Why are static datasets unsuitable for AI training?

Static datasets quickly become outdated, failing to reflect real-world changes. This leads to model errors and reduced effectiveness in dynamic business environments.

How does the new infrastructure layer improve AI performance?

It provides real-time access to structured, contextual data, enabling models to make more accurate decisions and reducing the risk of incorrect outputs.

What challenges does RAG technology address?

RAG (Retrieval-Augmented Generation) allows models to fetch external data during queries, but many systems still struggle with data freshness and reliability.

← All news

Artificial intelligence

New Data Infrastructure Layer for AI Accelerates Development

June 24, 2026

Photo: MIT Technology Review

Quick answer

A new infrastructure layer enables AI systems to access real-time, structured web data, addressing critical gaps in model training and decision-making.

The next phase of artificial intelligence development may hinge on a new infrastructure layer designed to deliver real-time web data access. This layer must process hundreds of millions of domains and billions of new URLs appearing weekly, overcoming technical barriers to provide current information.

Bright Data CEO Or Lenchner highlights the vast volume of web data but notes that without proper infrastructure, companies cannot leverage it effectively. Previously, AI progress relied on scaling training data and models, but now the focus has shifted to the dynamic and unstructured nature of web data.

Modern AI systems require a steady flow of fresh data to analyze pricing, consumer sentiment, and market trends. The infrastructure must handle millions of simultaneous requests to sites with varying languages, formats, and access rules. Delays in data retrieval reduce model efficiency, while outdated information leads to incorrect decisions.

Access to real-time, high-quality data also minimizes AI 'hallucinations,' boosting user confidence. A survey revealed that 56% of AI professionals consider real-time data access critical for improving result accuracy. However, despite technologies like RAG, many systems still fail to deliver timely and contextually relevant information.

According to Gartner, by year-end, 60% of AI projects lacking access to structured, up-to-date data will be abandoned. This underscores the urgent need for robust web data infrastructure.

Common questions

Why are static datasets unsuitable for AI training?: Static datasets quickly become outdated, failing to reflect real-world changes. This leads to model errors and reduced effectiveness in dynamic business environments.
How does the new infrastructure layer improve AI performance?: It provides real-time access to structured, contextual data, enabling models to make more accurate decisions and reducing the risk of incorrect outputs.
What challenges does RAG technology address?: RAG (Retrieval-Augmented Generation) allows models to fetch external data during queries, but many systems still struggle with data freshness and reliability.

Dzen feed: /feed/dzen.xml · RSS: /feed.xml