On 11 November 2026, Martijn Snoep, chair of the Netherlands Authority for Consumers and Markets (ACM), gave the first Louis Brandeis lecture on market power and democracy in the era of Big Tech. The ACM published the speech on 13 November. It is an unusually candid text from a sitting competition regulator. It says dependence on a handful of US platforms is high, that "each service has few alternatives, if any at all", and that no single policy tool fixes this.
Snoep's answer is not "either or" but all four of the choices he lays out: regulation, prohibition, laissez-faire and collaboration. This article argues that the framework is the most defensible one a European regulator has offered. It also argues that its weakest link is not conceptual but administrative.
The strongest case for the regulatory route
The case for the EU's rulebook is serious, and it deserves a fair statement first. Dependence on a few platforms is real. Individual enforcement actions are slow. Ex-ante rules exist because waiting for a market to tip, then litigating for years, can leave the damage done. The Digital Markets Act was built for this purpose. It became applicable on 2 May 2023, identifies "gatekeepers" by objective criteria, and allows fines of up to 10% of worldwide annual turnover, or 20% for repeat infringements. The DSA, the Data Act and the AI Act add further layers. Snoep does not reject any of this. He lists regulation as one of the four tools.
Where the framework is wise
The lecture's real contribution is to treat the other three choices as legitimate rather than as admissions of failure.
Prohibition, used sparingly. Snoep presents prohibition as a real option for specific uses, such as AI in decisions affecting fundamental rights, and as a way to protect children. A narrow, use-based ban is more proportionate than a general restriction on a technology. Bright lines for a few high-stakes applications leave the rest of the market free to innovate. That fits an evidence-based approach, provided each ban rests on demonstrated harm rather than on a general unease about the technology.
Laissez-faire, openly acknowledged. A regulator saying that some harms cannot be prevented at an acceptable cost to freedoms is worth noting. Speech and innovation are the usual casualties of rules written for the worst-case platform. Accepting limits is not weakness. It is the proportionality principle applied to enforcement.
Collaboration, the genuinely new element. Snoep adds a fourth choice: cooperation among competitors and buyers. The speech names common standards for assessing digital sovereignty, joint procurement or provision of sovereign cloud services, and a new e-wallet. It also says governments can act as a "launching customer" and that buyer-side collaboration "can create much-needed countervailing power."
This is the part most likely to matter. Regulation can penalise a dominant firm. It cannot conjure a competitor. Building a credible alternative requires demand to be aggregated so that European providers reach scale. The ACM has made the point before. In a September 2025 speech at De Balie in Amsterdam, Snoep said EU alternatives for the most crucial IT services are needed "with great urgency", and warned that exchanging dependence on a few US companies for dependence on one European company is "only marginally better."
That warning is the right test for any sovereign-cloud project. Collaboration among buyers is only pro-competitive if it keeps switching easy. Here the EU has already legislated. The Data Act has applied since 12 September 2025 and removes cloud switching charges, including data egress fees, entirely from 12 January 2027. Procurement consortia built on open interfaces and portable data would compound that effect. Closed, state-favoured champions would simply rebuild lock-in with a European flag on it.
There is a competition-law caution as well. Collaboration among competitors is something competition authorities normally scrutinise. The ACM is therefore well placed to say where cooperation is acceptable, but it must hold itself to the standard it applies to others. Joint ventures should be open to new entrants, time-limited where possible, and judged on whether they widen or narrow consumer choice.
The weak link: enforcement capacity
The lecture's sharpest practical warning is also its least conceptual. Snoep says the recently announced cutbacks to the Dutch civil service "will not help either", and that moving to a civil service 25-35 percent smaller "will also have an effect on the enforcement of existing rules." That figure refers to the civil service generally, not to the ACM specifically. The logic still holds. A four-pronged strategy needs staff to design bans, investigate harms, assess collaborations and police the DSA, DMA and Data Act at national level.
There is a political dimension too. The speech describes a DSA roundtable with Big Tech firms around the 2025 Dutch general election. It says the US House Judiciary Committee later published ACM employees' names unredacted, which Snoep calls an attempt to "intimidate ACM and individual employees." Critics of EU digital rules often cast them as censorship tools. Tech Policy Press contributor Courtney Radsch argues that the US government and large tech firms are campaigning against the DSA and DMA. That is an opinion piece, but the ACM's own account suggests regulators are under real pressure.
What policymakers should take from it
- Fund the enforcer before adding rules. Every new obligation without staffing is a paper right. Exempting competition and digital-regulation capacity from across-the-board cuts is cheaper than the dependence it addresses.
- Make bans narrow and evidence-led. Prohibit specific uses with documented harm, and sunset or review them.
- Tie sovereign procurement to openness. Public money should reward interoperability and portability under the Data Act, not national or regional preference alone.
- Be honest about limits. Where harm cannot be prevented without chilling lawful speech, say so and use lighter tools.
Snoep's framework is a sober alternative to both regulatory maximalism and the claim that Europe should simply wait for markets to fix themselves. Whether it succeeds depends less on the elegance of the four choices than on whether the Netherlands keeps the people needed to carry them out.