Brazil competition law tech

Brazil Narrows Its Gatekeeper Bill's Discretion Before CADE Ever Gets to Use It

Brazil's rapporteur tightened PL 4675/2025's designation test and merger rules, trading CADE flexibility for legal certainty before a single firm is named.

Brazil's Gatekeeper Bill, Narrowed People of Internet Research · Brazil R$50B Global revenue threshold Triggers CADE designation review a… R$5B Brazil revenue threshold Domestic-only alternative trigger,… 15→30 days Comment period, old to new Public consultation window doubled… Mar 18, 2026 Urgency motion approved Lets the bill skip committee revie… peopleofinternet.com
Brazil's Gatekeeper Bill, Narrowed People of Internet Research · Brazil R$50B Global revenue threshold R$5B Brazil revenue threshold 15→30 days Comment period, old to new Mar 18, 2026 Urgency motion approved peopleofinternet.com

Key Takeaways

Brazil's answer to the EU's Digital Markets Act has spent a year being rewritten before it has regulated a single company. On July 8, 2026, Deputy Aliel Machado (PV-PR), rapporteur for Bill 4675/2025 — Brazil's digital markets legal framework bill — filed a revised substitute text in the Chamber of Deputies that narrows several of the powers the bill would hand to Brazil's antitrust authority, CADE (Chamber of Deputies tracking page).

The bill, submitted by the Executive Branch on September 17, 2025, would let CADE designate "economic agents of systemic relevance" in digital markets — firms with global annual revenue above R$50 billion (~$9 billion) or Brazilian revenue above R$5 billion (~$900 million) — and then impose bespoke obligations on them through a new Superintendência de Mercados Digitais inside CADE (CSIS analysis). Unlike the DMA, which locks a fixed list of obligations onto any firm crossing a threshold, Brazil's model is case-by-case: CADE investigates, then tailors remedies to what it actually finds.

What the rapporteur just changed

Four changes stand out in the July 8 substitute, and they move in the same direction: less discretion for CADE, more predictability for regulated firms.

These changes are documented in contemporaneous reporting on the filing (Truth on the Market; aRede). The bill cleared an urgency motion on the floor on March 18, 2026, meaning it can go straight to a plenary vote without committee review — but it stalled before Congress's recess began July 18, and a public hearing on the bill's economic impact was only held August 12.

The case Machado is answering

It's worth taking the pro-regulation case seriously, because it isn't frivolous. CADE's own enforcement record shows the problem ex-ante rules are meant to fix: its investigation into Google Shopping took roughly five years and ended in a tied vote, long after any remedy could have mattered to the market it was meant to protect. Proponents argue that waiting years for a case-by-case antitrust finding against a dominant platform effectively cedes the market while the case is pending, and that dispersing entrenched platform power is what actually creates room for Brazilian SMEs to build competing products (ProMarket). CGI.br, Brazil's multistakeholder internet governance body, endorsed the underlying goal in a November 26, 2025 public note, calling for "proportional regulation of digital markets to defend national sovereignty [and] stimulate innovation" while committing to keep weighing in as the text evolves (CGI.br note). That is a coherent position: ex-post enforcement is too slow for markets that tip in months, and a sector-expert regulator with tailored tools can move faster without assuming every large platform behaves identically.

Why the narrowing still matters

The steelman doesn't resolve the core design risk, which is that Brazil is building an ex-ante regulator with open-ended discovery power rather than fixed rules, and asking firms to trust that CADE will use it narrowly. Every one of Machado's four changes responds directly to that risk rather than to the policy's goals. A closed, two-step designation test means a firm can read the statute and know in advance whether it's exposed, instead of discovering after the fact that some unlisted "other" factor swept it in — the single biggest predictability problem with open multi-factor tests, and one the DMA itself avoided by fixing its thresholds in euros and user counts rather than open criteria. Dropping the mandatory pre-merger filing duty for every sub-threshold deal matters because blanket pre-merger notification regimes impose real transaction costs on deals that were never anticompetitive in the first place; an informational communication to CADE's Digital Markets Superintendency lets it flag the genuinely suspicious ones without taxing all the rest. Scoping obligations to specific services rather than whole corporate groups prevents a rule aimed at a social network from incidentally capturing a cloud division or a payments subsidiary that shares a parent company but not a competitive problem.

None of this makes PL 4675/2025 a deregulatory bill — designation still rests on CADE's judgment, and the threshold figures remain low enough to catch a fairly small set of firms operating in Brazil. But the direction of every amendment since September 2025 has been toward constraining the regulator's own discretion, which is the right instinct for a new gatekeeper regime: better to over-specify now, while the rules are still just text on a bill, than to let CADE's Superintendency define its own jurisdiction case by case once the statute is live. Congress returns from recess in August; whether these four changes survive a floor vote is now the thing to watch, not whether Brazil regulates platforms at all — that question was settled in September 2025.

Sources & Citations

  1. Chamber of Deputies — PL 4675/2025 tracking page
  2. CGI.br public note on PL 4675/2025
  3. Truth on the Market — 'Brazil's Gatekeeper Bill Gets a Tuneup, Not a Rethink'
  4. aRede — rapporteur files substitute text
  5. CSIS — Unpacking Brazil's Latest Effort to Regulate Digital Markets
  6. ProMarket — Brazil's Fair Digital Competition Bill