Google's past mergers like YouTube sailed through regulators — what does that history mean for its push into AI shopping agents?

This explores why regulators let Google's earlier acquisitions go through, and what that history means for Google's push into AI shopping and agents. One thing first: the collection has no material on the merger reviews of YouTube or DoubleClick themselves. It can't tell you what the FTC weighed in 2007 or why it cleared the DoubleClick deal. What it does have is a sharp argument that the same blind spot is repeating with AI. Seen through that argument, the historical question becomes a live one.

The main thread comes from Matt Stoller. He argues that Google's strategy has always been to build chokepoints: positions every transaction has to pass through. It isn't about buying any single competitor. In his telling, Google is now extending its search-advertising chokepoint into agents. It is gathering personal data, distributing its chatbot, and signing retailer and payment-network deals that, taken together, position it as a 'central planner' of economic transactions Is Google building a chokepoint across data, chatbots, and payments?. The point that answers your question: none of these pieces looks dangerous alone. Traditional antitrust tends to judge each deal on its own terms, mostly by asking whether prices will rise. A strategy built from many individually harmless steps can pass every one of those tests and still end up controlling the whole market.

The pricing examples show why price-focused tests miss these harms. Stoller argues that Google's Gemini pricing service, which advises competing merchants, starts to look like automated price-fixing. It blurs the line between serving rivals and coordinating them, in ways 'existing antitrust frameworks may not catch' Does Google's AI pricing tool coordinate prices across rival merchants?. He also calls Google's promise that the tool won't raise prices unfalsifiable, because merchants can inflate the list price before any discount applies Can Google's pricing tool really prevent price increases?. A formal economic model sharpens the point. A platform can set its fees so that consumer welfare stays flat while value quietly shifts from sellers to the platform Does platform pricing of inspection tokens encourage excess search?. If the test is 'did consumers pay more?', this kind of harm comes back clean.

There's a counterweight. Lab experiments find that AI pricing agents collude when they are identical. Mixing different agents breaks the collusion Does agent heterogeneity prevent LLM pricing collusion?. That turns a classic antitrust worry on its head: the danger comes less from many firms using AI and more from many firms using the same AI from one provider. That is exactly the scenario a dominant platform creates. Lawmakers are starting to try rules that act on structure before any harm shows up, rather than reviewing deals one at a time. Senator Warner's draft AI AGENT Act would require large platforms to admit competing agents Could federal AI agent registries solve platform access unfairness?. Meanwhile, the Congressional Research Service notes that antitrust law can also block cooperation people might want, such as AI labs agreeing to slow down development What antitrust risks do different AI safety collaborations face?.

The takeaway you may not have expected: the collection's answer to 'why did enforcement fail?' amounts to 'because it was looking at the wrong unit.' Judging one deal at a time, by its effect on prices, can't see a company assembling a chokepoint piece by piece. For the actual record of the YouTube and DoubleClick reviews, look outside this library.