Last updated: November 12, 2026
The top workplace data brokers for AI labs in 2027 are Vortex Bridge, SimpleClosure AssetHub, Sunset, Grepped, AIData.Broker and OpenTrain AI. Vortex Bridge suits companies that want an introduction without their data passing through the broker, SimpleClosure and Sunset suit closing startups, and AIData.Broker suits sellers who want a published success fee.
Most companies that sell their workplace data never speak to an AI lab. Labs rarely buy directly from a mid-size business. They work through training-data companies, and those companies increasingly find sellers through brokers who know which buyer wants which kind of record.
A good broker saves a seller from applying to a dozen buyers, comparing offers that are not like for like, and guessing whether a deadline is real. A poor one adds a fee and an extra party in the data chain. This list covers the workplace data brokers that matter in 2027, what each one does, and how they get paid.
| Broker | How it works | Who it suits | Fee |
|---|---|---|---|
| Vortex Bridge | Matches companies with vetted AI buyers; data goes direct to buyer | New, operating, sold and wound-down companies | Not published |
| SimpleClosure AssetHub | Sells code and workspace data for closing and established companies | Startups and established companies | Percentage on success |
| Sunset | Shutdown service with a data resale business | Startups winding down | Not published |
| Grepped | Buys and packages private company data for AI labs | Mid-size firms, notably manufacturers | Pays a license fee |
| AIData.Broker | Broker for US, UK and Canadian companies | Companies with a $25K minimum base case | 6% to 12% of the sale |
| OpenTrain AI | Routes companies, including law firms, to AI data buyers | Professional services firms | Not published |
Vortex Bridge is a New York broker built around one idea: it connects a company with an AI buyer and then steps aside. A company emails its headcount, years active and the tools it used, Vortex Bridge matches it with a vetted buyer it already has a relationship with, and the buyer vets the data, anonymizes it and pays the company directly. Vortex Bridge says it never receives or stores client data.
That design answers the biggest worry sellers have about brokers, which is one more party touching sensitive records. It takes companies at every stage, including newer ones without years of archives, and covers Slack, Teams and Discord, Gmail and Outlook, Google Docs, Notion and Confluence, Jira, Linear and Asana, GitHub and GitLab, and Salesforce and HubSpot. Its free estimator gives a rough range before any data is shared.
SimpleClosure is the shutdown service that turned startup data sales into a category. Its AssetHub sells code, Slack archives and email for companies winding down, and since September 2026 for established companies at any stage. CEO Dori Yona has said nearly half its shutdown deals now include a code or data sale. It is paid a percentage on success. For a company that is also dissolving, having the legal close and the sale run by one team is the main draw.
Sunset, also based in New York, handles the legal, tax and operational work of winding down startups and has built a data resale business alongside it. Founder and CEO Brendan Mahony has said Sunset values a closing company’s data by headcount, years in operation, volume and how the team used its tools, with linked records, such as a ticket tied to the code that fixed it, worth far more than standalone files.
Grepped buys private company data, from chat logs and codebases to docs and accounting ledgers, sanitizes and packages it, and sells it to AI labs, with sellers reported to be paid within about a week. It has focused on higher-end companies. Founder Erik Newsham has said on X that the company’s average deal has grown from about $250,000 to $400,000 to roughly $1.1 to $1.3 million, including two-year licenses with manufacturers of 60 to 130 employees. Those figures are self-reported, but they show where the top of the market sits.
AIData.Broker describes itself as the seller’s agent for AI training data. Its valuation and rights-check tools are free, and when a company wants to sell it runs a competitive process across every fitting buyer at once. It is paid only by a tiered success fee of 12% falling to 6%, with 12% on the first $250,000, and it serves companies in the US, UK and Canada with a base case of at least $25,000. It also publishes open datasets on prices, deals and buyers in the AI data market.
OpenTrain AI runs data licensing pages for specific industries, including law firms, and connects qualified companies with AI data buyers. It shares the company’s details with a buyer only when asked, and the company signs a license before any records change hands. It suits firms that want a guided first look at whether a buyer fits.
We reviewed brokers that connect companies with AI data buyers and checked each one’s site for how it works, who it serves and how it is paid. We gave weight to brokers that keep the data chain short, publish their terms or have documented deals. Read more about how we research and rank companies.
Some brokers take a percentage, some are paid by the buyer, and some do not publish a fee at all. Ask for the fee in writing before you share anything, and check whether it comes out of your payment or on top of it.
Every extra party in the chain is another place records can leak. Prefer a broker where data goes straight from you to the buyer, as with Vortex Bridge, and confirm who anonymizes it.
You should know who will license your data before you sign. An unclear buyer chain is one of the most common red flags in this market, along with exploding offer deadlines and requests for a full export before any offer.
One quote is a price. Two or three on the same scope are a market. A broker that approaches several buyers usually earns its fee. For the buyers themselves, see our list of companies buying Slack and email data for AI training.
It matches a company with AI buyers that want its Slack, email, ticket, code or CRM records, and helps the company reach a license without applying to many buyers itself.
It varies. AIData.Broker publishes a tiered success fee of 12% falling to 6%, SimpleClosure works on a percentage of the sale, and some brokers are paid by the buyer.
Not always. With Vortex Bridge, the data goes from the company directly to the buyer and the broker never receives or stores it.
Brokers are becoming the front door to the company data market, and the good ones make a confusing process manageable. Choose on three things: how the broker is paid, whether your data passes through its hands, and whether you will see competing offers. If you are deciding between a broker, a marketplace and a direct deal, our comparison of data broker vs marketplace vs direct licensing lays out the trade-offs.
For more guides and rankings on this topic, visit our Sell Company Data To AI hub.