The European Union’s Data Act entered into force on 11 January 2024. Among its many provisions — covering data sharing from connected devices, business-to-business data access, and protections against unfair contract terms — the ones that matter most to infrastructure teams concern cloud switching: a set of obligations designed to make moving between cloud providers, or off the cloud entirely, genuinely feasible rather than nominally permitted. The Act’s switching rules are already reshaping provider behavior, with several hyperscalers announcing changes to their data-transfer pricing in the weeks since it took effect. This piece sets out what the Act actually requires, what cloud customers actually get, and on what timeline.

What the Data Act Is

The Data Act is a regulation — meaning it applies directly across all EU member states without requiring national transposition — that aims to ensure fairness in how value from data is distributed and to remove barriers to a functioning data economy. Its scope is broad, but its cloud-relevant heart is a chapter on “switching between data processing services,” which targets the practical obstacles that have made changing cloud providers difficult, expensive, and risky.

The Act entered into force on 11 January 2024, but its provisions do not all apply immediately. Most obligations, including the switching rules, become applicable after a transition period — roughly twenty months — giving providers time to adapt their contracts, technical interfaces, and pricing. So the correct framing in early 2024 is that the Act is in force (the legal instrument exists and binds) but its substantive switching obligations are not yet applicable; they sit on a defined runway. This distinction matters for setting expectations: the rights are coming, with a date attached, rather than being available today.

The Switching Provisions

The switching chapter addresses several distinct barriers that, together, have made cloud lock-in a practical reality even where it was never a contractual prohibition.

  • Removal of switching charges. The Act requires the progressive withdrawal of charges that providers impose on customers for switching — most prominently the egress fees levied when a customer moves data out of a provider’s network. The intent is that, after the transition period, providers may not use data-transfer charges as a financial barrier to departure.
  • Contractual switching rights. Providers must include terms allowing customers to switch to another provider or to on-premises infrastructure, with defined maximum notice and transition periods, so that an organization is not trapped by open-ended or obstructive contract terms.
  • Interoperability and portability obligations. Providers must enable customers to port their data and, where applicable, their digital assets in a structured, commonly used, machine-readable format, and must provide the information needed to achieve functional equivalence on a new provider — addressing the technical, not just the commercial, dimension of lock-in.
  • Assistance with the switch. Providers are obliged to support the switching process rather than passively permit it, reducing the operational burden that has often fallen entirely on the departing customer.

Taken together, these provisions attack lock-in on three fronts at once: the cost of leaving, the contractual freedom to leave, and the technical feasibility of leaving. It is the combination that is significant. Removing egress fees alone does little if the customer’s data is trapped in proprietary formats; mandating portable formats alone does little if leaving still incurs a prohibitive transfer bill.

Early Provider Responses

The Act is already visible in provider behavior. In January 2024, Google Cloud announced it would waive exit-migration egress fees for customers leaving the platform, citing the Data Act explicitly. In early March 2024, other major providers announced comparable changes to their own exit-egress pricing — a pattern that suggests the Act’s switching provisions are functioning as intended even ahead of their formal applicability date, by creating commercial and reputational pressure to lead rather than wait to be compelled.

The caution here is to read the scope of each announcement carefully. The early moves have generally addressed exit egress — the cost of fully leaving a provider — rather than the much larger volume of ordinary day-to-day egress. That is consistent with the Act’s focus on switching specifically, but it means the changes do not amount to the elimination of egress charges in general. The lock-in that lives in proprietary APIs and managed services is also untouched by pricing changes; the Act’s portability and interoperability provisions address that dimension, but those are the parts still on the transition runway.

A Regulatory Lever on a Structural Problem

The Data Act is interesting beyond its specific provisions because of how it works as a mechanism. Cloud lock-in is a structural feature of a market where the incumbents have every commercial incentive to make leaving difficult. Market forces alone had not dislodged it — customers complained about egress fees for years without the major providers removing them. The Act demonstrates regulation acting as a lever on a structural problem that competition had not solved, and doing so partly through the incentive it creates for providers to move first.

For organizations whose architecture spans the EU market, the practical reality is that a regulation written to satisfy European policy goals is likely to reshape provider products globally, because maintaining separate contractual and pricing regimes per jurisdiction is impractical at hyperscaler scale. The reach of the EU market means the Data Act’s effects will not stay neatly within EU borders.

What Customers Should Do Now

The switching provisions are not yet applicable, but the transition period is the time to prepare, not to wait.

  • Map your switching exposure. Understand, per workload, what would actually be involved in moving it — the data volumes, the proprietary-service dependencies, the contractual terms. The Act will lower some of these barriers over its runway; knowing which barriers it addresses (cost, contract, format) and which it does not (deep API coupling) lets you plan accordingly.
  • Track the applicability date. The switching obligations become enforceable at the end of the transition period. Renewal and architecture decisions made between now and then should account for rights that will exist by the time many contracts come up for renewal again.
  • Use the early egress moves. Where providers have already waived exit-egress fees, the practical barrier to a migration you were considering may have fallen. Revisit any analysis that was shelved on egress-cost grounds.
  • Don’t mistake the Act for full portability. The Act improves switching; it does not make all workloads instantly portable. Architectural choices that minimize proprietary coupling remain the customer’s own responsibility and the deeper hedge against lock-in.

The Data Act is a substantial development, but a phased one. What cloud customers actually get in early 2024 is a clear legal direction of travel, a defined date by which switching barriers must fall, and provider behavior already beginning to shift in anticipation. The full benefit arrives with applicability at the end of the transition period — and realizing it will still require customers to do the architectural work that no regulation can do on their behalf.

Further Reading

Back to Blog