The AGNTCon+MCPCon North America 2026 keynote schedule and the program’s Featured Keynote Speakers page function as a strategic map of the emerging agent infrastructure landscape. By analyzing the speaker assignments, we can identify how major vendors are positioning their platforms and where they are concentrating their influence. The stage in San Jose this October reveals a clear divide between those prioritizing broad platform governance and those focusing on specialized technical foundations. Our earlier read of this program mapped its 11 tracks as competing control planes. The keynote assignments add the second layer this piece reads: who got the main stage, and what each vendor chose to do with five minutes of it.
Google has secured the most significant presence, utilizing three keynote slots to signal a full-stack approach. Rao Surapaneni, VP/GM of the Business Application Platform at Google Cloud, will discuss shipping autonomous agents safely with open standards, observability, and control. Sarah Drasner, a Distinguished Engineer, is scheduled to address the “Agentic Web,” while Chi Wang, founder of AutoGen and a senior staff research scientist at Google DeepMind, will present Sutando, an open-source personal AI agent that runs on hardware its owner controls. This distribution suggests a strategy aimed at capturing the entire lifecycle of agent deployment, from cloud-scale architecture to individual user interaction.
Anthropic is simultaneously working to establish the Model Context Protocol (MCP) as the industry standard. With two keynote slots, the company features co-creator David Soria Parra in a Thursday session, alongside Kelly Fang, who will detail how Anthropic is adopting and building with MCP across its products and what the team has learned from running it at scale. This dual focus anchors the protocol layer firmly within Anthropic’s ecosystem. On the same stage, Solo.io, which also sponsors the event, holds a keynote slot for CEO Idit Levine and CPO Keith Babo. This appearance follows the company’s donation of its agentgateway to the Linux Foundation, providing a platform to showcase their infrastructure contributions to the broader community.
AWS has adopted a different approach, focusing its single keynote on the verification of distributed systems rather than platform governance. Shadaj Laddad’s presentation, “Abstractions are Dead, Long Live Abstractions,” centers on deterministic simulation and universal properties. This choice highlights a shift in focus: while other vendors are building expansive platforms, AWS is positioning itself as the provider of the rigorous, verifiable foundation necessary for high-stakes agentic operations.
The disparity in protocol adoption is legible in the speaker assignments themselves. MCP commands three keynote slots: its co-creator David Soria Parra, OpenAI’s Thomas Sarlandie with “MCP in the Wild,” and Kelly Fang on how Anthropic is building with it. The Agent-to-Agent (A2A) framework has no keynote of its own on the published schedule. Surapaneni’s abstract names A2A among the open protocols his talk covers, but the protocol’s real footprint is on the floor rather than the stage: hands-on workshops, an A2A v1 deep dive, and discovery sessions. What that allocation indicates is narrower than a verdict on the technology: on this program, MCP takes the main stage while A2A remains the work of the builders in the technical sessions.
The concentration of keynotes around MCP and cloud-native governance suggests that the industry is coalescing around specific control planes. The event schedule serves as a primary indicator of which standards are gaining institutional momentum. As the conference approaches, the divide between those defining the core standards and those participating in the technical workshops will serve as the most reliable metric for evaluating the long-term viability of various agent infrastructure projects.