On May 7, 2026, Canada's Competition Bureau asked the Competition Tribunal to compel evidence from former Google employees living in the United States, after none of the people it wanted to question agreed to testify voluntarily. It is a procedural footnote with a large lesson inside it: the Bureau's signature big-tech case — an application to break up Google's advertising-technology business — increasingly depends on facts, documents, and people that sit outside Canadian jurisdiction.
How the case got here
The Bureau filed its application with the Competition Tribunal in late November 2024, alleging that Google abused its dominant position in online advertising and asking the Tribunal to force Google to divest two tools at the centre of the ad-tech "stack": its publisher ad server, DoubleClick for Publishers (DFP), and its ad exchange, AdX. The Bureau's figures describe near-total control of the pipes: roughly 90% of publisher ad servers, 70% of advertiser networks, 60% of demand-side platforms, and 50% of ad exchanges in Canada (Competition Bureau backgrounder; Newsweek).
Google's first line of defence was constitutional. It argued the penalty the Bureau seeks — three times the benefit derived from the conduct, or, if that cannot be calculated, 3% of Google's worldwide gross revenue — was so large (Google warned of a figure approaching $91 billion) that it amounted to a "true penal consequence" triggering Charter protections. In early March 2026, Tribunal judge Andrew Little dismissed that challenge, calling the feared penalty "hypothetical at best" and noting that large fines may be necessary to deter non-compliance (Globe and Mail). With the constitutional off-ramp closed, the case turns to evidence — and that is where the witness problem surfaces.
The strongest case for the Bureau
It would be a mistake to treat this as a watchdog overreaching against an innocent firm. The conduct the Bureau describes is not speculative. In April 2025, US District Judge Leonie Brinkema found Google liable for unlawfully monopolizing both the open-web ad-exchange and publisher-ad-server markets, and for illegally tying the two together — citing a 91% share for its DoubleClick server and finding "substantial" harm to publishers and to the open web (Simpson Thacher analysis). Canadian publishers and advertisers transact in the same global stack, so the harm the Bureau alleges is real and locally felt. When a firm runs the auctioneer, the buy-side tool, and the sell-side server simultaneously, the conflict of interest is structural, not hypothetical. A national regulator declining to act simply because the defendant is foreign would be abdication, not restraint.
Where proportionality breaks down
Grant all of that, and the May 7 motion still exposes the core difficulty. The witnesses the Bureau cannot reach are former Google employees in the United States. The conduct was designed and executed largely in the US. And the most authoritative finding of illegality already exists — Brinkema's ruling, now in its remedy phase, where the US Justice Department is itself seeking divestiture of AdX and DFP. Canada is, in effect, litigating the same conduct, against the same products, toward the same break-up remedy, but with a fraction of the evidentiary reach.
That raises a proportionality question the Bureau should answer squarely. A structural break-up is the most drastic tool in competition law; layering on a penalty pegged to worldwide revenue compounds it. The Tribunal was right that the $91-billion figure is hypothetical — but a remedy calibrated to global turnover, for harm measured in a single mid-sized market, is hard to square with the principle that sanctions should track the injury. If AdX and DFP are divested globally as a result of the US case, the marginal benefit of a separate Canadian break-up order shrinks toward zero, while its costs — years of cross-border discovery, compelled foreign testimony, and legal uncertainty for every Canadian publisher mid-stream — do not.
A more proportionate path
The pro-competition goal here is sound; the instrument is heavier than the harm requires. Three adjustments would serve Canadian advertisers better than a parallel break-up race.
- Sequence, don't duplicate. The US remedy phase will likely reshape Google's ad stack worldwide. Ottawa can hold its structural claim in reserve and move first on conduct rules — interoperability, auction transparency, an end to self-preferencing between Google's own tools — that deliver relief to Canadian publishers now without waiting on witnesses it must subpoena across a border.
- Calibrate the penalty to Canadian harm. A fine anchored to demonstrable overcharges and lost publisher revenue in Canada is both more defensible and more likely to survive appeal than one indexed to global revenue.
- Coordinate with allies. The EU's 2025 ad-tech decision, the US case, and the UK's digital-markets regime are converging on the same defendant. A Canadian remedy harmonized with those processes avoids contradictory obligations and shares the evidentiary load that the witness motion is straining to carry alone.
None of this means letting dominant conduct slide. It means matching the remedy to the injury and to the jurisdiction's actual reach. The witness motion is a useful signal: a break-up case built largely on extraterritorial conduct will keep colliding with the limits of national process. Better to win durable, proportionate relief for Canadian publishers than to chase the most dramatic remedy and discover, late in a long trial, that the decisive evidence never crossed the border.