Most people interact with cloud storage every day without thinking about what it actually is, where it lives, or what it takes to run. A photo uploaded to iCloud, a video backed up to Google Drive, a gallery delivered through a client platform - all of these are stored somewhere physical, in a building, on hardware, drawing power.
The scale of that physical infrastructure in 2026 is difficult to comprehend without numbers. Here are the numbers.
The Scale of the Cloud
By 2026, the amount of data stored in cloud environments is expected to exceed 200 zettabytes globally.
To put that in context: one zettabyte is one trillion gigabytes. A single 4K wedding film takes up roughly 15GB. 200 zettabytes is enough to store approximately 13 trillion of those films.
Around 94% of enterprises worldwide use some form of cloud service, including storage, analytics, or infrastructure platforms. The shift from companies owning their own servers to renting capacity from cloud providers happened faster than almost anyone predicted.
Synergy Research Group reports 1,360 hyperscale data centers in operation at the end of 2025, now holding 48% of total data center capacity, while enterprise on-premises has fallen to 32%, down from 56% in 2018. By 2031, hyperscale is expected to reach 67% of capacity.
Who Runs the Infrastructure
With 4,011 data centers as of March 2026, the US has by far the most globally. The UK ranks second with 511, followed by Germany with 507, China with 368, and France with 344.
Amazon, Alphabet, Microsoft, Meta, and Oracle are collectively forecast to exceed $600 billion in capital expenditure in 2026 - a 36% increase over 2025. Roughly $450 billion of that spend is directly tied to AI infrastructure: servers, GPUs, data centers, and related equipment.
Dell'Oro Group forecasts that full-year 2026 global data center capital expenditure will surpass $1 trillion for the first time in history.
One trillion dollars, in a single year, building and upgrading data centers. The infrastructure powering the modern internet is the largest construction project in human history, measured by investment.
The Energy Problem
Data centers consumed 183 terawatt-hours of electricity in 2024 - over 4% of total US national electricity consumption - and the IEA projects that figure will grow by 133% to 426 TWh by 2030.
Large data centers use up to 5 million gallons of water per day for cooling purposes. A single hyperscale facility in a dry climate can consume as much water as a small city.
The energy problem is getting harder because of AI. The AI workload surpassing non-AI workloads in 2026 - with AI infrastructure forecast at 44 GW versus 38 GW for traditional computing - represents a genuine inflection point in what data centers are and what they are for. A standard server rack might draw 5–10 kW. An AI training cluster with dense GPU configurations can draw 50–100 kW per rack or more.
What Drives the Growth
The data storage growth curve isn't driven by people storing more photos. It's driven by three structural forces:
AI training datasets. Generative AI workloads significantly increased cloud storage demand, with over 70% of enterprises expanding storage capacity in 2024 to support AI training datasets.
Video. Video is the fastest-growing category of data by volume. A single hour of 4K video at broadcast quality is roughly 50GB.
IoT. By 2025 there were an estimated 75+ billion IoT devices in use worldwide generating data - much of which feeds into cloud platforms for storage and analytics.
The Object Storage Revolution
Most of the world's data - including every photo in a client gallery, every video in a delivery platform - is stored using object storage.
Object storage treats data as discrete objects with unique identifiers, rather than files in a folder hierarchy. This makes it massively scalable, highly durable (data is automatically replicated across multiple locations), and cost-effective at scale.
S3 - Amazon's Simple Storage Service, launched in 2006 - pioneered this model and became the de facto standard. More than 75% of enterprises now use object-based storage systems in cloud environments.
The durability of well-architected object storage is measured in "nines" - a system with eleven nines of durability (99.999999999%) expects to lose one object per trillion stored per year.
The Security Reality
According to Uptime Institute's 2024 Global Data Center Survey, three out of four data center operators have experienced a cybersecurity incident in the past three years. Over a third of those had serious or severe impacts.
For photographers and videographers using cloud delivery platforms, the protections that matter are: encryption in transit (TLS), encryption at rest, and access control that ensures clients can only access galleries they've paid for.
What 2TB Actually Means
For context on what 2TB of cloud storage means in practice:
- Average wedding gallery (photos only, full resolution): 15–25GB
- Average wedding highlight film (4K broadcast quality): 10–20GB
- Full wedding package (photos + highlight + reception edit): 30–50GB
- Complete wedding deliverables in 2TB: approximately 40–65 events
A photographer shooting 30 weddings per year and archiving galleries for 2+ years is working with 1.5–3TB of active storage. The 100GB caps common in legacy gallery platforms aren't sized for how professional photographers work in 2026. They're sized for how photographers worked in 2015.
DAT Drives uses S3-compatible object storage on OVHcloud infrastructure. 2TB per Professional account. Files stored at full quality, accessible permanently, delivered securely.
