(Preview) Google and Apple Avert Disaster, More on Kpop Demon Hunters and Hollywood, Mail on Microsoft, Adobe and ASML
Summary
- Judge Amit Mehta preserved Google’s default-search payments despite finding that its nonexclusive default agreements foreclosed competition. He rejected Google’s argument that users could simply change their defaults, yet left the payments intact in the remedies. Sharp described the Apple deal as giving Apple 36% of Google’s search revenue, while Thompson said Google can pay more than rivals and make recipients dependent: “Everyone is too dependent on this money, so we can’t take it away.”
- Apple is the biggest near-term beneficiary, retaining more than $20 billion annually in “pure profit from Google for doing nothing.” Thompson sees the inverse long-term risk: Apple has less reason to build search or AI capabilities and might instead use Gemini to improve Siri. “Apple is a short-term winner, and innovation at Apple, at a minimum, is a long-term loser.”
- The Apple Maps episode illustrates the competition that search payments can suppress. After difficult Maps negotiations, Apple built its own product; Thompson said Google lost valuable signals from affluent iPhone users, while Apple’s default helped consumers gain another mapping platform. He argues Google’s payments remove Apple’s incentive to make the analogous investment in search.
- Protecting Mozilla is defensible; treating cash-rich Apple as another dependent is not. Thompson said ending the payments would kill Firefox and hurt Android manufacturers whose handset profits largely come from Google. Sharp proposed preserving payments to Firefox while enjoining Apple’s deal, and called the ruling “gutless.”
- The data remedy is too weak to offset Google’s continuing distribution advantage. Challengers receive one-time access to Google’s index, including the long tail, but not its accumulated data or ongoing crawling capabilities. By leaving payments in place while limiting access, Mehta “chose from column A and column B,” producing a remedy Thompson thinks “is not gonna accomplish anything.”
- AI’s rise may justify caution, but it heightens the need to stop Google from carrying its search advantage into Gemini. Mehta cited AI competition and reserved the right to revisit the agreements in a couple of years. Sharp disputed whether hundreds of billions in AI capital directly target Google Search; Thompson countered that “we can’t have it both ways” after recognizing ChatGPT’s threat. Even without payments, he thought Apple would probably keep Google as the default because users might riot at a Bing-like alternative. His preferred guardrail is an open API letting any AI service ground results on Google’s index, rather than making the optimized path available only through Vertex.
Deep dive
1. Google’s patronage network survived the monopoly judgment
Sharp framed the contradiction: Mehta ruled on August 5 that “Google is a monopolist, and it has acted as one to maintain its monopoly,” yet his September 2 remedies left the decisive payments intact. Although Mehta rejected Google’s argument that nonexclusive defaults were legal because users could switch, he found that users do not change defaults in practice, so an agreement between two of the three biggest companies foreclosed search competition.
Thompson’s “overt generosity” thesis held that the payments do two jobs: lock out potential competition and make recipients so dependent that unwinding them becomes prohibitively disruptive. “Google has paid off everyone.”
Thompson corrected an important legal distinction: monopoly itself is not a crime; illegally maintaining one is. Courts can undo gains from that conduct, but Mehta prioritized downstream harm over ending the agreements.
2. Firefox presented real collateral damage; Apple did not
Thompson said a blanket prohibition would kill Firefox. Android manufacturers also sell hardware at cost—or even a loss—while Google payments generate much of their profit.
Sharp proposed a narrower remedy that would preserve Mozilla as a going concern while enjoining Apple’s agreement: “It’s the absurdity of the Apple deal that really rankles me.” He likened Mehta’s overly cute solution to something Adam Silver would have tried and called the ruling “gutless.”
Mehta cited the risk of “fewer products and less product innovation” from Apple. Thompson called that reasoning ridiculous: Apple has ample money, while the ignored downstream effect is that it simply pockets Google’s payments instead of competing.
3. Apple Maps shows what genuine platform competition requires
Thompson revisited Google’s Maps negotiations, when demands for data and advertising rights let Android’s ambitions “wag the dog.” Apple walked away and funded its own initially embarrassing Maps product.
Thompson said Google consequently lost an extremely valuable signal from affluent iPhone users; Sharp added that it lost “a ton of share.” Apple Maps gained distribution through default placement, and Sharp’s own behavior supplied the example: he remained an Apple Maps user because Apple created the app.
The search deal removes the incentive for that same investment cycle. Thompson’s core claim was not merely that Bing or DuckDuckGo loses placement, but that Apple loses any incentive to build competing search.
4. Apple’s cash victory could deepen its AI dependence
Apple keeps “$20 billion plus a year of pure profit from Google for doing nothing.” With legal uncertainty reduced, Thompson said it “probably behooves Apple” to use Gemini as a partner for rebuilding Siri.
He considers Google the more natural partner: the companies have settled their old Android tensions and now comfortably govern smartphones as a duopoly. OpenAI, by contrast, is “coming for Apple’s throat” over the long run.
The strategic cost is deferred capability-building. Thompson concluded that Apple need not invest in search and may decide it need not invest deeply in AI; Sharp connected that choice to hardware competitiveness 10 years out.
5. The hybrid remedy leaves Google’s AI leverage intact
Mehta also cited the emergence of AI companies as a reason for caution and reserved the right to revisit the agreements in a couple of years.
Competitors, explicitly including AI companies, receive one-time access to Google’s search index and its long tail. They do not receive Google’s accumulated data or continuing crawling operation, so they must find and refresh sites themselves.
Thompson saw two coherent choices: end the payments and give challengers an initial index head start, or retain payment competition while allowing ongoing “free riding” on Google’s infrastructure. The selected hybrid preserves Google’s compounding advantage.
Sharp argued that AI investment is not necessarily aimed at direct Google Search competition. Thompson disagreed: ChatGPT and LLMs plainly threaten Google, but even cutting payments might not change the default because users could “riot” if Apple replaced Google with Bing.
Thompson’s proposed remedy targets the next market: require an API through which any AI provider can ground answers on Google’s index. Vertex availability is insufficient if Google alone can integrate Gemini with Search in an optimized, streamlined way.