TABLE STAKESDATA

Guest post A list from LinkedIn, given more room

What AI is doing to the data business

Daniel Goldberg has been in alternative data since before it had the name. He built the first data products sourced from email receipts, founded Alternative Data Analytics and later sold it, and now advises data, analytics and AI companies on how to sell what they have. Last week he posted a list on LinkedIn answering the question he says he gets most: what is AI doing to the data business? I asked him if I could give it more room here. Here is his list, with the thinking behind each item.

Daniel Goldberg · Guest post · October 2026

Disclosure: Daniel is the outsourced Chief Commercial Officer of Financial AI, which sells AI-generated financial forecasts, and he advises data companies on monetization.

Recently I have been asked the same question many times: what is AI doing to the data business? I have spent a lot of time lately with data owners, buyers, marketplaces and AI platforms. Here is what I am seeing firsthand.

1. New data businesses are being created by companies that never viewed their information as a product

For years the sellers of data were the companies built to sell it. That is changing. Companies whose information was a byproduct of running the business are now asking what it is worth and who would pay for it. Demand from these first-time data owners is the highest I have seen.

2. Data diligence has become more forensic

Buyers used to ask whether a dataset was predictive. Now they ask where every row came from, who had the right to collect it, and what the data looked like on the day a decision would have been made. Governance, lineage and point-in-time accuracy are under more scrutiny than ever. A dataset that cannot answer those questions does not get past the first review.

3. Consumption-based pricing is beginning to replace traditional seat licenses

When the user is a model or an agent rather than an analyst at a desk, a per-seat license stops making sense. Pricing is moving toward what is actually consumed: queries, records, calls. That changes how a data business plans its revenue and how it sells.

4. AI-generated financial forecasts are increasingly challenging consensus expectations

Sell-side consensus has been the reference point for a long time. Forecasts generated by AI from large volumes of data are now differing from consensus often enough that investors are paying attention. They are also more timely. A sell-side estimate is revised when the analyst gets to it. An AI-generated forecast can be updated daily as new data arrives. Every one of those forecasts runs on data, which creates demand for it.

5. AI labs have emerged as serious new buyers of enterprise data

The labs building foundation models need data they cannot get from the open web: licensed, documented and clean. They have budgets, and they are buying. For a data owner, that is a buyer who did not exist a few years ago.

6. Competitive advantage is moving toward transforming unstructured content into structured intelligence

Most of the world's information is text, documents, transcripts, images and filings. The edge is shifting from owning a structured dataset to being able to turn unstructured content into structured, decision-ready intelligence, reliably and at scale.

7. Prediction markets are creating entirely new demand for data and signals

A prediction market needs something to price and something to settle on. Both create demand for data and signals, some of which had no market before.

The value of data is no longer determined by ownership alone.

It seems that the value of data is no longer determined solely by ownership. Increasingly it is determined by how effectively that data can power AI-driven outcomes. If you own data or buy it, the playbook remains fluid.