's power draw has moved, rather than an estimate from a market report.

Cost exposure sits on top of that footprint. Wholesale electricity moved from 9.8p/kWh in Q1 2026 to 10.0–11.5p/kWh by May–June 2026 — a swing that feeds directly into colocation billing, which is one likely reason AI/GPU racks (£230–£340) command a premium over standard racks (£180–£260) in current UK pricing. As datasphere volumes climb toward the 718.5 zettabyte 2030 projection, the electricity line of an infrastructure budget is no longer a rounding error.

Beyond 2026: Data Growth Trends to 2030

IDC’s live forecast product covers 2025–2029, and an IDC-derived tracker summarises a June 2026 forecast path that reaches 718.5 ZB by 2030 — more than 3.3 times the 218.4 zettabyte figure now forecast for 2025. That endpoint supersedes the legacy 221 ZB number still attached to 2026 on many competing pages; anyone quoting a 2030-or-beyond datasphere figure should check which forecast vintage it came from before repeating it, exactly as this piece has done for 2026.

Sources

Every figure in this article traces to the sources below.

• •IDC — Global DataSphere 2025–2029 forecast product
• •IDC — Global DataSphere regional forecast update, January 2025
• •IDC — Global DataSphere research page, updated August 2026
• •MarketResearch.com — IDC-linked 2022 forecast summary (2022–2026 doubling claim)
• •Exponential Tracker — IDC-derived Global DataSphere historical and forecast series
• •Ofgem — Energy price cap unit rates and standing charges
• •GOV.UK — Energy Trends June 2026: data-centre electricity consumption in Great Britain 2020–2024
• •GOV.UK — Quarterly Energy Prices, June 2026
• •Habo Energy — AI data centres and UK electricity bills, 2026
• •ArmaSon — UK data-centre power cost benchmarking, 2026