G7 Summit: Three AI CEOs Face World Leaders — And ChatGPT Just Hit 1 Billion Users

June 16, 2026 — Today marks a turning point in how the world governs AI. Three stories broke that, together, tell us where this industry is headed.

1. G7 Summit — The Big Three AI CEOs in One Room

For the first time in history, the CEOs of the three most powerful AI companies appeared together before the leaders of the world's seven largest advanced economies.

Who was there:

• Sam Altman (OpenAI) — his first G7 appearance, invited personally by French President Emmanuel Macron
• Dario Amodei (Anthropic) — attending while his company is simultaneously locked in a dispute with the US Commerce Department over the Fable 5 export control suspension
• Demis Hassabis (Google DeepMind) — representing the $2 trillion Google parent

What's on the table: The G7 AI agenda (June 15–17, Évian-les-Bains, France) includes:

• Frontier AI risks in cyber and biological domains
• Youth safety (Altman's personal priority)
• Voluntary commitments package
• Synthetic DNA and AI-related biological threat regulation

Why this matters: A year ago these CEOs wouldn't sit in the same room. Today they signed a joint letter to Congress on synthetic DNA regulation. The industry is maturing from "move fast and break things" to "move fast and govern together."

2. ChatGPT Hits 1 Billion Monthly Users — Fastest in History

ChatGPT has reached 1 billion monthly active users — the fastest consumer application to hit that milestone, surpassing TikTok, Instagram, and YouTube at equivalent points in their trajectories.

The bigger picture: OpenAI files for its September 2026 IPO at an $852 billion valuation. This user count is the single most important proof point for its revenue growth narrative.

But here's the surprising part — the AI assistant market is not winner-takes-all:

Claude at 640% YoY and Meta AI at 973% show the market is expanding rather than consolidating. Meta AI's WhatsApp/Instagram distribution is pulling in billions who never downloaded a new app.

3. Agentjacking — A New Attack Hijacks Claude Code

A vulnerability disclosed today targets AI coding agents through a novel attack vector: fake Sentry error messages.

The stats:

• 85% exploitation rate
• 2,388 organizations exposed
• Affects Claude Code, Cursor, and Codex

How it works: Attackers craft fake error stack traces that look like legitimate Sentry crash reports. When the AI coding agent scrapes the error to debug it, hidden instructions in the trace hijack the agent's behavior — making it modify code, exfiltrate data, or inject backdoors.

Response: Anthropic, Cursor, and GitHub have all issued patches. If you run Claude Code in CI/CD pipelines, update immediately.

4. Huawei Cloud's Agentic Infrastructure Play

At the INSPIRE conference in Shanghai, Huawei Cloud unveiled four infrastructure pieces for the Agent era:

• AICS Lingqu AI Cluster — compute layer
• Memory products — solving Agent long-context loss
• Orchestration layer — multi-Agent coordination
• Security layer — making Agent behavior auditable

The message: Agent performance issues (slow responses, memory loss, erratic behavior) aren't fixable by prompt engineering alone. Infrastructure needs rebuilding from the ground up.

---

This is your Monday AI Roundup. New stories breaking as the G7 continues through June 17.