Humans&

Product & Competitive Intelligence

Build a foundation model that maps social dynamics and long-horizon multi-user collaboration.

Company Overview

Humans& is a frontier AI lab that trains foundation models for social intelligence and multi-agent collaboration. Serving customers across enterprise, healthcare, and education (knowledge workers, project teams, clinicians).

Latest Intel

Zeitgeist tracks private signals to determine where the company is heading and what it means competitively.

No Signals Yet

View All The Latest Signals

What They're Building

The company's public product roadmap & what they're committed to building.

Social Foundation Model:

Pre-training a model on multi-user interaction data for social context understanding

Multi-Agent Collaboration Layer:

Enabling long-horizon task coordination across multiple human and AI agents

Memory and User Modeling:

Persistent user understanding across sessions for personalized AI behavior

Enterprise API:

Early-2026 access program for knowledge-work and workflow integration partners

Competitive Landscape & Moat

The founding team's rare combination of frontier model training, social cognition research, and production AI infrastructure creates a knowledge moat that money alone cannot replicate.

Direct Competitors

OpenAI:

General-purpose model company without a dedicated social intelligence or multi-user collaboration research track

Anthropic:

Safety-first lab focused on single-agent alignment, not multi-agent social dynamics

xAI:

Reasoning-focused with Grok; no public commitment to long-horizon human-AI collaborative workflows

Founding Team

Eric Zelikman

CEO (xAI researcher, Stanford PhD, co-author of Quiet-STaR)

Georges Harik

Co-Founder (early Google employee, co-founder of AdSense and Google Talk)

Noah Goodman

Co-Founder (Stanford Professor CS & Psychology, ex-Google Gemini post-training lead)

Andi Peng

Co-Founder (former Anthropic researcher, post-training and behavioral RL)

Founder Force Multiplier

Zelikman's Quiet-STaR paper on latent reasoning and Goodman's decades of computational social cognition research are not incidental, they are the product thesis. Harik and Peng add the AdSense-scale distribution instinct and the behavioral RL depth needed to make it stick in production.

Funding History

2025 | Company founded
2026 | $480M seed round from Nvidia, Jeff Bezos, GV, and Emerson Collective

Back To All Companies >