Why do software vendors charge for data access?

Partly to cover real costs: running an API, supporting integrations, carrying load on a system that was built for its own users. Partly because access is where switching costs live, and a customer who can take its data out cheaply can connect other tools, or leave. How much each counts varies by vendor. A small business-software vendor supporting integrations on customers' own servers has real costs. A large platform metering bulk extraction is pricing access as a product. The cloud egress episode of 2024 shows how fees behave when a rule changes, though not, on its own, why they were set.

What the usual answer says

The pages that rank for this question explain access fees as covering infrastructure and support, or as a deliberate lock-in, and some note that cloud providers waived exit fees under new European rules. The newer theme is AI: platforms metering access by other companies' AI agents. What the pages don't give is the regulator's own evidence on how the waivers worked, or a checklist for a buyer.

What happened to egress fees

Egress fees are charges for moving data out of a cloud provider. A data law that came into force on 11 January 2024 required that switching charges "cannot exceed costs incurred by the provider of data processing services that are directly linked to the switching process concerned", and that they end entirely from 12 January 2027.

A competition authority's 2025 cloud market investigation records what followed, in its appendix on free switching programmes. Google announced free egress for customers switching on 11 January 2024, AWS on 5 March and Microsoft on 13 March, each applying worldwide. Microsoft said it acted in response to the law. The other two providers' stated reasons are redacted in the published appendix.

Each programme came with conditions. AWS required pre-approval and gave a credit to move stored data once. Microsoft gave customers 60 days and required them to cancel all their subscriptions. The investigation found that the providers "do not publicise the programmes through direct communication", though information was on their websites for customers who looked.

What the episode shows

It shows that a fee on leaving could be dropped worldwide within about nine weeks, on terms that limited its use. Our argument is that fees which behave like that are doing more than recovering cost. The investigation offers a counter-reading worth stating. Uptake was low, and it suggests that "the egress cost to customers for switching may not be a factor influencing the decision to switch." AWS's announcement also noted that "over 90 percent of our customers already incur no data transfer expenses" out of its cloud. A fee few customers pay, on an event few customers choose, is cheap to waive.

Business software shows the same instinct more openly. On a February 2026 earnings call, HubSpot's chief executive said of other companies' AI agents using its customers' data: "that access, we will monitor it, we will meter it and we will monetize it." The platform, the executive said, is "open by design, but we're not a free data pipeline for everybody".

For a timber business, the effect is practical. Software that holds scale tickets, inventory or accounts is the source for every report and every AI tool that might use that data, so access terms are part of the price.

What to ask before signing

Put data access on the price list alongside the licence. In what formats can every table be exported, and how often, at no charge? Is there an API, and what does it cost per user, per call or per connection? Can the business read its own database, read-only, or through a replica? What does leaving cost, including a full export of history and help mapping the fields? Will these terms hold through renewals? And if the vendor's own AI assistant becomes the easy way to query the data, will access for other tools still be allowed and priced?

Get the answers in the contract. The investigation's finding that the cloud programmes were not communicated directly suggests that terms not written down are hard to use.

When it doesn't apply

Some access fees reflect real cost, such as high-volume real-time feeds, custom integration work or support for integrations on a small vendor's on-premise product, and are fair to pay. Open-source and self-hosted systems usually leave the data with the business already.

Quarri for finance and strategy teams is built for the people who close the month, explain the margin and answer the board.

Quarri is an AI-native data platform for the timber supply chain. It connects buying, production, sales and inventory for forest management, sawmill, wood products and pulp, paper and packaging operators.