We frequently hear that we have a data scarcity problem in AI. And when it comes to unique, Tolstoy-level literature and the like, that could be true.

But in the business world I think we have the exact opposite problem.

What we actually have is a, "There's absolutely nobody to look at 99.999% of our data" problem.

According to IDC's Data Age 2025 report, we're generating 149 zettabytes of data annually. That's 149 trillion gigabytes. Every single year.

Here's what's actually happening to it:

Surveillance and security:

• 1+ billion cameras worldwide generating 5.5 million terabytes per day (Grand View Research 2024)
• 95-99% of footage is never viewed unless an incident is reported (Grand View Research 2024)

IoT and industrial sensors:

• 21.1 billion connected devices producing 79.4 zettabytes annually (IoT Analytics 2024)
• 30-50% gets discarded immediately at the edge (IoT Analytics 2024)
• Of what remains, less than 5% is ever analyzed (IoT Analytics 2024)
• 99% of IoT data is lost before reaching decision-makers in industrial settings (McKinsey Digital)

Enterprise operations:

• Machine logs and telemetry: 80-90% never examined (Coralogix 2024)—just stored for compliance
• More than 90% of observability data is never read (Coralogix 2024)
• Documents: 41-80% never accessed after creation (NetApp 2024)
• Communications and emails: 85-91% never meaningfully consumed (Veritas Global Databerg Report)

As far as totals, we're talking about 149 zettabytes generated globally each year, only 12-15% is ever examined by humans or AI (IDC Data Age 2025). That's roughly 20 zettabytes.

My takeaway ​

As far as I'm concerned, yes, there might not be new high-quality literature being generated, which I suppose is a problem. Maybe we're running out of that, and I don't know where we're going to get more.

But practically speaking, I think the bigger problem is that businesses, companies, and people are generating I guess zettabytes of data, and nobody is actually looking at any of that data or at least a small percentage.

To me, this presents an extraordinary opportunity for AI to actually give us visibility and the ability to extract insights from all this data that nobody is looking at.

So sure, we have a data problem, but not the one that people think.