Divya van Mahajan

Does AI Search Threaten the Business Model of the Internet?

ai-daily-brief-podcast

Does AI Search Threaten the Business Model of the Internet?

Overview

This episode of the AI Daily Brief (recorded May 9, 2025) covers several interconnected stories about how AI is reshaping the economics of the internet, the restructuring of OpenAI’s deal with Microsoft, emerging agentic interoperability standards, Meta’s AI advertising ambitions, and U.S. AI chip export policy. The host (unnamed in the transcript) presents these as part of a broader narrative about AI-driven disruption across multiple industries and geopolitical contexts.

Source video: URL not provided. Episode title: “Does AI Search Threaten the Business Model of the Internet?” — AI Daily Brief, May 9, 2025.

Prerequisites

  • Basic understanding of how internet advertising and search economics work (cost-per-click, ad revenue, traffic-driven monetization)
  • Familiarity with the Google antitrust case and its remedies phase
  • Awareness of OpenAI, Microsoft, Apple, Cloudflare, Meta, NVIDIA, and their roles in the AI ecosystem
  • General knowledge of U.S. export control policy and the concept of chip restrictions on China
  • Understanding of what large language models (LLMs) and AI agents are

Main Points

AI Search Is Disrupting the Core Business Model of the Internet

  • Apple SVP Eddie Cue testified in the Google antitrust trial that Apple is actively exploring AI-powered search engines (OpenAI, Perplexity, Anthropic) as alternatives in Safari, calling the search landscape “more wide open than it’s ever been.”
  • Safari recorded its first-ever decline in search volume, which Cue attributed to increased AI usage.
  • Cloudflare CEO Matthew Prince stated that AI is fundamentally changing web monetization: whereas Google once sent one visitor for every two pages scraped, it now sends one visitor per six pages scraped, because ~75% of queries are answered directly on Google’s results page without a click-through.
  • Prince concluded: “The business model of the web can’t survive unless there’s some change.”
  • Content creators who rely on ad revenue, subscriptions, or readership traffic are losing the primary mechanism that drove that value.

Market Reaction and Divergent Interpretations

  • Google shares fell more than 8% following Cue’s testimony, reflecting investor fears that AI has structurally undermined Google’s search dominance.
  • Skeptics (e.g., commentator Talon Sharpedge) argued the market is reacting to “headlines and vibes” rather than fundamentals, noting Apple has not shipped major software innovation in years and its AI moves are largely rebranding of third-party models.
  • Bloomberg’s Mark Gurman offered an alternative reading: Apple may be strategically portraying Google search as obsolete to persuade the court that their revenue-sharing deal is no longer competitively significant and therefore does not need to be broken up.
  • A regulatory commentary thread noted that by the time regulators act, market forces often already drive the relevant shifts.

OpenAI–Microsoft Revenue-Sharing Negotiations

  • OpenAI is restructuring its for-profit conversion, with Microsoft as a key holdout among investors.
  • Existing terms: Microsoft receives 20% of OpenAI’s top-line revenue plus a 49% share of profits capped at $92 billion.
  • OpenAI has told investors it expects to reduce the revenue share to 10% by 2030.
  • If OpenAI hits its 2030 revenue projections, Microsoft’s take from the revenue share alone could reach $97 billion.
  • The contract term was renegotiated from “until AGI is achieved” to a fixed end date of 2030; Microsoft wants extended access to OpenAI technology beyond that date as a quid pro quo.
  • Legal observers were skeptical of OpenAI’s leverage, noting Microsoft is unlikely to accept greater antitrust risk by converting revenue share into an equity deal.

Microsoft Adopts Google’s Agent-to-Agent (A2A) Protocol

  • Google launched Agent-to-Agent (A2A), an interoperability protocol enabling AI agents to communicate with each other, distinct from Anthropic’s Model Context Protocol (MCP), which governs agent access to external tools and data.
  • Microsoft announced support for A2A on Azure AI Foundry and Copilot Studio, and joined the A2A development consortium.
  • MCP already has broad industry support (OpenAI, Google, Microsoft, and others). A2A has less buy-in currently, but Microsoft’s adoption is a meaningful signal.
  • The emerging pattern is that AI companies are converging on shared standards rather than fighting format wars, reducing vendor lock-in at the infrastructure layer.

Meta’s Vision for Fully Automated AI Advertising

  • Mark Zuckerberg outlined plans for an end-to-end AI advertising platform: businesses state their objective and budget, connect a bank account, and Meta’s AI handles creative generation, audience targeting, and optimization autonomously.
  • The system would generate thousands of AI-created ad variations, test them across Meta’s social networks, and scale what performs.
  • Critics (e.g., ad agency founder John Hornby) argued generative AI cannot replicate the “leaps of imagination” behind major brand-building campaigns.
  • The host distinguishes between brand advertising (creative, high-concept) and direct response advertising (performance-driven, measurable)—Zuckerberg’s system targets the latter, which is already largely a data optimization exercise.
  • The host frames full automation of direct response advertising as “completely inevitable,” beneficial especially for small businesses.

U.S. AI Chip Export Policy: Repealing the AI Diffusion Rule

  • The Biden administration’s AI Diffusion Rule divided the world into three tiers with varying chip export restrictions; most allied nations (India, Israel, South Korea) were placed in a restrictive Tier 2.
  • The Trump administration announced plans to repeal the rule one week before its scheduled implementation, calling it “overly complex, overly bureaucratic.”
  • NVIDIA CEO Jensen Huang lobbied officials to reconsider, arguing China’s AI progress stems from scientific investment, not chip smuggling: “China’s not behind. China’s right behind us.”
  • The replacement policy is expected to be simpler; the administration stated it will continue strict enforcement of chip controls specifically on China and will crack down on suspected diversion via Malaysia and Thailand.
  • Chip access may be folded into broader trade negotiations, with the Gulf states (Saudi Arabia, UAE) as a region of particular focus—Trump signaled loosening restrictions there ahead of a diplomatic trip to the Middle East.

OpenAI’s “AI for Countries” Initiative

  • Following the Stargate announcement, OpenAI launched AI for Countries, a program to co-fund AI infrastructure (data centers) with foreign governments.
  • OpenAI aims to pursue at least 10 international projects initially, though specific locations were not disclosed.
  • The initiative carries explicit geopolitical framing: OpenAI described the goal as spreading “democratic AI”—development and deployment of AI that “protects and incorporates long-standing democratic principles”—as a counterweight to Chinese AI deployment globally.
  • The initiative aligns with the second objective of the original diffusion rule: not just preventing Chinese chip access, but actively spreading U.S. AI technology to friendly nations.

Key Concepts

  • AI Diffusion Rule: A Biden-era export control framework categorizing countries into three tiers based on permitted volumes of advanced AI chip imports, intended to prevent chip leakage to China while expanding U.S. AI access globally.
  • Search Economics: The advertising and traffic model underpinning much of the web, whereby Google’s search results drive clicks to third-party sites, generating ad revenue and readership for content publishers.
  • Zero-Click Search: The phenomenon where search engines answer queries directly on the results page, eliminating the need for users to visit external websites, thereby reducing traffic to content creators.
  • Model Context Protocol (MCP): An Anthropic-originated open standard that governs how AI agents access data from external tools; now broadly supported across the industry.
  • Agent-to-Agent Protocol (A2A): A Google-originated interoperability standard enabling AI agents to share data and communicate with one another across different platforms and ecosystems.
  • Direct Response Advertising: Performance-driven advertising measured by specific user actions (clicks, purchases), as opposed to brand advertising focused on awareness and creative impact.
  • Innovator’s Dilemma: The phenomenon whereby an established market leader struggles to adopt disruptive innovations that threaten its existing business model, used here to characterize Google’s position relative to AI search.
  • AI for Countries: OpenAI’s initiative to partner with foreign governments to co-build national AI infrastructure, framed as a vehicle for spreading U.S.-aligned “democratic AI.”
  • Chip Diplomacy: The emerging use of access to advanced AI semiconductors as a geopolitical bargaining tool in international trade and diplomatic negotiations.

Summary

The central argument of this episode is that AI is simultaneously disrupting multiple foundational structures of the digital economy and global technology competition. In the near term, AI-powered search is eroding the click-through traffic model that has funded the open web for two decades, threatening content creators and raising existential questions about web monetization—a shift evidenced by Apple’s testimony, Safari’s first-ever search volume decline, and Cloudflare’s data on deteriorating traffic returns. Concurrently, the business arrangements underpinning AI development itself are under negotiation, as seen in the contentious OpenAI–Microsoft revenue-sharing dispute, while the industry is quietly converging on shared agentic standards (MCP and A2A) that reduce fragmentation. Meta is moving to fully automate the direct response advertising cycle, and the U.S. government is recalibrating chip export controls and supporting OpenAI’s international infrastructure push as tools of geopolitical competition with China—reflecting a world in which AI infrastructure, trade policy, and national strategy have become inseparable.