# Stop the AI Race March: Legitimacy Outside the Same Offices Shipping Models

Times of AI Desk · 2026-07-11 · Policy

[https://timesof.ai/2026/07/sf-stop-ai-race-protest-openai-anthropic-deepmind](https://timesof.ai/2026/07/sf-stop-ai-race-protest-openai-anthropic-deepmind)

> About 200 protesters marched from OpenAI to Anthropic and Google DeepMind in San Francisco, organized by Stop the AI Race, calling for a collective pause on new frontier training. Small vs capital stack; durable signal that safety, labor, housing, and power are merging locally.

July’s model week put **capability** on the front page; July 11 put **legitimacy** in the street outside the same offices. The protest is still small relative to the industry’s capital stack—but it is a durable signal that **safety, labor, housing, and power** are merging into a single SF political coalition.

On **Saturday, July 11, 2026**, roughly **200** demonstrators marched through **San Francisco** from **OpenAI** past **Andreessen Horowitz** to **Anthropic** and **Google DeepMind**, demanding the labs **pause new frontier training**. The action was organized by **Stop the AI Race**, led by former AI researcher **Michaël Trazzi**. Signs and speeches hit jobs, rents, energy/environment, and existential risk—the largest local anti-AI march in recent SF coverage of the campaign.

## What the march demanded

From San Francisco Chronicle reporting and organizer framing:

| Element | Detail |
|---------|--------|
| **Size** | ~200 people (students, AI workers, longtime SF residents) |
| **Route** | OpenAI → VC corridor / a16z → Anthropic → Google DeepMind |
| **Lead organizer** | Michaël Trazzi (Stop the AI Race; prior DeepMind hunger-strike activism) |
| **Core ask** | Collective pause: keep current models, **no new training runs** of larger/more general frontier models; reassign teams to narrow AI or alignment |
| **China clause** | Global agreement framing—pause only works if major labs and China participate |
| **Local politics** | Speakers (including former supervisor Dean Preston) tied AI boom to rents, housing, political influence |

Participants included AI-safety researchers (e.g., MATS fellows quoted by the Chronicle) who said they are not “anti-AI” but oppose an unmanaged **race**. Labs did not issue same-day formal replies to Stop the AI Race’s demand; organizers cite earlier comments from industry leaders about international collaboration (e.g., Hassabis on a CERN-like model) as partial rhetorical alignment without operational pause.

For labs, “stop the race” is unworkable without coordination; for cities, the march is a preview of **local regulation** fights that do not wait for Washington.

## Limits

- Crowd size (~200) is Chronicle/organizer-scale reporting, not a city permit census.
- No same-day formal lab replies; earlier “collaboration” quotes are not pause commitments.
- Global-pause framing depends on China and peer-lab participation organizers acknowledge as necessary.

## Sources

- San Francisco Chronicle: “S.F. protesters march on OpenAI, Anthropic and Google DeepMind to demand: ‘Stop the AI race’” (July 11, 2026).
- Organizer framing via Stop the AI Race (pause definition: no larger/more general frontier training runs).
- Contemporaneous social/video reports of the July 11 San Francisco march.