Microsoft CTO Kevin Scott on the birth of the agentic web

Microsoft CTO Kevin Scott discussed the company's new open-source tool, Model Context Protocol (MCP), designed to enable AI-powered natural language search on individual websites and facilitate the creation of an "agentic web." MCP, initially developed by Anthropic, aims to standardize how AI agents interact with databases and services, potentially reducing the reliance on centralized search indexes like Google and Bing. Scott emphasized the importance of aligning business models with technology to ensure content and service providers are incentivized to participate in the agentic web, allowing them to control data access and monetization strategies while fostering greater competition and innovation in the search landscape.

Analysis

Microsoft is championing a shift towards an "agentic web" through its open-source tool enabling Model Context Protocol (MCP) integration for websites, a standard initially developed by Anthropic and now seeing broader industry interest, including from Google. CTO Kevin Scott outlined this vision, emphasizing that MCP, in conjunction with Microsoft's Natural Language Web (NLWeb) project, aims to provide a decentralized alternative to current centralized search infrastructures dominated by Google and Bing. This approach would allow individual websites to implement AI-powered natural language search, control their data, and choose their AI models, potentially lowering search costs and fostering competition. The successful rapid integration by Tripadvisor was cited as an example of the technology's ease of use. A key challenge acknowledged is the development of new, sustainable business models for content and service providers in this new paradigm, as MCP itself is business-model agnostic, allowing for diverse monetization strategies such as subscriptions or novel advertising methods. Scott suggested the current antitrust pressures on Google, combined with the maturity of AI technology, create a favorable environment for such new web standards to emerge. He characterized the current AI landscape as being in its "middle innings" with a significant "capability overhang," where the power of AI models surpasses their current application in products, particularly outside of software development. Microsoft's relationship with OpenAI remains substantial, with OpenAI being a major Azure customer and a partner in core AI infrastructure, despite OpenAI's independent consumer product ambitions. The overarching goal is to establish open protocols that enable a more dynamic and potentially more equitable web ecosystem.

Key Decisions for Investors

• Investors should monitor the adoption trajectory of Model Context Protocol (MCP) and related technologies like NLWeb, as their success is pivotal to Microsoft's 'agentic web' strategy and could reshape the competitive landscape for search and AI services.
• Consider the long-term implications for Alphabet (GOOG/GOOGL), as widespread adoption of decentralized agentic search could erode its search dominance, while Microsoft (MSFT) stands to benefit from increased Azure utilization and a strengthened position in the evolving AI ecosystem.
• Evaluate the significant uncertainty surrounding the emergence of viable business models for content creators and service providers within an agentic web; successful monetization strategies will be critical for broad participation and the ecosystem's overall health.
• Focus on companies demonstrating an ability to translate AI's 'capability overhang' into tangible product-market fit, particularly in developing useful, general-purpose AI agents, as this will be key to unlocking value in the AI sector.