Agent Platform Scorecard

Who is building the agent operating system?

Twelve dimensions across seven providers. Re-graded as announcements land, with the market's reaction shown alongside.

When a keynote runs, we grade it in real time. Each announcement gets a plain read from an agent and control-plane perspective, the affected scorecard cells move, and where there is a public ticker we note how the market reacted. The market is a second opinion, a live vote on whether investors believe the announcement matters.

WWDC 2026

live nowApple · 2026-06-084 graded

Apple's developer keynote is the single most consequential event on this scorecard's calendar. Apple enters with the widest gap between its assets (on-device silicon, personal context, distribution) and its assistant capability. The question is whether this is the year App Intents, a rebuilt Siri, and a real model strategy turn latent platform power into actual agent capability, or another year of catch-up.

Meta Connect 2026

upcomingMeta · 2026-09-17watching

Connect is where Meta's ambient-AI thesis shows up: the next Llama, Meta AI across the app family, and the Ray-Ban / AR glasses as an on-device agent surface. Watch on-device AI, conversational UX, and whether open weights translate into a coherent platform play.

Latest graded announcements
  1. Google

    Gemini becomes the brain of a billion iPhones

    The flip side of Apple’s concession is a landmark distribution win for Google. Gemini will power Siri’s frontier reasoning across Apple’s installed base, extending Gemini’s reach into the one large ecosystem Google could never touch, and reinforcing its position as the model everyone else builds on.

    Third-Party Integrations7→8▲+1
    source: MacRumors: Gemini-powered Siri
  2. Apple

    Apple concedes the frontier: Siri’s hard reasoning runs on Google Gemini

    Apple confirmed that Siri’s most demanding reasoning is powered by Google’s Gemini models. Pragmatically it is the fastest route to a capable assistant, and precisely why it does NOT lift Apple’s Model Strategy grade. Renting frontier intelligence from a direct competitor is the clearest possible statement that Apple does not own its model destiny. Intelligence up (graded above); independence unchanged, so Model Strategy holds at 4.

    source: MacRumors: Gemini-powered Siri
  3. Apple

    Siri gains personal context and acts across apps

    Siri can now pull in information from your phone (Messages, and deeper AI woven through Photos and Mail) and understand what apps can do. This finally activates the latent advantage that always separated Apple: the device already knows you. It is the clearest step yet toward real cross-app orchestration via App Intents, and the single most MyAGI-relevant thing Apple announced.

    Personal Context7→8▲+1Cross-App Actions5→6▲+1MyAGI Alignment4→5▲+1
    source: CNBC live updates
  4. Apple

    Siri, rebuilt: conversational and “profoundly more capable”

    “Siri is now a profoundly more capable assistant, more conversational, so you can go back and forth like never before and get detailed, engaging answers.”

    The personal-context Siri Apple promised back in 2024 finally ships in iOS 27, with a new, customizable expressive voice and a major jump in system-wide dictation accuracy. It is the most consequential Siri change in a decade, though the real-world quality only proves out as iOS 27 reaches devices this fall, so we move the grades with a verification caveat rather than to the top of the field.

    Assistant Intelligence4→6▲+2Conversational UX4→6▲+2Agent Capability4→5▲+1
    AAPL▲ +0.4%intradayA shrug, not a pop. The market had largely priced a Siri overhaul in.
    source: Engadget liveblog