What the ACM found
On 28 October 2026 the Dutch Authority for Consumers and Markets (ACM) published a study of how Dutch energy suppliers buy cloud services. It sent a short questionnaire to all Dutch energy suppliers and interviewed IT staff at ten of them. The ACM's news release reports a gap between ambition and behaviour. Suppliers say they want more digital autonomy, yet few have changed cloud provider. They fear high costs and disrupted processes.
The ACM says cost, functionality and integration with existing IT systems usually decide the choice, and that ease of use and broad functionality weigh heavily. Most firms stay with one or a few large providers. Suppliers that moved fully or partly from a hyperscaler to a European alternative were positive. They reported improvements in cost, functionality and flexibility, although some found the migration itself difficult.
The release gives no percentages, and the sample is small: ten interviews. It is a qualitative snapshot, not a market measurement, and it should be read that way.
The strongest case for tougher intervention
The case for more aggressive action is serious. Energy suppliers run billing, customer data and grid-facing systems, so they are critical infrastructure. If three U.S. firms (Microsoft, Google and Amazon) dominate the underlying cloud, a pricing dispute or outage at one of them becomes a systemic risk. Switching costs also compound. Once a supplier builds workflows around one provider's proprietary tools, leaving gets steadily more expensive, and the provider can raise prices knowing that. A regulator that waits for evidence of abuse may wait a long time.
What the Data Act already does
But the regulatory toolkit is not empty. Chapters 6 to 8 of the EU Data Act, Regulation (EU) 2023/2854 set rules on switching between data processing services, parallel use, interoperability and international access to non-personal data. The ACM supervises them in the Netherlands.
In October the ACM put a draft Leidraad Cloud (cloud guidance) out for consultation. The deadline is 6 November 2026, three days from now. According to PPC Land's reading of the draft, the 52-page document says that from 12 January 2027 a cloud provider may not charge for switching. It also sets out contract notice and transition expectations, and includes a 28-question checklist for contracts. PPC Land notes that charges at cost for data transfer in parallel use remain allowed.
Board member Manon Leijten said the Data Act can advance digital autonomy because it makes switching easier, and that users can report providers that obstruct a switch to the ACM. This is a sensible division of labour. Hard rules are set by the legislature, the regulator interprets them, and users supply the enforcement leads.
Why rules alone will not move the needle
The survey's most useful insight is that the stated obstacle is not a fee line on an invoice. It is the fear of disrupted processes. A ban on exit charges removes one cost, but it cannot remove the engineering work of re-platforming billing systems, retraining staff and running two environments during a migration. The ACM itself notes that some switchers found the process difficult.
That matters for policy design. If regulators judge success by switching rates, they risk concluding that the Data Act has failed when it has simply removed only one of several barriers. The reverse mistake is worse: layering on further obligations, such as mandated interoperability beyond what the Data Act requires or national sovereignty procurement rules, would raise compliance costs for every provider. European alternatives are the least able to absorb those costs, and they are the very firms the policy aims to help.
The finding that switchers were satisfied is encouraging for competition. It suggests European alternatives are competitive on substance, not just on sovereignty sentiment. But it also comes from firms that chose to switch, so it is subject to selection bias. Suppliers that stayed may have had good reasons, including the functionality and integration advantages the ACM identifies.
A proportionate path
A pro-innovation approach would make the following priorities clear:
- Enforce before adding. Give the Data Act's switching and fee rules a full year of application, from 12 January 2027, before judging them. The ACM should publish how many reports it receives about blocked switches and how it resolves them.
- Keep the guidance practical. The Leidraad Cloud should help a procurement team write a better contract. Its 28-question checklist is a good model because it translates legal text into questions a buyer can ask on the first day. Respondents should push the ACM for worked examples of what counts as an obstacle to switching.
- Target the migration problem. Standardised export formats and clear technical documentation reduce risk more than fee bans do. Industry-led tooling and shared migration playbooks, not new mandates, should carry this work.
- Measure outcomes. Repeat the energy-sector survey in 2028 with larger samples and quantified answers, so the ACM can show whether fear has given way to action.
The ACM's study deserves credit for asking buyers directly rather than assuming the answer. Its message is that Dutch energy suppliers are not locked in by law but by risk. Regulation can lower the risk at the margin, but suppliers, providers and migration specialists will have to do the rest. If the ACM resists the temptation to judge the Data Act before it has taken full effect, it can keep the market open without choking off the investment that makes European cloud alternatives viable.